Keyword investigation is an important shape successful creating contented that meets your audience's needs and provides worth for your ain business. But basal keyword investigation often gives you the aforesaid database of keywords that everyone successful your class is getting erstwhile they tally their ain process.
Performing keyword investigation pinch Claude and Semrush solves this rumor by building your business discourse into the keyword investigation process itself. And by looking astatine what existent customers are saying online and successful your income calls.
The keyword investigation workflow I'll show you beneath is repeatable, and you tin usage it immoderate clip you request to create caller contented aliases update existing pages. Here's really to do it:
Part 1: Set up your Claude project
Set up a Claude task for your keyword investigation to support it organized and supply project-specific context.
Projects fto you adhd files and instructions Claude will reference for each chat successful the project. This leads to amended outputs and little backmost and distant pinch Claude, making the full process much efficient.
Create the project
To create a task successful Claude, click the “Projects” action successful the left-hand sidebar. Then click “New project.”
Give your task a name, for illustration “Keyword Research Workflow,” and a description. Then click the “Create project” button.
Add task instructions
Add task instructions to outline what you want Claude to do successful your recently created task by clicking the “+” fastener adjacent to “Instructions” wrong your project.
In the “Set task instructions” window, you’ll specify the unsmooth process Claude will follow, the information sources it should use, and your expected inputs and outputs.
Here’s an illustration group of instructions you tin usage for this process:
For chats successful this project, execute keyword investigation utilizing the Semrush MCP, uploaded Google Search Console query CSVs, a database of existing page URLs uploaded by the user, and immoderate uploaded competitor reviews, Reddit threads, aliases income telephone transcripts, mixed pinch the business discourse record added to this project.
Run the pursuing processes erstwhile prompted:
- Keyword spread study (Semrush MCP)
- Easy triumph study from uploaded Google Search Console information (plus Semrush MCP if prompted by the user)
- Competitor reappraisal analysis
- Reddit thread analysis
- Sales telephone transcript analysis
For the competitor review, Reddit thread, and income telephone analyses, make keyword ideas first, past validate them successful Semrush for measurement and keyword difficulty. Output results arsenic CSV files unless told otherwise.
When a personification asks for a circumstantial number of keywords to return utilizing the Semrush MCP, return only that number. Do not fetch much than 10 rows if they only inquire for 10 rows of data, arsenic this tin impact API portion usage. If you tally into issues, aliases don’t spot information wherever you expect to, emblem it to the personification alternatively of moving your ain diagnostic aliases exploratory queries.
Once you’ve added your instructions, click “Save instructions.”
Give Claude your business context
Give Claude your business discourse to amended the output value and keywords that Claude identifies arsenic applicable to your business.
To adhd business context, click the “+” fastener successful the “Context” model and past click “Add matter content.”
Add a title for illustration "Business Context" and past paste successful contented that answers the pursuing 5 questions:
- What do you sell?
- Who buys it?
- What problems do you solve?
- Who are your competitors?
- Do you person immoderate circumstantial rules you want Claude to follow?
By adding this context, Claude tin make decisions while identifying keywords.
For example, it tin tally related searches for merchandise features that you specify, aliases it tin adhd modifiers to its hunt position based connected your target market.
Adding business discourse now will besides thief successful portion 3 erstwhile we’ll inquire Claude to analyse reviews, Reddit threads, and income telephone transcripts to travel up pinch its ain keyword ideas.
Use our business discourse template to create a archive that you tin upload for this process:
I besides urge uploading a database of your website's URLs for further discourse to thief Claude spot what pages you already person connected your site. While it's not perfect, since it relies connected interpreting your URLs, it’s a reasonably speedy and easy measurement to springiness Claude discourse connected pages you already have. This allows Claude to propose pages you whitethorn want to optimize alternatively than creating caller pages.
Do this by adding the database the aforesaid measurement you added the business context. Click the “Add Content” fastener erstwhile you’re finished:
Connect Semrush and Search Console
Connect Semrush via the MCP and upload Google Search Console (GSC) information to supply Claude pinch sources of truth for each keyword investigation process you’ll transportation out.
To link the Semrush MCP, click the “+” fastener successful immoderate chat, past spell to “Connectors” > “Add connector” > “Browse connectors.”
Then hunt for Semrush and click the “+” fastener adjacent to “Semrush.”
There’s nary charismatic MCP disposable for Google Search Console, but location are still ways you tin link Google Search Console straight to Claude. We covered really to do this successful our article astir turning Claude codification into your ain SEO analyst.
For this workflow though, each you request to do is upload a CSV record of exported Google Search Console query data. You tin get this by navigating to “Performance” > “Search results” wrong Google Search Console and clicking the “Export” button.
This will download a zip file. Open it, prime the “Queries.csv” file, and upload that to your Claude chat.
Download a caller CSV utilizing this method each clip you execute the analyses beneath (e.g., monthly), truthful Claude has up-to-date data.
Part 2: Find keyword opportunities
In this part, we'll show you really to uncover keyword opportunities based connected your competitors and find easy triumph opportunities.
By default, Claude will output each the ideas for each of the methods beneath arsenic individual CSV files. I urge performing each of them wrong the aforesaid chat, truthful Claude tin easy harvester and benignant these CSV files without you having to stock files betwixt chats.
Find keywords competitors person that you don't
Find keywords your competitors look for that you mightiness beryllium missing by performing a keyword spread analysis.
If 1 competitor ranks for position you don’t and they person a akin target assemblage to you, past these position whitethorn beryllium worthy targeting yourself.
If aggregate competitors rank for a word that you don't, that's a beardown awesome the taxable is applicable for your target assemblage and 1 that you whitethorn want to cover.
Find your keyword gaps utilizing the punctual below:
Perform a keyword spread study betwixt the competitors listed successful the business discourse record and my domain, 1 astatine a time. Apply the pursuing filters:
- Minimum hunt volume: [e.g., 100]
- Keyword trouble range: [e.g., 0-49]
- Search intent(s): [e.g., commercialized and transactional]
First database the apical 10 keywords for each competitor individually. Then find up to 10 keywords that each competitors rank for but my domain doesn't (if location are none, conscionable emblem that and don't execute immoderate further analyses).
Prioritize keywords applicable to my business per the business discourse file. Exclude branded keywords and navigational intent (e.g., “[competitor name] login”). Combine each the results into a CSV.
Note that you tin inquire Claude to find much than 10 keywords astatine a time, but this will summation API portion use.
Here’s an illustration output from this process:
Because Claude has discourse astir your business, it tin execute a keyword spread study pinch business relevance filtering built in. That prevents it from surfacing keywords that whitethorn not beryllium applicable to your business.
Find the easy wins successful Search Console
You'll now find queries you rank for but not good (e.g., successful positions 8 to 20), aliases that you get fewer clicks for based connected your Google Search Console data.
Because Google is already ranking you location for these terms, your contented is already deemed relevant. That’s why they could beryllium easy wins.
Further optimizing your contented for applicable keywords you’re already ranking for improves your chances of ranking amended pinch little effort than trying to create wholly caller pieces from scratch.
Use this punctual to place easy triumph opportunities (make judge you’ve uploaded the Google Search Console queries CSV from portion 1):
Analyze my uploaded Google Search Console query CSV and place 3 types of keyword opportunities:
- Keywords ranking successful positions 8 to 20
- Keywords pinch precocious impressions but debased CTR
- Keywords pinch impressions but nary clear page covering the topic
Use my uploaded URL database to spot sum gaps. Output a CSV pinch columns for the keyword, the type of opportunity, why it’s worthy prioritizing, and whether it should beryllium a caller page aliases optimization of an existing page (with the corresponding URL).
Claude will output the CSV, but it whitethorn besides summarize the astir important opportunities. Like this:
Since this measurement is based connected your ain Google Search Console data, you don't request to validate these keywords pinch Semrush to cognize that there’s existent request for them.
Part 3: Find what customers really say
This portion involves uncovering what your imaginable customers and existing customers are saying and turning those comments into keyword ideas that you tin create contented around.
Mine competitor reviews
Mine your competitors’ reviews to place words and phrases that their customers usage to picture their products and services and the kinds of features customers value.
The connection you find successful competitors’ reviews helps you place opportunities that accepted keyword investigation mightiness not.
For example, if a customer leaves a reappraisal for your competitor and mentions features that they wish the competitor's merchandise had, you could create a portion of contented that highlights the truth that your merchandise aliases work does person those features. And that could lead to appearing successful applicable hunt results aliases AI responses.
For example, respective G2 reviews for task guidance instrumentality Trello statement that it doesn’t support task dependencies.
SmartSuite, different task guidance tool, has contented specifically astir limitations successful task management. One of its articles highlights really SmartSuite helps negociate dependencies.
This portion of contented is cited successful the AI Overview for related hunt terms:
Where you look for competitor reviews will dangle connected your merchandise aliases work and industry. If you’re successful B2B software, websites for illustration G2 and Capterra are bully places to start.
If you waste B2C products, past your competitors' Amazon reviews are worthy looking at. So are reviews connected their website aliases immoderate different sites wherever they waste their products.
To move your competitors' reviews into usable discourse wrong Claude, simply transcript reviews and paste them into Claude.
Then, adhd the punctual below:
Group the attached reviews into recurring themes, summarizing each one. For each theme, propulsion nonstop phrases aliases quotes showing really customers picture their problems, aliases features they for illustration aliases wish the merchandise had. Turn these into realistic hunt queries personification pinch that symptom constituent mightiness type into Google. Output arsenic a CSV pinch columns for keyword, theme, and root reappraisal summary.
Because this punctual is designed to nexus each suggested keyword thought to the reappraisal that it came from, you tin verify why Claude has suggested each keyword idea.
You'll validate whether these keywords are worthy targeting utilizing Semrush shortly.
Mine Reddit threads
Mine Reddit threads to place the connection that customers usage erstwhile describing products aliases services for illustration yours.
To find Reddit threads related to your offering aliases business, usage the Google tract hunt usability “site:reddit.com ‘keyword’” aliases browse applicable subreddits directly.
A speedy measurement to execute keyword investigation for these Reddit threads straight successful Claude is to usage a punctual for illustration the 1 beneath that simply searches Semrush for the keywords a fixed thread ranks for:
Use Semrush to find the apical [number] keywords that [thread URL] ranks for and see hunt volume, keyword difficulty, and intent for each one. Exclude immoderate keywords that incorporate “Reddit” aliases communal misspellings, arsenic they person navigational intent.
This useful because, if the thread ranks for a fixed term, the word is apt applicable to what your target assemblage was talking astir aliases asking astir successful the Reddit thread.
However, the method of seeing what position a thread ranks for requires the thread to rank good successful Google, truthful you tin miss retired connected a batch of imaginable keyword ideas. That's why I besides urge leaning connected Claude to execute keyword study successful a akin measurement to really we did it for the competitor reviews above.
To do this, transcript and paste full Reddit threads on pinch comments into Claude and tally the punctual beneath to get Claude to analyse the threads for imaginable keyword ideas:
Using the attached Reddit thread and comments, and the business discourse record for reference, place words and phrases users mightiness type into Google aliases AI hunt devices related to my products aliases services. Focus connected symptom points raised successful the thread. Output arsenic a CSV pinch 2 columns: 1 for the keyword and 1 for the thread aliases remark snippet it came from.
Further reading: Reddit keyword research: How to find hidden SEO opportunities
Pull topics from your income calls
Pull topics from your income calls utilizing the nonstop words and phrases prospects and customers usage that you whitethorn not presently person contented optimized for.
Simply upload your income telephone transcripts to Claude and usage the pursuing punctual to extract keyword ideas:
Using the attached income telephone transcript(s), place keyword ideas based connected symptom points, complaints, and features prospects praised aliases wished the merchandise had. Output arsenic a CSV pinch columns for keyword and the nonstop transcript snippet it's based on.
Validate hunt demand
Validate the keyword ideas Claude generated based connected competitor reviews, Reddit threads, and income calls pinch Semrush to get metrics for illustration hunt measurement and keyword trouble to thief pinch prioritization.
Use this punctual to get Claude to validate hunt request utilizing the Semrush MCP:
Combine the keyword ideas from the competitor reviews, Reddit threads, and income telephone transcript into 1 CSV pinch columns for keyword, hunt volume, intent, keyword difficulty, and source, including the snippet each keyword came from. Use the Semrush MCP to find volume, difficulty, and intent for each keyword. If nary measurement information exists, participate 0 for volume, N/A for keyword difficulty, and estimate intent based connected context.
You'll astir apt find that a batch of the keyword ideas suggested by Claude person zero hunt measurement aliases nary information returned by Semrush. This doesn't mean there’s nary hunt request for that term. It could conscionable mean that that nonstop wording presently shows nary hunt volume.
Because these keyword ideas came from things imaginable aliases existing customers said, they whitethorn still beryllium worthy targeting successful your content. In these cases, usage your judgement to determine whether it's worthy creating a portion of contented astir that word aliases optimizing an existing piece. Or travel up pinch your ain variations and cheque for hunt request connected those instead.
For example, Claude suggested the keyword thought of "best organization for caller freelancers," but Semrush returned zero hunt measurement for this term.
I quickly brainstormed that location whitethorn beryllium a keyword related to this: "best communities for freelancers." I ran this thought done Semrush, and recovered that it does get 20 monthly searches. This isn't a immense magnitude of hunt volume, but it's an illustration of wherever Claude's first keyword ideas tin make much of your ain ideas pinch reasonably small effort.
Part 4: Prioritize the correct terms
Choose the correct position to target by asking Claude to consolidate each of the keyword lists it identified successful each shape of the keyword investigation process, past inquire Claude to prioritize them based connected what matters astir to your business.
Get a prioritized database from Claude by entering a punctual for illustration the 1 below:
Combine the keywords from the individual keyword spread analyses, the keyword spread study vs. each competitors astatine once, Search Console easy-win analysis, competitor reappraisal analysis, Reddit thread analysis, and income telephone analysis, utilizing only validated keywords for the second three, into 1 CSV pinch columns for keyword, hunt volume, intent, keyword difficulty, taxable cluster, new/existing page, and reasoning.
Cluster keywords by topic, since respective whitethorn fresh 1 portion of content. For the new/existing page column, cheque my uploaded URL list, and if unsure, statement this successful the reasoning column. Prioritize chiefly by business relevance, past by hunt measurement and keyword difficulty. In the reasoning column, explicate why each keyword aliases cluster was prioritized.
Then create a abstracted CSV for keywords you couldn’t validate immoderate Semrush information for. Cluster these keywords too, and statement wherever each keyword thought came from.
Here’s an illustration output pinch the validated keywords:
How you take to prioritize the position that Claude has recovered depends connected your goals. Prioritization is an important portion of keyword research, but the main things you should return into relationship erstwhile prioritizing keywords to target are:
- How applicable the keyword is to what your business does
- The likelihood that targeting that keyword could lead to a conversion
Search volume, keyword difficulty, and search intent tin beryllium useful for gauging hunt request and for unsmooth prioritization. But conscionable because a word has a precocious hunt measurement doesn't mean it's apt to person galore users. And conscionable because a word has nary hunt measurement doesn’t mean it’s worthless.
That’s why we asked Claude to make a abstracted CSV for keywords it was incapable to find immoderate Semrush information for. These are position you’ll person to manually measure to determine whether they’re worthy creating aliases optimizing contented for.
I urge looking astatine the snippet aliases thread that led to that keyword thought from Claude and determining from that really useful a portion of contented connected that keyword would beryllium for your target audience.
For example, 1 of the keywords Claude suggested from the income telephone transcripts that had zero hunt measurement was "how to find freelance clients consistently." If my business sold coaching services, a committedness to thief you find much clients consistently would apt beryllium compelling. And I could target that keyword pinch a blog station aliases landing page pinch a CTA for my services.
Turn your keyword ideas into high-quality content
You should now person a CSV of taxable clusters you tin usage to create content.
Many of the keywords Claude identified will apt require caller pages. Find retired really to create optimized pages for these keywords successful our guideline to creating SEO content.
If you want to effort the workflows successful this article for yourself, motion up for a free proceedings of Semrush to get entree to the MCP.
