How to Decide What to Write Next (Before You Open the AI Tool)

The worst way to start writing is to open an AI tool and ask it what you should write about. It will give you a list. The list will be ten topics that sound reasonable and mean nothing to your business. You will pick one, generate a draft, and publish something that exists only because the tool suggested it.

Three months later you have twelve posts, no traffic pattern you can explain, and a vague feeling that content is not working. Content is working fine. Planning is what broke.

The tool is useful once you know what you are writing and why. Before that point, it is a distraction dressed up as productivity. This article is about everything that happens before you open it.

Start with what people already ask you

You have a list of topics and you do not know it. It lives in your inbox, your DMs, your client calls, and the questions people ask in the first fifteen minutes of a project.

Write them down. Not the cleaned up, professional sounding versions. The actual words. “Do I really need a website or is Instagram enough.” “How much should this cost.” “What’s the difference between the two options you mentioned.” “Can you just explain what SEO actually does.”

These are better than any keyword tool for three reasons. First, they are real. Someone with money or attention to spend actually wanted the answer. Second, they are phrased the way a real person talks, which is usually how they search. Third, you already know the answer, which means the draft will have something in it.

If you have been in business for a year, you probably have fifteen to twenty of these. If you are just starting, you have fewer, and that is fine. Start with five. You do not need more than five to begin.

Sort by what your business needs, not what sounds interesting

Not all topics do the same work. A post can do one of roughly four things for a one-person business:

Bring in new people. These answer the questions strangers type into search. They tend to be the broadest, most introductory pieces. “What does a brand strategist actually do.” They reach people who do not know you exist.

Build trust with people who already found you. These show how you think, what you have done, and what you believe. They are the posts someone reads after they land on your site and before they decide to email you. They tend to be more specific and more opinionated.

Handle objections before the sales conversation. These save you time. “Why I don’t do rush projects.” “What’s included and what’s not.” “How long this takes and why.” Every time you answer the same question in a call, that is a post you have not written yet.

Stay in front of people who are not ready. Newsletter content, usually. Not meant to rank. Meant to keep someone thinking about you until the timing is right.

Look at your list of questions and tag each one. If everything on your list is category one, you are writing to attract strangers but giving them nothing to read once they arrive. If everything is category three, you are writing for people who are already almost sold and not reaching anyone new.

You do not need an even split. You need to notice what is missing.

Pressure test each topic before you draft it

A topic that sounds good in a list can fall apart when you sit down to write. Before you commit, run each one through three questions.

Do I actually have something to say about this? Not “could I research this.” Do you have experience, an opinion, a result, a story, a detail that is yours? If the answer is no, the draft will be a summary of what everyone else has written, and there is already a summary of what everyone else has written.

Is the reader going to do something different after reading this? If the answer is “they will be more aware,” that is a warning. Awareness is not nothing, but it is hard to write a piece with structure and direction when the outcome is vague. A useful test: finish the sentence “after reading this, the reader will ___.” If you cannot finish it with a verb, the topic might not be ready.

Can I write this in one sitting? Not “can I generate a draft.” Can you, the person who knows the subject, sit down and explain this in 20 minutes of talking or typing? If the topic is so broad that you would not know where to start, it is either three topics pretending to be one, or it is a topic you know less about than you thought.

Kill topics at this stage without guilt. A list of seven topics you can actually write is better than a list of thirty you will stare at.

Where AI helps (and where it does not)

Once you have your list, the tool becomes useful for specific tasks.

Organizing what you already collected. If you dumped forty questions from client calls into a document, the tool can cluster them, spot overlaps, and suggest which ones are the same question asked different ways. This is sorting, not strategy, and the tool is fast at it.

Expanding a topic into an angle. You know the subject but you have not decided the angle. Give the tool your topic, your reader, and three things you believe about it. Ask for five possible angles, each stated as a single question. Most will be obvious. One might be the piece you actually want to write.

Checking whether you forgot something. Give the tool your final list and your reader description. Ask what questions a person in that position would have that you did not cover. Treat the answer as a prompt to think, not as a list to accept. Half of what it suggests, you will have already covered. The other half, you will either add or deliberately skip, and deliberate skipping is fine as long as it is deliberate.

What it cannot do. It cannot tell you what matters to your business. It cannot tell you which topic will bring the client you want. It cannot tell you what you actually believe. It cannot look at your calendar, see that you have a quiet month coming, and suggest the trust-building piece that would fill it. That is the thinking, and the thinking is yours.

Build a list you actually use

The failure mode here is not a bad list. It is a good list you never open again. Two things prevent that.

Keep it short and visible. A spreadsheet with forty topics sorted by category, priority, keyword, funnel stage, content type, and estimated word count is a spreadsheet you will never open. A list of eight to twelve topics in a document you already use daily is a list you will write from. Add a new topic when you hear a good question. Remove one when you finish it or when it stops seeming important. That is the whole system.

Decide the next one, not the next ten. Sequence matters less than people think. Pick the topic you can write right now, because you just had the conversation or because you just solved the problem. Freshness of material matters more than editorial calendar logic. The best time to write about a mistake is the week you made it, not the quarter it was scheduled.

A worked example

Here is someone who skipped this step:

She opened an AI tool and asked for 20 blog post ideas for a freelance web designer. She got a list. She picked “Top 5 Web Design Trends” because it was first. She generated a draft, edited it lightly, and published it. It said nothing she believed and attracted no one she wanted to work with.

Here is the same person after doing the work:

She looked at her last ten client emails and her last five discovery calls. She found that four people asked some version of “should I redesign my site or is it fine.” Two asked “how long will this take.” One asked “why is my current site so slow.” She also noticed she kept explaining, in calls, why she does not build sites on a certain platform, and clients always seemed relieved to hear it.

Her list:

1. How to tell if your website needs a redesign or just a fix (brings in new people)
2. Why I stopped building on that platform (builds trust, handles an objection)
3. How long a website project takes and what makes it take longer (handles objections)
4. What makes a website slow and what you can check yourself (brings in new people)
5. What I need from you before we start (handles objections, saves her time in onboarding)

Five topics. Each one tied to a question someone actually asked. Each one something she can write from experience without research. She will not run out of things to say for two months, and by then she will have heard ten more questions.

Mistakes that waste the most time

Keyword-first planning. Starting with a keyword tool and writing whatever has volume is a strategy for media companies with editorial teams. A one-person business needs to start with what it knows and check whether anyone is searching for it, not the other way around. A post you can write with authority on a topic twenty people a month search for will outperform a post you researched on a topic two thousand people search for, because the first one has something in it and the second one does not.

Planning more than you will publish. If you write twice a month, you need a list of six to eight topics. Not thirty. Planning is productive up to the point where you have enough to work with, and after that it is procrastination.

Treating the plan as fixed. A client will say something on Tuesday that is more interesting than anything on your list. Write that. The list is a tool, not a contract.

Confusing a content calendar with a content plan. A calendar tells you when. A plan tells you what and why. Most people build the calendar first and fill it with whatever fits, which is backwards. Know what you are writing and why it matters, then decide when.

Common questions

How many topics do I need before I start writing? Five to eight is plenty. You can add more as you go. The goal is to start writing, not to finish planning.

Should I use a keyword tool at all? Yes, after you have your list. Use it to check whether people are searching for the thing you already plan to write about, and to see how they phrase it. That can shape your title and your framing. It should not decide your topic.

What if I genuinely do not know what my audience wants to read? Ask them. Send a short email, post a question, or pay attention to the next five conversations you have. The answer is in the room, not in a tool.

How often should I revisit the list? Glance at it every time you finish a piece. Do a real review every two to three months, or whenever your business changes. Drop topics that no longer matter. Add the ones that came up since last time.

Does this work for a business that is brand new? It works with less material. You will have fewer client questions to draw from, so lean more on what you know from your own experience and the conversations happening in communities where your future clients spend time. The list will be shorter and that is fine. Start with three.