Keyword research used to mean staring at a spreadsheet from a pricey tool and guessing which words your customers actually type into Google. AI takes over the slow part. It drafts keyword ideas, sorts them by what the searcher wants, and explains them in plain English in minutes.

This guide walks a non-technical small-business owner through a full keyword research session in under an hour. You will finish with a short list of search terms worth writing about, plus a rough sense of how hard each one is to rank for.

One ground rule before we start. AI guesses well, but it does not see live Google data. So we pair a chatbot for ideas with one free tool that shows real search numbers. Use both, and you stop writing for words nobody searches.

What AI gets right about keyword research

A chatbot is a fast brainstorming partner. Give it your business and your city, and it will list dozens of phrases a real person might search. It is also good at grouping those phrases by intent, which means sorting browsers from buyers.

Here is where it falls down. AI does not know how many people search a term each month, and it tends to invent confident-sounding numbers if you ask. It also underrates how crowded a keyword already is. Treat its volume claims as fiction and check them against a real tool.

The three tools you need

You can do this whole session on free plans. Two tools cover ideas, one covers real numbers. None of them needs a credit card to start.

ChatGPT or Claude (free tier)

Either chatbot handles the brainstorming and the intent sorting. The free versions are plenty for keyword work. ChatGPT Plus and Claude Pro cost about $20 a month each, but you do not need them for this task.

Skip it if you already pay for a marketing suite like Semrush or HubSpot that has its own AI keyword feature. No reason to bounce between two tools when one does the job.

Google Keyword Planner (free)

This is the closest thing to source-of-truth search volume, straight from Google. It is free, but you have to create a Google Ads account to reach it. You can set up the account without ever running an ad.

Skip it if the signup friction stops you cold. Planner also shows volume in wide ranges (like 100 to 1,000) unless you spend on ads, which is fine for rough sorting but vague for fine calls.

Ahrefs Free Keyword Generator (free)

This free web tool gives you a clean list of related keywords with a difficulty score and a search-volume estimate, no account needed. It is the fastest way to sanity-check the AI's ideas.

Skip it if you want deep data on every term. The free version caps how many keywords it shows and nudges you toward the paid plan, which starts near $129 a month.

The step-by-step session, under an hour

Block out 45 minutes and work through these four steps in order. The timing below is a guide, not a stopwatch.

1. Describe your business to the AI (10 minutes)

Open your chatbot and give it the full picture. Tell it what you sell, who buys it, and where you operate. A plumber in Austin gets very different keywords than a plumber serving the whole state. The more specific you are, the less generic the output.

2. Ask for keywords grouped by intent (10 minutes)

Ask the AI to list 30 keyword ideas and group them into three buckets: people researching, people comparing, and people ready to buy. That last bucket is gold for a small business, because those searchers are close to spending money. Save the full list in a plain text file or a spreadsheet.

3. Check real search numbers (15 minutes)

Now take your best 10 to 15 phrases and run them through Keyword Planner or the Ahrefs free tool. Throw out any term with near-zero searches, no matter how clever it sounds. Flag the ones with steady volume and a difficulty score you can realistically beat.

4. Pick your shortlist (10 minutes)

Choose five to eight keywords to actually write about. Favor longer, specific phrases over short broad ones. "Emergency drain cleaning in north Austin" is easier to rank for and pulls better customers than the single word "plumber."

A quick example: a local dog groomer

Say you run a two-person grooming shop in Denver. You describe the business to the AI, ask for intent-grouped keywords, then check the numbers. After 40 minutes you might land on a shortlist like this:

  • dog grooming Denver (broad, high competition, worth one strong page)
  • mobile dog grooming Denver (buyer intent, lower competition)
  • how often should you groom a doodle (research intent, easy blog post)
  • cat-friendly groomer near me (niche, low volume but loyal customers)
  • dog nail trimming cost Denver (price intent, ready to book)

Notice the mix. Two terms catch ready-to-buy searchers, two catch people earlier in the process, and one is a small but loyal niche. That spread keeps your content from chasing only the hardest words.

Three mistakes to avoid

  • Trusting AI search volumes. The chatbot will make up numbers. Always verify with Keyword Planner or Ahrefs before you commit a term to your shortlist.
  • Chasing only the biggest keywords. The one-word terms are buried under national brands. Specific long-tail phrases are where a small business actually ranks.
  • Ignoring intent. A great post for a researcher is a wasted post if you needed a buyer. Match each keyword to where the searcher is in their decision.

The bottom line

AI cuts keyword research from a half-day chore down to a 45-minute habit. The trick is using it for what it is good at, which is ideas and sorting, while letting a free real-data tool handle the numbers. That split gives you a shortlist you can trust without paying for a heavy SEO platform.

Run this session once a quarter, or any time you launch a new service. You will always have a fresh list of terms your customers are actually typing, and a clear sense of which ones are worth your time.