Which tool to choose
To generate the topics, a general-purpose AI assistant in its latest version works fine: it knows people's language well enough to guess the typical questions. If you also want what is searched for and discussed in real time, a tool connected to the web helps, and Grok is useful when you're interested in catching the hot topics and current trends on X in your field. For exact search volumes (how many people search for something) you need dedicated SEO tools: the AI alone gives you the topics, not the precise numbers.
How to do it
- Define the customer and their central problem in two lines.
- Ask the AI for the questions that person asks themselves, split by stage of the journey.
- For each question, have it tell you whether it's for an article, a fixed page, or just a short answer.
- Sort the topics by priority: first those close to the purchase, then those that attract people who don't know you yet.
The operative syntax:
I'm [your job] and my typical customer is [who they are]. List the concrete questions this person types into Google in the three stages: 1) when they have the problem but don't know I exist, 2) when they're comparing solutions, 3) when they're almost ready to buy. Write the questions the way they would write them, not in insider jargon. For each one, indicate whether it's better to answer with an article, a page on the site, or a short FAQ.
Feedback check: if the questions seem like stuff you already know, you're on the right track: they are exactly the answers the customer is looking for. If instead they seem abstract, push harder: "Make them more concrete and specific, the way a frustrated person searching for a solution right now would type them."
A concrete example
Giulia has a pet shop. Instead of writing the post she felt like writing ("the story of our business"), she asks the AI for dog owners' questions. Out comes a list: "how much should a three-month-old puppy eat," "why does my dog scratch all the time," "which food for an elderly dog with kidney problems." Each one is an article that answers a real doubt. Since she started writing on these topics, people arrive who were searching for exactly those answers and then buy in the shop. The content plan practically wrote itself.
When it does NOT work (and how to fix it)
If the suggested topics are too generic
"Gardening tips" isn't a topic, it's a category. Ask the AI to go down a level: "Turn every generic topic into three precise questions that a person would really type." The more specific the question, the less competition it has and the easier it is to answer better than anyone.
If you don't know which topics bring customers and which only bring visits
Not all traffic is worth the same. Ask the AI to mark each question with its distance from the purchase: someone searching "how much does it cost to redo a bathroom" is closer to buying than someone searching "ideas for decorating my home." Start with the first kind: they bring fewer visits but more customers.
If you have a hundred ideas and don't know where to start
A hundred topics with no order are paralyzing. Ask the AI to pick ten to do first, with reasons: "Given my goal of finding customers, which ten of these topics would I tackle first and why?". Better ten articles published than a hundred on a list.
A tip from someone who really uses it
The best questions are the ones customers always ask you out loud. Before querying the AI, write down the ten questions you hear most often in your work and give them to it as a starting point: they're gold, because they come from real people. The AI expands them and finds similar ones, but the seed has to be yours. Anyone who builds the content plan only on the AI risks a plausible calendar that's disconnected from real customers.
Frequently asked questions
How often should I publish?
Consistency counts, not frequency. One useful article a month, every month, beats five a week for two months and then silence. Build a plan you can actually keep up: the AI can prepare topics for six months for you, but writing them is up to you.
Can the AI also write the articles, not just the topics?
It can do the draft, but on the topics where you count as an expert you have to put in your concrete experience, otherwise you write the same articles as anyone else using the AI. Use it for the structure and the first draft, and add the real cases and details that only someone who does that job knows.
Doesn't chasing people's questions make me the same as everyone else?
The opposite: the question is the same for everyone, the answer isn't. A hundred sites answer "why does my dog scratch" by copying each other; you answer with the cases you've seen in your shop and the products that really work. You choose the question with method, you make the difference with the experience others don't have.