A detailed buyer profile can make a campaign feel ready before you have spoken to a customer. The profile names a frustration, describes a busy day, and suggests an appealing promise. A customer interview may produce something much less tidy: a half-remembered purchase, an unexpected workaround, and a reason for choosing that has little to do with the features you wanted to advertise. Both inputs can help. They answer different questions.
For a small business preparing a creative brief, the useful comparison is customer interviews versus AI buyer profiles. Interviews provide accounts of particular people and situations. AI-generated profiles propose a way to organize and explore a market. Your job is to keep the origin of each statement visible as you turn those inputs into campaign decisions.
I work with The Rosh Group, which sells CampaignPress.
Compare the kind of knowledge each method supplies
An interview can tell you what one person remembers doing, noticing, or worrying about. It can uncover details you did not think to put in a prompt. It cannot establish how common that experience is across a market. A participant may also reconstruct a decision after the fact. Treat the account as useful evidence with a specific scope, rather than a complete explanation of every buyer.
An AI profile is useful for proposing possibilities: which situations might matter, which questions deserve attention, and which alternatives a buyer might consider. Unless its statements are tied to real, relevant evidence that you can inspect, the profile remains a hypothesis. A vivid occupation, daily routine, or imagined quotation does not convert it into customer research.
The GOV.UK guide to learning user needs makes a useful methodological distinction: opinions that do not come from users should be treated as assumptions to investigate. That principle travels well into a marketing brief, even though the guide was written for public services.
Use an interview when the situation is unclear
Consider a fictional bookkeeping service for independent retailers. The owner believes buyers want a more attractive monthly report. Before building a campaign around that belief, the owner asks a recent customer to describe the last time their records caused a problem. The customer recalls being unable to answer a supplier’s question while standing at the register.
That account changes the next research question. Was the difficulty report design, access to information during work, or something else? It does not prove that every retailer needs the same thing. It does identify a situation worth investigating before the campaign commits to an argument about beautiful reports.
Ask for a concrete episode before asking for an opinion about your proposed solution. “Tell me about the last time you needed that information” leaves more room than “Would a clearer dashboard help?” Follow the sequence: what happened first, what the person tried, what interrupted them, and how the episode ended. Write down what remains uncertain.
Do not recruit only your most enthusiastic customers. Include relevant people whose experiences may differ, when you have permission to speak with them. Record the limits of your small set of conversations. An interview program is a learning activity, not a vote that turns three comments into a market-wide percentage.
Use an AI profile to prepare questions and alternatives
When you have a rough business description but no structured brief, an AI-generated profile can make your assumptions easier to inspect. Ask for several plausible situations, then mark which ones your existing knowledge supports. Separate suggestions from facts before showing the profile to a colleague. Otherwise, a polished document can acquire authority simply by being the first coherent version in the room.
For the fictional bookkeeping service, possible situations might include preparing for a supplier conversation, checking cash availability, or closing the month. These are starting points for investigation. They are not findings. A useful next step is to identify which situation would most change the campaign if your assumption were wrong.
The GOV.UK research-planning guide recommends converting unfounded assumptions into questions. Apply that discipline to the profile: replace “buyers hate complicated reports” with “When did a report last get in the way of a decision, if ever?” The second sentence gives a conversation a job without supplying its answer.
Keep drafting assistance separate from evidence
AI assistance is increasingly available where people already produce content. In its May 8, 2026 changelog, WordPress.com described expanding its opt-in AI Assistant to current paid plans, with support inside the editor and Media Library. That development makes assisted writing and design part of an everyday publishing workflow. It does not establish that generated descriptions accurately represent your customers.
A tool can help shape a brief while the business retains responsibility for the underlying assertions. Keep two columns in your working notes: the statement you are considering and the source that supports it. “A participant described checking records at the counter” belongs beside the dated interview note. “Most retailers want instant answers” needs broader evidence and should not quietly inherit the authority of that single conversation.
Remove identifying details before moving research notes into a tool. Often the useful input is an anonymized situation and a bounded observation, not a name, account record, or full transcript. Follow the permissions under which the information was collected.
Choose a sequence that matches the uncertainty
If you cannot explain the buyer’s situation without inventing it, start with conversations and observation. If you have reliable notes but struggle to organize them, use a profile as a synthesis aid and check each assertion against those notes. If you have neither evidence nor access to customers yet, draft a provisional brief that openly separates what you know from what you still need to learn.
- Write the decision the campaign needs to influence.
- Describe one relevant situation using the evidence you have.
- List the assumptions that would change your creative direction.
- Choose the next question that could challenge the most consequential assumption.
- Revise the brief after learning something, retaining the limits of the evidence.
For CampaignPress, the relevant product connection is the order of work: define the buyer, establish strategy, and then develop copy and imagery. A buyer profile is a decision aid that still deserves your judgment. Explore the buyer-first campaign workflow after you have identified the facts and assumptions your brief needs to preserve.
End with a brief someone can challenge
A useful brief makes disagreement specific. A colleague should be able to ask where a statement came from, name an alternative explanation, or point out a missing group. If the document answers those questions, it can improve as you learn. If it only presents a confident portrait, it may hide the uncertainty that matters most.
Customer interviews and AI profiles can work in sequence, but they should not be treated as interchangeable evidence. Use real accounts to learn about particular situations, use generated possibilities to expose questions, and keep the difference visible all the way into the campaign.
Once the buyer assumptions are explicit, use the pre-production proof guide to decide which claims the evidence can support. The buyer research practice collection provides ten separate worksheets and a print decision card for applying these distinctions to your next brief.