17 tech executives reveal how AI agents are changing work and hiring

This week, AI Insider hosted a dinner, sponsored by Dash0, in San Francisco for tech executives. I started alongside a uncomplicated question.

"Who had an AI agent assistance them at activity today?"

Every hand went up.

Then I asked a follow-up.

"Who had additional than five AI agents assisting them at activity today?"

Again, all hand stayed in the air.

Around the array were Zoom CTO XD Huang; Figma VP of Engineering Josh Clemm; Coursera CTO Mustafa Furniturewala; ServiceNow administrator Ashraf Karim; Glean CTO T.R. "Vish" Vishwanath; Netlify CEO Matt Biilmann; PagerDuty engineering chief Manu Gurudatha; Yahoo Mail GM Kyle Miller; Lime CTO Jon Jenkins; Proofpoint AI safety chief Ravi Ithal; Temporal cofounder and CTO Maxim Fateev; Inworld CEO Kylan Gibbs; Xero engineering chief Diya Jolly; MinIO CTO Ugur Tigli; Tribe AI partner Pooja Brown; Basis cofounder Creston Brooks; and Empirik CEO Kartik Chandrayana.

No one debated whether AI volition slay us. These are informed tech leaders from the trenches wrestling alongside genuine questions, specified as: How do you create AI agents really helpful inner a company?

Context

One expression kept coming up all evening: context.

Zoom's Huang stated AI agents won't rotate into genuinely helpful until they comprehend far additional concerning your company, your colleagues, your meetings, and your work.

"Context is all you need," XD quipped, riffing on the celebrated AI investigation document "Attention is All You Need."

Glean's Vishwanath agreed. He stated AI agents frequently neglect since they're connected to too small data — or worse, the incorrect information. They confidently create plausible answers that aren't really correct. Giving the correct environment to AI systems at activity could now be additional crucial to construction smarter models.

That theme cropped up repeatedly.

Ashraf Karim, who leads ServiceNow's inner rollout of its own AI products before customers get them, stated the business archetypal used AI to assistance individual assistance personnel before gradually allowing AI agents to grip simpler client requests themselves. Humans using AI are now delivering equal higher client fulfillment scores than humans operating solitary since they determine problems faster and create additional personalized answers.

Biilmann, the CEO of Netlify, asserted that penning code is quickly becoming a commodity. The scarce skills are increasingly domain ability and application judgment. Companies, he said, may eventually reorganize engineering teams about those strengths fairly than about group merely penning code.

Hiring

Hiring sparked among the liveliest debates of the evening.

Several executives described what one attendee called a "barbell" strategy: hiring extremely informed engineers who can inquiry AI-generated activity alongside younger graduates who are naturally fluent alongside AI tools.

Clemm, Figma's engineering VP, warned that engineers in the center of their careers could visage the most force as organizations flatten and managers increasingly favor elder hires.

Inworld's Gibbs stated startups don't have that luxury. His business is leaning equal harder into informed engineers who can resolve the difficult specialized problems AI motionless can't. If AI can already do item in a day, he argued, it likely isn't anywhere your rivalrous advantage lies.

Optional software

At one point, Tigli, the CTO of MinIO, showed me item that captured another theme of the evening: How AI is changing the application endeavor and making several application unnecessary.

Earlier that day, he'd asked Claude Code to create a presentation. Not in PowerPoint. Not in Google Slides. Claude generated a sole HTML document instead. He opened it on his laptop, afterward AirDropped the identical document to his iPhone. It worked absolutely on the two devices. No display software. No haze service. Just a file.

This is frequently how application gets disrupted. Not alongside a theatrical merchandise launch, but alongside group quietly discovering a simpler way to get the identical job done. If AI can build a display from a immediate and bundle it into a portable HTML file, how much do you really need Microsoft PowerPoint or Google Slides?

Predictions

I ended the meal alongside one final question: "What volition we chuckle concerning doing manually two years from now?"

Several group predicted position meetings would mostly disappear since AI agents would already cognize what everyone had accomplished. That echoes what OpenAI is already seeing internally: several investigation teams have stopped holding troubleshooting agency hours since researchers increasingly depend on AI instead.

Then Lime CTO Jon Jenkins produced among the night's biggest eyebrow-raisers.

He suggested application engineers won't be the lone group wondering whether AI can do their jobs.

Within two years, perchance C-suite executives (like those about the table) volition be asking themselves the identical uncomfortable question.

Sign up for the AI Insider newsletter here. Reach out to me via email at [email protected]