The “blank page” problem
Most small business owners I talk to don’t struggle to write content. They struggle to decide what to write about. Sitting down to a blank document and trying to think of a topic is one of the slowest, most demoralising parts of marketing. AI tools can take that part off your plate completely, if you use them properly.
The mistake people make is asking AI for “blog post ideas about [topic]” and accepting whatever comes back. The output is generic, recycled and forgettable. To get genuinely useful ideas, you need to give the tool more context and treat its output as a starting point, not a finished list.
Step 1: Feed it real context
Before asking for ideas, give your AI tool of choice (ChatGPT, Claude, Perplexity, whichever) a proper brief. The more it knows, the better the suggestions.
A useful starting prompt:
“I run [business] in [location]. My customers are typically [description]. They usually come to us because [reason]. Their biggest frustrations are [list]. We want to be known for [positioning]. Our existing blog covers [list of recent topics].”
That single paragraph is the difference between bland output and ideas you actually want to write. We talked about why context matters in our piece on using ChatGPT for marketing: sorry, that one’s still being scheduled, but the principle is the same as the rest of our content advice.
Step 2: Ask for breadth, not quality
Now ask for volume. Forty ideas. Sixty ideas. A hundred. Don’t try to get five perfect ones; you’ll fall into “average” territory. The goal is breadth so you have something to filter from.
Useful angles to ask for:
- “Forty blog post ideas based on questions my customers commonly ask”
- “Twenty topics that compare two options my customers might be choosing between”
- “Fifteen ‘how to’ posts that solve specific problems for my audience”
- “Twenty ‘mistakes to avoid’ posts in my industry”
- “Ten myth-busting posts that challenge common assumptions”
- “Fifteen ‘behind the scenes’ or process-revealing topics”
You’ll get a lot of duplicates and clichés. That’s fine. The next step is filtering.
Step 3: Filter ruthlessly
This is where you take over from the AI. Read the list with two questions in mind:
- “Would my actual customers care about this?”
- “Could I write something genuinely useful and specific about it?”
If the answer to either is no, cross it out. You should end up cutting at least 80% of the suggestions, and that’s a good thing. The best content lives in the small pile of ideas that survive your filter.
The job of AI is to help you generate possibilities. The job of you is to apply judgement. Don’t outsource the judgement.
Step 4: Add your own angles
Now go back through your filtered list and rewrite each idea so it’s specific to your business. Compare:
- Generic: “How to choose the right interior designer”
- Specific: “How to choose an interior designer in West Sussex (and the questions to ask before you sign anything)”
The specific version will outperform the generic one because it’s more targeted, more useful and more obviously written by someone with first-hand experience. We covered this idea in our post on how keyword research transforms strategy.
Step 5: Cluster ideas into topics
Once you have a list of 15–30 strong ideas, group them by theme. You’ll notice patterns emerging: “five posts about pricing”, “four posts about process”, “three posts about local projects”, and so on.
This is where you go from a content list to a content strategy. Each cluster becomes a topic area you can return to over and over. Each post within a cluster can link to the others, building topical depth that both Google and AI search engines reward. We talked about this in our guide to content marketing for small businesses.
Step 6: Schedule and start
The biggest mistake at this stage is over-planning. You’ve got your ideas. Pick one. Write it this week. Pick another. Write it next week. Don’t spend three months building the perfect content calendar before you publish anything. When you do sit down to write, our content writing checklist keeps each post clear, useful and easy to find.
A simple monthly plan looks like this:
- Week 1: One in-depth, useful “pillar” post
- Week 2: One quick, practical post or guide
- Week 3: One opinion piece or behind-the-scenes story
- Week 4: One round-up, FAQ or update of an older post
That’s four posts a month, with variety, and it’s enough to start moving the needle.
Things to do (and not do) when using AI for ideas
Do
- Give your AI tool detailed context every session
- Ask for high volumes of ideas, not finished output
- Always filter through your own judgement
- Add specificity that only you can add
- Cluster ideas into topics, not isolated posts
Don’t
- Accept the first list it gives you
- Use AI to choose your strategy. Only to support it
- Publish AI-generated ideas without making them yours
- Forget to verify any facts or stats it suggests
- Use it as a substitute for actually talking to your customers
The thing AI can’t do
The very best content ideas come from real customer conversations. The questions clients ask you in meetings. The objections that come up on calls. The misunderstandings you correct over email. AI can help you build on those, but it can’t replace them. We touched on this in our post on “it’s about you, not us”: the best marketing always starts with a real understanding of your customers.
Pair that human understanding with the speed and breadth of AI tools, and you’ve got a content engine that punches well above its weight. If you’d like a hand setting up a content plan that uses both, get in touch.
Frequently asked questions
Because the prompt gives the tool nothing to work with. Asking for “blog post ideas about [topic]” returns recycled, forgettable output, since that is all the request supports. The fix is context: tell it about your business before you ask for anything.
Tell it what your business does and where, who your customers typically are, why they usually come to you, what frustrates them most, what you want to be known for, and what your blog already covers. That single paragraph is the difference between bland suggestions and ideas you actually want to write.
Ask for volume rather than quality. Forty, sixty, even a hundred. If you ask for five perfect ideas you get five average ones. The point is breadth, so you have enough raw material to filter down to the few worth writing.
Angles that map onto how customers actually think. Questions they commonly ask. Comparisons between two options they are choosing between. How-to posts that solve a specific problem. Mistakes to avoid in your industry. Myth-busting pieces. And behind the scenes posts that show how you work.
This is where you take over. AI is good at breadth, at getting you past a blank page and generating more options than you would alone. Deciding which ideas are worth writing, and bringing the experience and opinions only you have, is the part that makes the piece worth reading.
