The SaaSpocalypse that wasn’t, alongside Atlassian CEO Mike Cannon-Brookes

Today, I’m talking alongside Mike Cannon-Brookes, who is cofounder and CEO of Atlassian.

Atlassian is one of those companies that all another business runs on — it makes crucial phase tools akin Jira and Trello that authorize group to arrange and oversee big teams, create shared databases of business information, and mostly authorize activity to happen. As you’ll comprehend Mike say, all of Atlassian’s products are really distinct expressions of a sole center platform, which really shapes how Atlassian itself is organized and how those products are built.

All of this method Atlassian is additionally correct in the center of the way AI is changing how all these companies activity — AI tools power be capable to appearance at all the distinct tools and systems you have and fair peruse them for you, making a big immigration to Atlassian’s phase small enticing. Or, if you buy the idea of the so-called SaaSpocalypse, AI power fair build all of these tools for you, destroying this complete category of businesses.

Obviously, Mike had a lot of thoughts pushing rear on this. So we spent several genuine period talking concerning the value of design, individual users, how AI is really expanding the use of Atlassian’s tools, and what this all looks akin a few years from now.

Of course, AI has additionally changed Atlassian. Like so many another tech companies, Atlassian did a round of layoffs before this year, alongside Mike saying he felt the business needed a distinct “mix of skills.” So I asked Mike what he thinks that mix of skills is and how AI is changing what it method to run a application company.

But Mike was additionally refreshingly straightforward concerning what AI can’t do, and he was consenting to shove rear on his own allies akin Cloudflare CEO Matthew Prince, who was just on the show talking concerning AI eliminating what he called “measurement roles.” This was a fun one — and shoutout to Mike, who recorded this from Atlassian residence basis in Australia, which meant he had a profound conversation concerning org charts at 5AM.

Okay: Atlassian CEO Mike Cannon-Brookes. Here we go.

Mike Cannon-Brookes, you’re the co-founder and CEO of Atlassian. Welcome to Decoder.

Thanks for having me, man.

I am really enthusiastic to conversation to you. I accept you’re in Australia?

Yes.

It’s tomorrow for you. What’s going to happen tomorrow?

It’s tonight. It’s a awesome day here. It looks beautiful. It looks akin it’s going to be an amazing day.

At this item in the AI news cycle, I awareness akin I can fair petition to cognize what’s going to happen tomorrow, and everyone understands the urgency rearward that question. There’s a lot going on alongside Atlassian. There’s a lot going on alongside your products. There’s a lot going on in the extremely idea of SaaS businesses and endeavor processes generally.

Let’s commencement at the extremely start. I think group cognize Atlassian. They cognize Jira. I started my occupation as a Trello person. Tell group what you think Atlassian is today.

Atlassian is a phase that enables businesses to collaborate and oversee activity throughout their teams. We nexus their endeavor teams and their innovation teams to grip the most challenging activity problems in a distinctive phase throughout any institution that is technologically driven.

For any institution that has application and innovation as its center rivalrous advantage, we create a broad phase that allows them to collaborate on content, oversee projects, and unleash the cognition of their teams and organizations throughout their business, their strategy, operations, all the way through to their assistance teams in all areas of their business. It’s a extremely ample phase now that we provide.

Just in conditions of what activity is today and what it power be in the forthcoming as we add additional and additional AI to these enterprises, perchance this is too reductive, but I have continually idea of Atlassian as a business that makes tools to assistance teams fig out what they’re going to do tomorrow. Jira is a ideal example of this. You don’t do the application engineering in Jira, but Jira helps you arrange ample teams of application engineers and document tickets and prioritize problems and tasks.

I was a Trello individual a lengthy period ago. My goal in existence now is to never use any endeavor software. I awareness akin that’s a sign of true achievement if you’re fair an iPad individual in archetypal class. I’m operating on it. I’m not saying I’m there yet, but that’s one of my goals.

But there was a period whenever I was the managing publishing company of The Verge and my job, straightforwardly, was to appearance at Trello all day and create certain everyone was doing the correct item at the correct time. And I’ve continually idea of Atlassian as that category of products. We’re going to arrange the processes of the company.

There’s item concerning AI that’s changing that. Has that concept of Atlassian changed for you? Is it changing faster since of AI?

Understanding what activity you have to do at an idiosyncratic flat akin Trello, at a squad flat akin Jira, or at an organizational flat akin our Strategy Collection for large-scale scheme and operations — that is certainly a big part of what we do.

Understanding anywhere processes are at — It’s small concerning activity to do than processes. If you think concerning a endeavor as a scheme of processes that are put together, everything in a endeavor is several kind of system. It’s a procedure in the gathering of all these, and how fine we execute them is what your endeavor is doing. The fundamentals of that haven’t changed. The way those processes run, the figure of them running, and how to fig out what’s going on hasn’t changed.

We think concerning Jira as a individual citation to work. When we say that, it’s a really engaging term because, as you said, the activity is not done in Jira. Developers don’t live in Jira. Marketers don’t live in Jira. Finance group don’t live in Jira. Jira is a workflow motor alongside extremely collaborative parts that enables teams to comprehend what they, their colleagues, and their broader institution are doing. That’s why agents and AI and another things are going to addition the speed of those processes. In a lot of ways, they volition addition the reliability, the quality, and the consistency of output of certain processes.

They additionally put additional accent on anywhere the humans, the judgment, the intuition, the individual bits that are unique, and the initiation arrive from. It’s our job to create certain that that’s all motionless understandable so person can fair appearance at any flat of their endeavor and cognize what the fuck’s going on. What are group doing? Are we successfully doing any it is that we’re trying to do, whether that’s assistance function in a backing revenue deal, elimination assistance array to activity out how many exceptions did we provision out in revenue this quarter, or it’s an engineering squad who’s trying to activity out, “There’s material happening all complete the place. What did we build this month? Or what are we going to build next month?” All of those things are incredibly difficult to do at measure and quantity in a business.

One of the additional crucial features of the tools you’ve built is legibility. It’s person experience. You have a big repository of exceptions stated out by salespeople, and you need to appearance at a dashboard and see how the company’s doing. Or you have a entire bunch of tasks, and Jira volition let you see at assorted distinct levels of idea how many tasks are being completed at any stated time.

That was item that was extremely crucial for individual managers. A lot of the capability unlock of businesses as they added application akin the application Atlassian makes is to fair be capable to see and coordinate additional things at a higher scale.

The rotate that’s coming is that perchance those databases aren’t going to be looked at by individual beings anymore, or they’re going to be fed into additional synthesized databases, or AI is going to create several another benevolent of data at distinct levels of abstraction. I cognize you’ve got a merchandise called Rovo which feels akin the new visage of Atlassian in several ways.

What’s the way here? What does the end province appearance like?

There’s no doubt AI is going to assistance alongside a lot of those processes. There’s no doubt it’s going to choice up several steps in a process. Most businesses at the instant have a process, and they’re using AI to automate 80% of this step. It may not be 80% of the distance. It can be the width. If you think concerning that revenue agreement elimination procedure that you fair mentioned again there, if AI can obtain 80% of the elimination — say this client wants 45 days, not 30 days — and we can compose a comparatively readable document, AI is beautiful fine at study the document and saying, “Yep, this one’s fine.”

But there’s going to be a client that needs a lot additional complex exceptions. They say, “I desire to pay in 90 days in this way, and I desire this item and that thing.” That needs to go to a individual for intuition and judgment. So 80% of those may be easier done alongside AI now since it’s extremely capable at handling our rules in this particular part of our business, and that’s great. It’ll addition the speed of the process, the norm of it.

It doesn’t obtain distant the need to cognize what’s going on. Someone in the endeavor volition motionless need to know, “How many exceptions did we give?” What happens in most of these areas is businesses discover way additional ways to create themselves measure and efficient, and they discover way additional ways to create higher-quality products and services alongside AI. I see this all the period anywhere they are using AI in ways that group don’t expect. It’s not all concerning mechanistic efficiency. There’s a huge amount of norm that is capable to be measured and understood that was never before. The agreement elimination procedure is a fine example since it’s really a extremely individual process. It’s not normally rule-driven. The rule-driven parts are easy. It’s the non-rule-driven parts that are additional complex and necessitate judgment.

The endeavor motionless needs to know, for that process, “Do we need to add additional group to this process? Do we need to alter it? How much is going through it?” You motionless have a lot of activity to comprehend the stream of activity about an enterprise, which gets additional complex as the flow’s quantity goes up a lot.

I’m asking since I’m headed towards a sequence of questions concerning the expected SaaSpocalypse, and I’m extremely inquisitive anywhere you think tool providers akin Atlassian fit next to the frontier models are fair going to eat additional and additional capability. Eventually, Claude volition fair do this for you. You’ll say to Claude, “Run my endeavor for me.” In the backend, Claude volition scorch a bunch of tokens, and it’ll fair do it all for you, which is a commitment that many, many group appear to accept in.

You’re describing a slightly distinct way anywhere the tools get smarter or additional capable since of AI, and you’re motionless adding a lot of individual judgment, but at no item is a frontier example fair coming and operating your complete endeavor for you.

If a frontier example is capable of coming and operating the complete endeavor for you, if it’s really operating the entire thing, equal assuming four or five years value of increase, you have a comparatively uncomplicated business. Most businesses aren’t simple. They are earth conglomerations of rules, compliance, laws, personnel in distinct places, individual inconsistency — which is the identical as saying individual ingenuity — that they’re putting it all together to try to provision products and services for their customers in their industry: healthcare, university, finance, automotive, or any the industry is.

They’re trying to vie alongside several another business. They are motionless going to have a huge figure of humans, I would contend additional cognition workers, additional developers, in five to 10 years period than today since the capability to do tasks to vie volition go up.

The bar for competition volition go up in nearly all these industries. You volition motionless have a huge amount of things to go do. Most of the things that we think are uncomplicated volition get done for you. That is the former of technology. I’m a big fan of saying AI is fair technology. What we can do alongside spreadsheets, what we can do alongside the internet, what we can do alongside ordering products online… You think, “Once Amazon arrives, we’ll all fair never have to go to the shops again.” It doesn’t happen.

I heard person talking concerning financial institution branches as an analogy, which I idea was a really fine one. Around 2000 whenever mobile banking came along, everyone idea financial institution branches were dead. Bank branches started decision down. Everyone wasn’t going to go to their branch anymore. It was a beautiful negative place. At smallest it was in Australia. I’m beautiful certain it was in America as well.

If you appearance complete the final 10 years, financial institution branches are increasing. People are like, “Huh, this is a narrative violation. Why is this?” And the answer is financial institution branches did near down. People did go to net banking. They did use their mobile app. They don’t go into the branch anymore alongside a passport to say, “I’d akin to dispatch $100 to Nilay, please. Here are his details,” and compose them all down.

But financial institution branches have adapted. The logic they’re expanding is there is a huge client assistance component to a financial institution branch. The services they provision today are totally distinct than what they provided 25 years ago because of the innovation of the financial institution and the preparedness of things. They are much additional concerning assisting you alongside higher-level processes and services than they always have been.

That turns out to be, presumably, profitable for the banks, which is why they’re beginning new branches. Bank branches are going up. The things those branches are doing are totally distinct to what they were 25 years ago. I think you’ll see the identical item alongside cognition workers and everything else, that it volition motionless be incredibly important.

I don’t see that part changing.

I am operating my way up to getting Jamie Dimon on the display to discover out what happens in all the Chase Bank branches in New York, since I don’t really know, but we’ll set that aside. Jamie volition arrive on the show. We’ll ask, and we’ll fig it out.

One of the reasons I’m asking all these foundational existential questions is to comprehend how you awareness concerning Atlassian’s association to businesses and how businesses power alter as we add new technology. I concur alongside your broad framing that automation innovation arriving to a endeavor is a beautiful acquainted phenomenon, but there’s item alternatively going on alongside AI that is either making that faster or additional dramatic. It’s evidently happening to your endeavor as well. You’re the CEO of Atlassian, and you’ve made several changes about AI.

I desire to ask you the Decoder questions now. How many group is Atlassian today? And how is it structured?

We are, I think publicly, 12,000-13,000-odd group about the world. We are organized globally, I would say. We are “Team Anywhere.” Employees have the choice to arrive into an agency or not. About 60 percent of group arrive in three days a week or more. About 25 percent of group don’t arrive in a sole day per week. We run as a globally distributed business as we have done since the start. When you commencement in Australia and San Francisco, you benevolent of get used to the Pacific Ocean being a small bit of a distribution. This was before Zoom. We used to have these elephantine Polycom systems in gathering rooms. We don’t have those anymore. Technological advancement is nice.

We have customers all about the world. We have personnel all about the world. We are organized functionally, I conjecture you would say, alongside a lot of matrices internally. I don’t cognize if that’s the answer you’re looking for.

No, that’s entirely the answer I’m looking for. What do you average by “functionally alongside a lot of matrices”?

We have a CRO. We have a CFO. We have two CTOs at the moment. We have two ample merchandise groups, all of which has a CTO and a chief merchandise officer, one in the endeavor and emerging flank and one in the forthcoming of teamwork side.

Our products are organized into collections. We have collections of apps that are all built on a sole platform. Well complete fractional of our R&D is on the platform. The apps are increasingly a smaller amount of the proportional total funding in building. More and additional is on the distinctive Atlassian platform. Customers employ us as a phase throughout their endeavor to get activity done, not for a sole application.

That requires a matrix internally. We don’t have salespeople per product, if you desire to think concerning it that way. The sales, client success, and the FDEs are all customer-related motions. They’re organized as per the customers and their geographies and their size and scale. Then we evidently have to have designers and merchandise managers on particular products who attention concerning the intent of a product, so you end up immediately alongside the matrix between how the customers are organized and how the products are organized.

We have a elephantine platform. We have the specialized phase that runs throughout all the products. It’s a extremely ample phase team. We afterward have backing and endowment and another things that are additionally broad processes. Naturally, you end up alongside collaborative areas that you have to get together, and we oversee that through our functioning processes.

Do you use the Atlassian phase to run Atlassian?

We do. We run entirely on the Atlassian platform. I used to say we’re the figure one person — we’re not anymore — of our scheme and operations products to comprehend what our endeavor is focusing on, what our large-scale goals are, how the people, the investments, and the specialized systems all arrive together, and what is going correct and incorrect throughout all of those. We are a big believer in our Strategy Collection to run a large-scale business.

One of my favorite questions to ask endeavor application CEOs is how much they personally use the endeavor software. How much do you personally use your products?

An awful lot.

What’s your figure one characteristic petition for your teams?

My figure one characteristic petition is normally continual UI congruence. I conjecture that’s the category that group would put it in. I’m a bit of a stickler for scheme and consistency. We desire to accomplish eventual consistency in design, in which we’ve arrive a huge extend in the final five years. In the final two years, we’ve been really world-class at doing it. That’s normally my biggest area anywhere I use a lot of our products on a regular basis. As such, I’m moving about the platform, and I’m looking for ways anywhere this item complete current happens this way. They do it improved complete there. We’re trying to accomplish that eventual consistency throughout a ample merchandise exterior area.

I use Dia, our web browser. I think 96 to 97 percent of Atlassian now uses Dia on a regular basis. I mean, that’s a regular basis, so I exactly expend many, many, many hours of the day on it. We are construction a web browser for ourselves since we’ve reached the item that we feel, “This item should fair be improved for cognition workers.” We’ve already built, I would argue, the finest browser for cognition workers in the world, and it’s getting much, much, much better. Keep observing this space.

So, yes, I would say there’s no 24-hour duration that goes by anywhere I don’t use one of our products. I doubt there’s a six-hour duration that goes by anywhere I don’t use one of our products at the moment.

I desire to conversation concerning Dia in a second, since I’m extremely inquisitive concerning that acquisition. Obviously, we covered The Browser Company extremely deeply at The Verge. That was a big acquisition for you guys.

Setting that apart for fair one second, among the things I’m particularly inquisitive concerning current is the difference between the phase and the products that run on the platform. Again, perchance this is fair reductive, but it appears akin there’s a big tier of capability, and the products are expressions of those capabilities in distinct ways, but they’re all operating on the identical center platform.

Is that how you think concerning it?

Yes, absolutely. Yes. We have to build a distinctive R&D platform, whether that’s how we conversation to AI gateways, whether that’s how conversation works, whether that’s how automation and character and logging and governance and compliance and satisfied classification work, whether that’s how our residence and our hunt motor work. There are many, many additional specialized capabilities that have to be shared.

Then you have all of the next tier up, and customers desire that to activity the same. It doesn’t matter if I’m in Confluence or Jira or Loom or the Service Collection, automation shouldn’t be distinct per location in the world. That’s extremely expensive. It’s extremely hard. It takes a lot of time, but it delivers a huge amount of client value and loyalty. Most vendors don’t try to do that. They exactly don’t try.

Then you have a tier up, which is important, which is the consistency of procedure and how satisfied can be mixed. Operations and processes can be blended throughout applications. If I open a Confluence leaf inner of Jira, I desire that to awareness akin a Confluence page, but I desire it to be inner of Jira. I don’t desire to depart Jira to open it. I fair desire to peruse it, near it, keep going. There’s a lot of UI layer, a kind of recomposability, if you desire to think concerning it that way.

Lastly, all use has to do its job. If you blend endeavor application together, you end up in a earth of hell. People have tried this many times. You desire a tool to do a job. You desire to choice up a screwdriver or a hammer or saw and awareness akin that thing’s going to do the item you hired it to do. The interface is designed to assistance you do that task. That’s anywhere there is a difference in the interface tier and the communication layer, between those tools.

The blend of those is anywhere a lot of our scheme challenges live. For example, our conversation layer, Rovo, is amazing. I would contend it’s among the best, if not extremely near to the finest endeavor conversation products in the world. Our customers continually say that. “How arrive you’re improved than insert-famous-brand-here?” There are a lot of reasons why we execute improved in evaluations and evaluation alongside Rovo. It’s not since we create a foundational LLM, I think, which is the flaw of thinking.

However, using Rovo inner a product, if you’re going to Rovo for the purpose of Rovo, is awesome. You have a inquiry to ask, you desire to hunt throughout everything, fine, you go to a particular interface. If you’re inner of Loom or Jira or the Service Collection and you fair desire to open up chat, ask a question, you desire it to comprehend the environment you’re in, but you desire that procedure to be accordant throughout the platform. That’s what we’re striving to provision to customers.

One of my changeless tropes on the display is that if you depict to me your org chart, I can inform you 80% of your problems. The construction of the business leads to several benevolent of natural government in several benevolent of way.

The strain between “we have products that are designed to do a job” and “we have a phase that contains the center capabilities” is beautiful fine known.

Mm-hmm.

Maybe your merchandise teams are going to desire phase capabilities that don’t be yet or they volition desire distinct phase capabilities from one another. How do you interrupt those ties? How do you create those calls?

We have a figure of ways that we are distinct and we try to do that. First, I would say most importantly, we have a platform-centric CEO and co-founder who thinks in conditions of platforms. Thinking in Systems is my favorite book, man. I have a copy on all desk.

You have to put the phase archetypal collectively, and afterward you go resolve a bunch of problems. That does not average genericism wins, but it does average whenever you create trade-offs, you have to expend the additional attempt and period to create it phase centric, to create it work. And that is a everlastingly task. It helps to have a CEO, if you think concerning the referee throughout all the functions, that’s phase centric. They cognize anywhere I’m going to district and so we’re going to end up there.

Secondly, we have scheme as a very, extremely elder role. We have a enormous funding in design. Our chief scheme authoritative reports to me, sits on our administrator team. We now have three, four, five erstwhile chief scheme officers in the endeavor from another extremely ample businesses, several larger than us. We have a huge funding in scheme and experience. Especially in the AI era, that’s getting always additional crucial actually, much as Claude Design or item would inform you the opposite. I don’t think so. I really think [when it comes to] scheme and the cognition of software, if the disbursal of construction goes down, the norm and differentiation is on the cognition and the design, and that’s not item that’s uncomplicated to do. That requires a huge amount of sensation and judgment.

Lastly, I think we aren’t fearful to put things together in different ways in the institution anywhere we need to resolve problems. The latest example of that likely is our inner IT and engineering function now reports to our chief group authoritative and I’ve blended those two roles, which is unusual. “Why is HR operating IT? That doesn’t create any sense.”

It really does, I would argue, at the instant since of the things we need to do to change our endeavor internally continually, which is that if I appearance at AI and becoming additional AI native as a business, I end up alongside two piles of projects. A heap of projects that connect to talent, how I need to alter and develop the group we have, get them using these technologies, get them thinking in certain ways. I have a huge endowment problem. Hiring, growing, training, changing group is extremely difficult to be competitive.

I have a huge scheme problem. I need to alter our inner systems to have MCP servers. I need to create certain that everything is operable. I need to create certain that we have all of these scheme changes so that AI rolls out. How are we managing our token spending? There’s a huge amount of scheme change.

I ended up refereeing a bunch of, “Is this a endowment issue or a scheme problem?” As I say, that doesn’t create any sense. Usually the answer is this one’s 60 percent talent, this one’s 40 percent system, and this one’s the contrary way around. I’ve put them all together inner the endeavor so that our scheme changes to use AI internally and our endowment changes to use AI internally are in the identical function correct now, which is a small odd.

The group who provision the laptops and the group who oversee compensation activity for the identical group, but it really makes logical awareness according to what we need to alter concerning our endeavor at the moment. What is this three-year epic that needs to change? We arrange about the problems we need to resolve and the goals we have as an institution whenever we need to.

This is a brilliant segue, since I wanted to ask concerning this administrator you have. I accept the ceremonial heading of your HR individual is now “chief group and AI enablement officer.” That’s fairly a title. And the item you’re making concerning the systems of the endeavor needing to alter foremost to several alter in your endowment mix is coming true. In March, you laid off concerning 10% of the company. I watched the video you made in Loom concerning that.

Your item there was you’re not necessarily replacing group alongside AI, but you had identified a alter in the mix of skills that the business needed and you were going to create these changes proactively.

Walk me through this. What made you say, “AI is current sufficiently that the mix of skills is different”? And what changes did you create to your systems to create losing 10% of your group productive or worthwhile?

We have to commencement alongside the environment. The surroundings we’ve lived in in innovation for a during is continual hiring and growth. The markets have changed in their perspective on that. Because among the ways that we’ve managed changing businesses complete period has been hiring group in distinct areas. If you have red dots and blue dots and you say, “We need additional green dots,” you fair say, “We’re going to employ all green dots for a during until we have a blend.” You oversee it that way, that is no longer a feasible path, firstly.

Secondly, we have certain areas of the business, our AI roots and area-

Wait, actually, can I fair ask one inquiry concerning that? You’re benevolent of describing the COVID era over-hiring that all tech business was doing, and you’re saying that’s benevolent of over?

Yeah. People employ at the charge of the finest decisions they cognize at the stated period and what they’re encouraged to do by others. Hopefully they don’t appearance at everybody about them, but I cognize that a lot of CEOs do appearance at the group about them and say, “I should do that.” I continually think those are mediocre choices.

Over-hiring, I don’t know. Pre-COVID, we were a few thousand people. You cognize what I mean? So we are four times, I don’t know, three times the size we were pre-COVID probably. Again, we’re accelerating as a business. We grew about 30% final fourth on a $7 milliard run rate, which is the fastest we’ve grown in concerning two years. Customers are opting for additional acceleration.

I’m fair trying to comprehend the “you can’t fair employ a bunch of green dots” comment a small additional plainly since I’ve heard from many of your peers, “During that duration we fair hired everybody that we could.” It appears akin that’s not specifically what you were experiencing, but it rhymes in several way.

It does rhyme, but there are periods anywhere that is the way to go. That’s not the current three- or four-year duration of time. That’s the archetypal item I say since afterward you need to activity at how we alter our accomplishment mix.

We have a huge amount of inner programs to try to do that. We additionally have certain areas, our AI products and services specifically, and our endeavor revenue areas are expanding very, extremely fast. They’re doing fine and we desire to allocate additional in those areas. When you have certain areas you need to allocate additional in, you run into extremely difficult choices anywhere you say, “I need to allocate additional in these two areas. I can’t develop the general in a important way. I have to oversee how to passage the institution between that.” Which leads to a lot of extremely difficult choices.

We are motionless hiring really difficult in those areas in command to alter the accomplishment mix, but you additionally have a inquiry of how accelerated we need to do that. It doesn’t, I continually say, obtain distant from any “internal training programs.” I say that in air quotes in conditions of how AI is adopted and used. It’s additional about learn, play, share, and mutual learning. It’s small akin an L&D, in conditions of “I’m going to sit downward and observe a category on how to do this,” but that doesn’t obtain distant from that. It’s a blend of all these factors as we’ve massively changed the endeavor to be really earth foremost at AI merchandise betterment in the final two years. There’s no doubt that’s been a fast journey and we’re trying to move as accelerated as we can to get there and to remain ahead.

One of the things I’ve heard from a lot of CEOs in your stance is, “The value of our elder group has increased.” There are additional elder application engineer job listings out in the market correct now since as you’ve said, that judgement and that sensation is extremely crucial and perchance the tools can fair build that material for you.

I’m inquisitive for your perspective on design. The idea that all of our designers volition not be enabled to fair container code is extremely tempting to a lot of people. There’s a third perspective that says merchandise managers roles do everything and all these roles are going to arrive crashing into all other.

What mix of skills specifically did you acknowledge and how are you prioritizing in this kind of traditional triumvirate as additional and additional group get the capability to do all the another jobs?

It’s extremely uncomplicated for group to have extremely binary takes on these things. I discover it amusing that if you put an engineer, a designer, and a merchandise director in a room, they’d all sit there and they’d all say, “Oh man, it sucks to be you. Your job’s going to go away.” I say, “Ah, that likely is your answer correct there.”

We have a lot additional merchandise managers and designers who are penning code from prototypes to really shipping code than we always have before, that is for sure. We have a lot additional marketers and backing group and HR group who are doing the identical thing, by the way. The code that they’re penning and shipping builds on two things. One, a phase that has the engineering robustness to be capable to measure and provision the customers a European-compliant data inhabitant resolution that is necessary, the lawful constraints of nation X, et cetera.

You have to build a specialized surroundings that is capable to get out to a client alongside the required flat of client safety. That is a non-trivial problem. That’s not item you’re fair going to vibe code. If I’m vibe coding my own resolution to item for my household, fine, it can fair activity fine enough. I’ll go fix it if it doesn’t. If I’m doing that for a ample organization, I need several another controls and rules about it.

Secondly, the specialization of those areas all doesn’t go away. The individual judgement required in distinct areas is akin the edges of the roles blending. They’re capable to do a small bit additional of person else’s job. I have a lot of designers who are capable to container prototypes that are operating code prototypes alternatively of a image or a counterfeit prototype in conditions of a mock-up. That’s great. I motionless see them sitting in Figma a lot and doing that.

I have a lot of engineers who are sitting in Cursor or Claude Code or any on a regular basis, penning features and construction things, and they’re capable to do additional scheme than they were before. That doesn’t substitute the need for the another job. It increases the norm of the general software. It increases the speed of collaboration between group if you desire to think concerning it. If they all comprehend all other’s role, they can have a much additional fulsome conversation additional quickly. The specialized nature of what they do goes up. It becomes additional spiky, which is really, really important.

The capability of construction platforms — that’s going to get overused current — becomes additional important. Just this week, for example, one of our designers wrote an engaging blog article concerning how we now have an MCP (Model Context Protocol) and a CLI (command row interface) for our scheme scheme internally. We have an award-winning scheme system, atlassian.design. It’s extremely public. We publish all of our scheme system.

I’ve continually stated in the function of design, 50% of their job is to create our scheme scheme and to keep it scalable, flexible, and buildable. 50% of their job is to resolve the hardest 5% of problems in scheme since the fractional anywhere they’re construction the platform, what used to be templates and scheme documents that are now CLIs and MCPs to do scheme internally in the way that we desire scheme to activity so it’s accordant and coherent, that is a really difficult issue to build that.

However, they’re unleashing thousands, if not nearly tens of thousands of engineers to build better-designed things whenever they do that. Their leverage is a lot higher in construction a scheme scheme than it is in solving, “Help me alongside this screen.” It’s the identical in security. It’s the identical in client assistance and client assistance actually. If you think of bug fixing as the phase on which you decrease client service, additional and additional of that function is required on construction those scalable templates, platforms, ways that the another functions can build on top of so they can do that thing.

If a merchandise director is vibe coding a feature, as in they’re fair stabbing about several kind of code generating area and it’s an area anywhere it’s not infrastructural, it doesn’t need to scale, it’s fine. It should leverage our complete scheme system. Turns out, out of the box, that’s not extremely workable. We’ve fair shipped MCPs and CLIs internally, which are on V4 that are amazing at understanding our scheme scheme so the designers can develop that scheme and the features that are being built volition continue to equivalent what customers anticipate in that congruence and consistency. That gets us huge leverage overall, but it’s a non-trivial thing. It doesn’t average the designers are going to go away, the engineers are going to go away, the merchandise managers are going to go away. I don’t think any of those roles disappear.

I’m continually joking that Decoder is basically a display concerning org charts, which method I think I cognize what the next five years of the display is going to be concerning since we’re on the cusp of several of the weirdest org charts in former in AI.

Matthew Prince from Cloudflare was just on the show. We talked a lot concerning his Wall Street Journal op-ed anywhere he said, “I’m going to blaze all the measurers at the business since I can substitute them alongside AI. They can appearance at all of my systems and do all the auditing for me and we’re going to have small center managers since what they’re really doing is measuring and auditing things and AI can do that.” There are reports that Mark Zuckerberg wants managers to have teams of 50 empowered by AI location inner of Meta. This strikes me as not a awesome idea, but we’re going to see how it goes.

Where are you? Are you in traditional org chart? Are you at untamed ideas? How is this coming together for you?

I would say I cognize the two of those group extremely well. I’m certainly additional measured than that and I think the actuality of how the earth volition be. I think it’s uncomplicated to item to extremities. Is person location a director alongside 50 people? Sure. Are they effective? Maybe. People continually conversation concerning Jensen Huang. He’s evidently extremely effective. He’s a fine dude. He runs things in a extremely weird way. Is that scalable and repeatable? I’m not certain it’s scalable and repeatable. I’m not certain if you took an complete set of ample companies and made them all activity that way, they wouldn’t have worse results as a collective.

Part of endeavor is scalability and consistency, and being capable to be understood and managed that doesn’t necessitate uniquely gifted individuals to comprehend a extremely odd construction as they move about a business. I doubtful you are going to get additional function blending and group capable to grip multiple types of skills. I doubtful they’re going to be additional generalist. The edges of the function volition blur into the roles about it. A merchandise director is capable to do a bit additional promotion and a bit additional engineering.

That applies nearly all throughout the way. Finance group are likely capable to do a bit additional lawful than they were before. They’re capable to comprehend the lawful frameworks we have internally alongside agents and that’s great. I don’t think that changes the roles themselves. It may create teams a bit broader. I do concur alongside Zuck’s broad direction that organizations volition rotate into a broader triangle, a fatter, wider triangle. The companies volition get wider and hence small deep.

Broadly that’s a true direction. Do you go to 50 individual teams? Then the extend of authority is akin three people. The organizational degree is akin three or something. If you multiply them by 50 at all level, if there are two and a fractional thousand group alongside two levels of reports, I would inquiry that. That’s likely not an average. I don’t cognize concerning this term, “measurers.” I don’t cognize that anybody could sit about in a cognition activity job and say, “My role, my intent is to measure things. I fair stroll about alongside a ruler and I’m like, ‘That’s a foot.’”

I cognize there are a lot of jobs anywhere — I cognize what Matthew Prince is saying — that becomes the productive output of a role. We’re measuring it for a purpose, I would believe, and we are going to activity out how to create that additional scalable. My probable way of looking at it is a small bit different. There are two types of jobs. If you desire to simplify an organization, there are supply-constraint and there are demand-constraint jobs inner a business. Input border and output bound.

But what is the constraint on a job? Look at a promotion or a innovation function, a imaginative function versus a lawful or a client assistance function. Legal and client assistance are input bound. If I deploy a lot of AI, I don’t abruptly get a lot additional contracts. The figure of contracts coming in, if I’m looking at NDAs or leases or trying to comprehend the company’s lawful stance in certain areas, I can possibly procedure that activity faster alongside AI. Through systems, I can be additional productive in how I do that. That is one set of roles that you can appearance at nearly everything in a endeavor procedure and say that’s input bound.

I’m going to oversee that alternatively than things that are output bound. If I think concerning promotion or engineering, unless I run out of roadmap items — I’ve never heard any engineering squad in my former of 25 years of doing this anywhere they say, “Actually, we’ve reached the end of the roadmap and we’d akin several additional ideas of what to do next, please.” Said no one construction application always that has a ample client base. Their output, their ingenuity is border by their own initiation and their capability to arrange groups of group and processes and everything else, which is what the Atlassian platforms do. That is their limiting factor. It is not the individual capability to create ideas. It may be that you have small input border functions and areas as a proportion of your total employee basis and you have additional imaginative output border functions.

How much is this individual ingenuity and ingenuity and imagination? A possibility. That’s apt the way things change and trend complete time. That’s going to obtain much longer than group think. We’re big in technology. “Next twelvemonth everything volition be akin this!” and it takes 10 years to happen, but it volition happen. I think I’m extremely eager this term. I keep definition to put up a rant online concerning it. I’m stealing another friend’s citation here, but we used to conversation concerning data asymmetry and how much certain group cognize things and that is their origin of power or advantage.

I do think we’re moving to — he stated it and it really hit me akin a bolt — imagination asymmetry. We’re moving from data asymmetry to imagination asymmetry anywhere your capability to create and think is going to be far additional crucial as a rivalrous advantage for your team, your job, and your endeavor than your admission to information, your authority of information, your understanding of information. That part is going to be cheap. That’s what LLMs are really fine at. What they’re not fine at is imagining things. They didn’t decide to commencement Atlassian. What endeavor has been started by an LLM? Zero in the earth ever.

Anthropic has run a unsuccessful vending device for multiple years now. Come on.

But it requires a individual to say, “Here’s what we’re going to do. Let’s neglect at operating a vending device today.”

They were likely sitting in forefront of a vending device cursing the item whenever it didn’t autumn their candy bar and they said, “I’m going to do this improved alongside AI,” which is likely the way that most of those companies think concerning all sole issue as they’re staring at a vending machine. It required a individual being to say, “Let’s go do this.” It required them to compose a bunch of material to go create the vending device procedure code and try it. We all learned a lot hence of this specialized experimentation.

Let me ask you the another Decoder question, afterward I desire to use all of this to what is broadly happening to endeavor application and SaaS. The another Decoder inquiry I ask everybody: How do you create decisions? What’s your example for doing that?

Someone told me you were going to ask this question. I said, “What a broad, open-ended question.”

I would say I create decisions by talking a lot. That’s what group internally would really say. I discover the need to conversation things from distinct angles extremely important. I akin to let another group talk. I akin to get all the data together. I do accept there’s evidently a period anywhere we need to halt and fair move in a certain direction. The additional crucial part for me, fairly than how I create decisions, is I really akin to explain decisions to people. At Atlassian we have a value: We’re an open company, no bullshit. I inform group all the period that does not average transparency. We don’t owe you as an employee all of the particulars rearward all decision. All the sausage making is not useful. It’s overwhelming to people. It doesn’t assistance us move fast.

We do owe workforce clarity on why we are making a decision and an clarification of stated decision. That’s among the most crucial things managers can do for training, for assisting their endowment develop is not, “We’re going left,” but, “We idea concerning it. We looked at the distinct options. We considered straight, right, and left. Here are the pros and cons. We’re going left. This is the logic why we’re going left.” Often group don’t do that since group don’t akin the decision. I say that all the time. When you talked concerning our endowment mix changes in March, we had to create a lot of difficult decisions to run this business. I desire to create certain that we are continually explaining those decisions to group in the clearest and simplest way we can, why we’ve made a decision and what the decision is without any dressing.

I really don’t akin these company emails anywhere I peruse the entire item and think, “So what are you saying exactly? You’ve got a lot of words here.” I say, “Here’s what we’re doing. We need to alter our accomplishment mix. Here’s why. You should have the correct amount of empathy to explain decisions, particularly difficult ones, but you additionally should not anticipate group to necessarily akin it. We create a lot of difficult decisions anywhere I’m certain for 60% of the company, 7 percent of the business think it’s a awesome decision and 40 percent think it’s a awful decision. We build rely as leaders by explaining our decisions and being apparent concerning why we’ve made it and how we’ve made it. Hopefully we enhance decision making. I continually say decisions get to my array whenever they’re presidential decisions. I think it was Obama that called it that, anywhere it’s 51 to 49, that’s why it made it to my desk.

If it was 90% certain what we should do, it should not have made it to my array in the archetypal place. I lone get difficult decisions. That’s a well-operating company. Any chief have to be in that situation. If it’s an apparent decision, why did it create it to you? That afterward requires additional idea and additional clarification concerning those decisions. That’s likely the most crucial part than any example we use to create them.

Let me put all of this under several pressure. You’ve described Atlassian as a platform, and a lot of an organization’s decision-making and data and capabilities are in that platform. “What are we going to do next? How are we going to do it? Is it working?”

The complete item of the SaaSpocalypse was that we don’t need your beautiful person interface. We don’t need your scheme surfaces or congruence of scheme throughout products. All we need is your data. If I unleash Codex upon your data store, it won’t equal matter that it’s the identical platform. You can have 15 distinct vendors and the AI volition create itself several middleware and peruse and synthesize all the data and that’ll be great. Actually your endeavor can jump onward in meaningful ways.

My favorite example of this from a former era was I continually had Daniel Dines from UiPath on the show. Instead of modernizing your healthcare billing system, he would fair build robots that would use Windows 3.1 for you. This is a extremely first type of what I think all AI business is marketing correct now, which is, “Instead of modernizing your backend and spending all that prosperity on a migration, our robots volition fair use your horrible old databases and current to you a new person interface.” That’s one type of SaaSpocalypse.

It is.

I see it coming for a lot of businesses. How do you anticipate to alter all of Atlassian against that? Because a lot of what Atlassian sells is upgrade, upgrade the entire platform, use one merchandise and use all the products, use the entire platform, modernize your complete business. The another answer is that it’s akin $200 a duration and you can fair tokenmaxx your way into the future.

There are a lot of distinct parts of that answer. You’re talking concerning the headless world, I guess. I say that headless is brainless. It is exactly the identical thing. It is submission in dilatory motion to do specified a thing, if you’re a application vendor. I say that extremely carefully since again, it’s one of these answers that requires nuance. We have among the most-used MCP servers in the world. We’ve fair shipped a elephantine upgrade to it, rewritten it from scratch. It is massively used on a regular basis by millions of group and it’s amazing. [Along with] MCP servers, we’ve shipped three or four CLIs that are additionally used incredibly heavily.

The capability to have inter-operation of tasks is very, extremely crucial for customers to get jobs done. I totally concur alongside that. However, group who use our MCP server and our CLI really develop faster. They create additional issues in Jira, they develop their seats faster. We’ve stated this on community earnings calls, I think it’s akin at smallest 5% faster. They develop their ARR at twice the charge of customers who don’t use our CLI and MCP servers. The customers using these things are expanding their Atlassian expend faster than they have before. That earth I totally accept in.

When I comprehend the term “headless,” though, I think that is the lone earth that we’re going to be in. We fair market an MCP server. That is the brainless part concerning it since afterward your application really has no value. If you really fair have a database, fine. If your interface sucks, and I think a lot of endeavor application has horrible, horrible, horrible interfaces—

This is why I ask all the endeavor CEOs if they use their own products. It’s really asking, “How seriously does your interface suck?”

I’m a big fan of record the Loom and saying, “Guys, this doesn’t create any sense.” If you do that, that yak-shaving complete and complete again, you end up alongside amazing interfaces and that’s why tens of millions of group awaken up to our software.

The funding in scheme and cognition that I talked concerning is the antithesis of that. The earth is extremely distinct than those views. We’ll see in a few years period who’s right. The way the Atlassian phase plant is not to be a sole platform. I am likely the lone CEO that goes to customers, enormous large-scale customers, and says, “Do not centralize everything on my platform.” I do not accept that is a viable scheme for a CIO. I don’t think it is.

Our phase is extremely connected. It continually has been. One of the things that Jira, Confluence, Trello, and the Service Collection do is engage alongside many, many, many, many another applications. You appearance at any activity item in Jira that’s been going for a while, it is complete of links to Workday and Salesforce and ServiceNow and Figma, links to origin code, links to Google Drive and Teams. Links are what build our environment graph. The logic we have the finest environment chart in the earth is that we have a lot of links to commencement alongside and we have had a 10-year former of having to comprehend how to be collaborative alongside another applications.

Our applications are by nature extremely connected to the remainder of your IT world, the remainder of your innovation world. We don’t desire to be a authority tower. It’s a foolish analogy. Every CIO has been told, “Just rely us. We’ll agreement alongside all the problems.” It’s never worked. I don’t think it always will. The actuality of innovation is you’re going to have more—

Once you add AI to the mix, perchance it’ll activity this time.

No.

This is the item that is being sold: You have now a mighty intelligent orchestration tier that volition eventually acknowledge the dream.

Three parts snake oil, two parts bullshit. The actuality is it volition assistance all the applications if built fine to activity improved alongside all other. There’s no doubt. Look at Figma and Atlassian. We have a profound partnership. We use all other’s MCP servers a lot, and we can build improved solutions for promotion teams and designers and innovation teams between the two platforms than we always have been capable to before, since we can comprehend all other’s data in ways that we never have been capable to before, and we can resolve workflows on distinct ends.

If you’re sitting in a profound design, you’re sitting in Figma. It’s a ocular experience. If you need that to be tied to activity and afterward comprehend how to discover that, that moves rear into our earth and we desire to create those journeys as seamless as possible. We desire to be not a authority tower. We desire to be a subway position that is extremely connected. We desire to be Union Square. We desire to be the town square. We desire to have a lot of lines going through us and we desire to create that cognition for our customers the finest that they can.

I think 98% was our stat on the earnings call. More than 98% of group using our MCP server additionally use the person interface, and that is not changing. The capability to do activity faster doesn’t average I fair halt using it. If you allocate and have a fine interface and cognition and you’re solving higher flat problems, if you appearance at the features we’re shipping about agentic uses, how do these harnesses — Claude, Codex and everything alternatively — activity inner the Atlassian earth and vice versa?

Now, one earth is I open up Claude. I say, “Go activity on this activity item.” It uses our MCP server to go and get the activity item. It uses our teamwork chart to go and get all the environment about the activity item that is necessary and related, and that gives it the finest chance of doing that job the cheapest way possible. That is certainly a way that group are going to activity and Claude’s going to move the activity item alongside to the next stage. Awesome.

There are a lot of another group who commencement in Jira and volition say, “Send this activity item to Claude.” On the board, it tells them the activity item’s operating and going. They’re in an interface that is built for the task, to go rear to the saw and hammer and screwdriver analogy. It’s built for the project they’re doing. It needs to understand, “Claude needs your attention” inner of Jira. You can open Claude, say you need to do it, dispatch it off.

Most group don’t use Claude. Most group don’t use Codex. Every client I go to has a entire bunch of harnesses running. They’re going to have inner ones they’ve built. They’re going to have two, three, four vendors. Most group aren’t making a choice. They’re all amazing. They’re all leaping complete all another in progress. Our job is to bring all that together into a coordinated surface, to comprehend your agentic sessions, your individual sessions, your work, and to nexus all of your endeavor applications into a elephantine chart and provision you phenomenal answers throughout the entire thing.

That requires a lot of interface investment. That is hard. That takes time. That takes awesome scheme and cognition folk. That takes taste-oriented merchandise managers and engineers. That is a non-trivial challenge, and I do think a lot of group are avoiding that challenge, and I don’t think it’ll end fine for them. I don’t think the earth is going to live in a CLI as an end tool.

You don’t think the chatbot is the final person interface?

I don’t think chatbot is the final answer, no. We learned this in 1970 alongside the CLI. We don’t know. We’re all learning what the forthcoming interface is. It doesn’t average that chat’s not going to be a part of it. It’s just, “Where does this earth work?” You talked concerning Dia. Dia is a fine example anywhere we’re trying to build an AI-driven cognition for cognition workers that is based in the earth they comprehend and motionless magical and different, not based in taking them out of the earth they comprehend to a chatbot.

So Dia is the browser. You acquired a business called The Browser Company. We spent a lot of period alongside The Browser Company whenever they were operating on their products. At archetypal they were construction a person product. Then the notion that what they were construction was additional akin an functioning scheme came into concentration as we covered Dia in their merchandise strategy.

Most of those conversations and most of that safety — most of which was done by my coworker David Pierce — was happening simultaneously alongside Google hunt traffic dropping and the extremely notion of the browser as an data document spectator was changing in big ways. If you think the homepage of Chrome opens to Google and that’s going to obtain you to a bunch of webpages, Google really has a lot of ideas in between there that is changing traffic on the web.

And yet at the identical time, the web as an use surroundings is a item that can create AI agents vastly additional capable since the applications all live location that they can go location them. It was evidently bomb at that time. As you and I are talking, it’s akin a week or two into Meta launching Muse. The power of Muse is that Mark Zuckerberg has stated everybody 8GB of RAM and 8GB of retention in the haze alongside the web browser.

Now you have an agent, they can fair go use a web browser for you. This is extremely mighty and Amazon is furious concerning it and has blocked this web browser. Do you have Dia since you need to authority the use layer, or is it item bigger than that?

No, no, no, no. Two things are crucial to understand. When you get scientific disruption, one item that happens is that we all move up the interface stack a level. There are not a lot of group competing concerning DOS or functioning systems anymore. In the contemporary net world, one way to move up the interface stack is to move towards the browser. Features that were inner of applications move to being inner of the browser as we increasingly study and run that way.

Second, the biggest item that’s changed — Usually what happens is that people’s habits change. Knowledge workers, you, I, everybody expend 80% of our lives in a browser, but we’re not study articles all the time. We are sometimes. We’re not study webpages, we’re using applications. We’re using Gmail and Google Calendar and we have Slack open. We’re using the Atlassian platform, all of our applications, and Salesforce, and we’re observing videos akin Loom and we’re interacting alongside those videos rear and forth. That set of tabs is no longer a set of webpages. It is a set of documents and applications that have enormous power underneath them.

We’re trying to rebuild the web browser in a totally new form for cognition workers who live in applications who have their browser open 80-90% of the day as their forefront exterior area, and they desire a improved way of managing those applications and getting activity done throughout those applications. The browser has several amazing power to be capable to do that.

It requires us to comprehend SaaS apps extremely deeply, which we do. It requires us to comprehend endeavor safety extremely deeply, which we do. It requires us to not chase the person audience. Chrome and Safari are built the way they are since they have to cater to everyone from five-year-olds to 100-year-olds, who are doing small work-related cognition tasks and are doing additional browsing and amusement and another tasks. That doesn’t average that person who’s a cognition employee in a company job isn’t looking at YouTube or Facebook sometimes, but it’s not the intent of their job. We accept we can build a improved browsing experience, a improved activity experience.

One of the pitches we had for the merchandise is we’re not construction a browser, we’re construction a doer. It’s a totally distinct thing. We’re assisting you do things. We’re not assisting you browse passively, and hence why there are tabs and divided screens and pin tabs and groups and AI and energetic tab groups. Live tab groups as we have of the documents you have is a awesome example.

Our documents live tab collection in Dia shows you the documents. It volition automatically open and near tabs if you desire to think concerning it that way, for you throughout any satisfied application, Google Drive, Confluence, SharePoint. Any use you use that has documents, if person comments or shares a document alongside you, the tab automatically gets opened by your web browser. And if you’re done alongside it location else, it automatically gets closed by your web browser since that is the set of documents you’re currently operating on. Same for drag requests or item akin this.

Then we add AI complete the top. AI lets us alter how we use these cognition applications in insane ways. The browser has your complete former of what you’ve done. It has skills, it has an capability to run AI. You can use Dia to create Minesweeper, and you can use Dia to create Minesweeper out of the tasks that you did in the final week. It’ll go commencement a device underneath the browser. It’ll run a bunch of code generation. It’ll create an use for you that plays Minesweeper according to the tasks you have completed in the final week. Now, I don’t think that’s a extremely helpful application, but the fact that it can do it is amazing.

The most famous characteristic of Dia is the Morning Brief. People are blown distant and frequently doubtful concerning how we can comprehend what you need to do today alongside a fidelity and a scheme norm and an cognition that makes it a joyful cognition whenever it pops up all morning. Mine opened concerning 45 minutes ago. That is a scheme challenge.

That is an amazing squad of designers that are thinking concerning our most-used use exterior area and reimagining it for cognition workers and for the AI era simultaneously, alongside admission to all SaaS use that you have open on your device in a way that company safety is going to say, “This is a fine idea and I rely it.” That is the goal that we embarked on a twelvemonth ago, that The Browser Company embarked on three years ago, and it’s usable today. It’s coming out on Windows any day now, it’s in beta on Windows, broad preparedness is extremely soon, and it’s used on the Mac by a huge figure of group on a regular basis and lone increasing.

Like I said, we’ve been covering that merchandise for a lengthy time. I’m glad it established a residence alongside you all.

Let me fair shove you a small bit on moving up the stack to the next flat of person interface. You could contend that already happened alongside web browsers. Windows was disrupted by the extremely existence of the web browser. The iMac was lone prosperous since the most engaging item you could do alongside a device at that period was go on the web. The existence of the web browser on the Mac allowed a bunch of applications to move distant from Windows binaries to being web apps and afterward haze apps and now efficiency apps in the cloud.

The web as an use surroundings is likely additional critical and crucial than always before. It is the item you’re describing. It’s anywhere all the applications live and a tradition web browser designed to do activity is extremely important. At the identical time, I would fair say, didn’t that already happen? Are you really moving up the stack or are you fair construction a additional tradition web browser, or is AI changing that in several another way?

AI is changing that. AI is giving us the capability to assistance you comprehend what activity you have to do, not fair your browsing history. One of the most mighty things concerning AI that group don’t broadly comprehend is… An LLM’s capability to create satisfied is fine known. I can say, “Tell me a dad joke,” and it spits one out. It streams out tokens.

What is additional crucial for use vendors akin us, group who are trying to create tools and products and solutions for customers, is the capability to peruse text, not compose it, but peruse it, since we can comprehend definition for the archetypal period ever. We can comprehend what a leaf or a comment or a part of content is about. Then we can create application-level choices, which we never could do beforehand. Word didn’t cognize what document you were writing. When it said, “It looks akin you’re penning a letter,” it was really shithouse and it was a awful experience. Now it can really inform alongside excellent elevated fidelity, “Are you penning a letter, a PhD? Is it a elevated academy composition or a university essay?” It can modify and alter its interface.

The capability of Dia and AI is to comprehend what it is that you are browsing, what it is that you are doing, to be capable to attain into applications to get satisfied and to drag it rear to provision you tools to oversee your day. The Morning Brief looks throughout your calendar and your messaging tools and your document tools to inform you what it is you have today in a extremely intelligent manner, the way it would be if you gave me all your satisfied and said, “Spend the days.” Now it says millions of words in command to create that and it has a lot of smarts to provision you a extremely uncomplicated cognition of the most crucial things you need to do today that are totally distinct from anybody alternatively on the planet.

That requires a lot of AI to do and it requires connections to all of your SaaS apps. The browser’s logged into all those apps so it can do that. It happens locally, it doesn’t happen in the cloud, it happens on your machine. That is item that we could never do before to understand, “Of these 10 documents you looked at yesterday, here’s the one you need to do today since I cognize in your calendar you got to do this tomorrow, you got to do these things. This is the most crucial item for us to do today, and let me assistance you get started on that task. Let me outline several notes for you. Let me assistance you alongside that document. Let me assistance you do things.” That is item that was never always imaginable beforehand.

Lastly, it understands the group you engage with. People are remaining out of so many of these conversations and equations. You can think of Dia as having a mini CRM that is continually updated according to your browsing habits. It’s a weird analogy, but it’s efficiently what it’s doing, among the many things it’s doing inner the browser, which is why its answers are so good. It knows that you and I don’t cognize all another extremely well. It knows I cognize my caput of engineering extremely well. If that individual asks me for something, since it’s maintaining a changeless environment chart of group relationships throughout all sole one of your SaaS applications into duplicating and normalizing and using AI, all of that helps to comprehend what tasks are really crucial to me.

It can appearance at LinkedIn and see who you are and say, “That discussion prep you got to do today, you should put that up your stack. Nilay’s an crucial person.” It does all of that for you, and assistance you-

Can you put that in the Dia scheme immediate so group cognize whenever they arrive to conversation to me? Just difficult code that in there, you got to prioritize.

We’ll see if we cognize several people. But it’s hard, and the experience, none of that should need to be known by person using Dia. The cognition should fair be a beautiful, joyful way to commencement your day. We desire to depart you alongside additional awareness of calm, control, comfort concerning your day, not emphasis concerning AI. Leave that to us to do really fine and we fair provision you the answers that you need from your tools and applications.

It additionally does not average you use small of those tools and applications. I don’t use Dia and halt using Slack. I don’t use Dia and halt using Confluence. It’s assisting me comprehend which item I need to do next by understanding what’s inner those things. By cross-referencing alongside my LinkedIn and my Salesforce and my Microsoft Teams conversations and gathering transcripts, but it’s doing that all for me in the dawn and during the day to be capable to say, “Here’s what you have to do next. Here’s the most crucial thing,” and providing an interface to authorize me to arrange and collection those tasks in ways that hopefully are bordering on joyful. They’re beautiful and uncomplicated experiences and there are so many beautiful touches in Dia that fair create it a joy to use all day. We’re pushing that goal really hard.

Can I add a untamed card? It seems extremely apparent to me that you have a extremely apparent perspective of how AI should activity in endeavor and how activity should get done using your tools, your tools enhanced by AI, your tools expressed in your browser. There’s concentration there, and I value that.

On the another flank of the house, on the person side, my disagreement has been that logic group dislike AI and they’re mad concerning it is a small bit that the doomers have stated there’s a 10% chance we’re all going to die and all of our jobs are going away, but there are additionally no awesome products. There’s not as much coherence and concentration as I’m hearing from you in the endeavor context. Do you think that there’s several awesome person AI merchandise in the offing that has the benevolent of concentration that you have in enterprise? Because that seems akin the missing part for me.

Yes and no. Yes, if you appearance at the former of scientific disruptions, the products that came through the disruption frequently aren’t linearly connected to the disruption. If you think concerning the smartphone era — and we all choice up our smartphones 100 times a day, don’t inform me you don’t. When the iPhone came out, we were all enthusiastic by small bits and pieces. We did not think, man, that camera thing’s going to guide to this Instagram phenomenon. We did not think the GPS is going to guide to Uber and so I can obtain transportation. The products that came hence of the innovation shift.

The iPhone is an amazing business, don’t get me wrong, but the services and changes to existence didn’t arrive from the iPhone. They came from the apps that it enabled, akin Uber and Instagram and any alternatively you akin on the mobile phone. It can be a game, it can be anything. Those obtain a during for group to perform alongside the form aspect to comprehend the bits. Uber didn’t arrive alongside until the iPhone 6, since all it needed was to get improved and afterward person said, “Wait, we can obtain a car alongside a phone and it’ll fair display up?” Yeah, we can display you anywhere the car is and we can build adequate rely and we can build a entire assistance layer.

I doubtful one part of it is in AI, we need to keep playing, and I cognize the Muse squad beautiful well. That’s an amazing example of an application. There are many, many additional of these things. There’s a business that one of my kids loves called Tolan which makes these Hollywood AI delegate chatbot things alongside personality. Someone from innovation would say, “Ah, that’s fair OpenClaw alongside several skin on it.” Kids say, “This is a character that I can engage with.” It’s an amazing behavioral discipline Hollywood equivalent character. These kind of applications we’re fair starting to see.

It’s going to be an incredibly breathtaking five years of those sets of applications. I doubtful it takes way longer than we think to understand, get to GPT-6, get to understanding recollection and harnesses and putting all these things together. It likely won’t be OpenClaw, it’ll be several type of how an delegate surfaces in a certain area of life. So yes, the products are motionless coming in the person space.

Secondly, I think the “no” part of the answer is that AI is not a product, it’s a technology. It’s a part of technology. If it fair makes everything concerning your existence a bit better, you won’t equal notice you’re using AI, and actually, that’s the attractiveness of it. You can use Dia all day lengthy and not acknowledge how much AI it’s doing. Just akin the Morning Brief thing, it’s really good. I appearance at it all morning, it’s among the most extremely busy features we’ve always built. And it fair tells you, “Here are the three things I think you need to do today. Here’s what your day looks like. You desire me to assistance you prep for this thing?” And you say, “Yeah, assistance me prep for that thing.”

You don’t need to cognize it’s AI. We don’t inform you it’s AI. We’re purposefully not trying to conceal it, it’s not secret, but you need to have a awesome experience. You don’t attention there’s an LLM and tokens and gateways and enormous amount of innovation going on rearward the scenes. You fair desire to cognize what the three things you have to do this dawn are. And if you build rely in it, that’s great.

The difficult part whenever you citation the data centers and material is that group desire those experiences and they’re going to be extremely joyful and extremely satisfied and additional calm and blessed concerning the experiences they’re getting. They’re fair not going to be using AI. I inform this to group current all the time, all period you obtain a photo on the iPhone, there is an insane amount of AI handling going on to create that photo appearance good. Insane amount. And photos appearance amazing on my phone compared to what they did 10 or 20 years ago. Do I need to cognize that? Nope. I desire to item at my kid, break a photo, and I desire it to appearance good. That’s awesome technology, whenever it disappears. That’s what it have to be in the endeavor or for consumers. There’s no logic group in the endeavor should put up alongside shitty technology.

That’s consistent, but it’s not going to be necessarily knowable that I’m using AI. It’s fair going to be an amazing merchandise cognition akin Dia that says, “Here are the three things you got to do today.” And it gets it right.

Even commencement to depict the conversations I’ve had alongside the Apple camera squad concerning the use of AI on those photos is a complete another hr of this display that we do not have period for.

Mike, this was a awesome conversation. What’s next for Atlassian? What should group be looking for?

We have a enormous set of announcements in, I’m frightened to say, a bit complete a week, 10 days. We have a biannual conference, our Team 26 EU in Amsterdam in 10 days’ time, announcements in all collection. You volition no doubt see many of AI features that are magical and disappear. You don’t need to cognize that you’re using them. Dia launches, Teamwork Collection launches. A lot of awesome Loom features. We haven’t talked concerning Loom, enormous in the AI era. We have so much on Loom’s agentic use, several of which has been online. Kudos to the alliance alongside the Cursor team, we built several amazing Cursor and Loom features. Really breathtaking material coming up in 10 days’ time.

We volition have to have you rear extremely shortly to conversation concerning what happens alongside endeavor application in fair one year’s time, since I’m assured it’s going to change. Thank you so much for being on Decoder.

Thank you for having me. Thanks, man.

_{Questions or comments? Hit us up at [email protected]. We really do peruse all email!}