Search used to offer a list of doors. AI increasingly walks through those doors, reads the rooms and returns with an answer. That changes how companies get discovered, but it does not make good marketing obsolete. It makes vague marketing much easier to ignore.
The click is no longer guaranteed
A growing number of searches end with a composed answer. The person may never visit the ten pages that helped create it. For marketers, this feels unfair. We did the writing, the machine gets the conversation.
Complaining will not restore the old search page. A better response is to become the source that deserves to be used, cited and remembered. That requires substance, structure and a recognisable point of view.
More content can make the problem worse
Generic content was already weak. AI has made it almost free. If your article could be published by any competitor after changing the logo, it adds volume but very little discovery value.
The useful question is not ‘How often should we publish?’ It is ‘What can we explain better than most people in our category?’ That answer usually lives close to customers, product decisions, sales objections and work that went slightly wrong before it went right.
You do not win by sounding like the internet. You win by being the clearest useful source inside it.
Martin Palmet
Build an answer inventory
Make a list of the real questions people ask before they trust you. Not broad keywords such as ‘B2B marketing’. Use the awkward, specific questions: What should a first marketing hire own? Why is our demo traffic not converting? When is a rebrand a distraction? What does a fractional CMO actually do in month one?
Give each important question one strong home. Answer it directly near the top, then add context, examples, trade-offs and a next step. A useful page should help a reader even if they never buy from you.
- Collect questions from sales calls, support and customer interviews
- Group them by the decision the reader is trying to make
- Choose the questions where you have real experience
- Update the answer when your view changes
Make expertise easy to recognise
Machines and people both benefit from consistency. Use the same company name, service descriptions and expert names across your website, profiles and reputable third-party pages. Explain who wrote the piece and why their experience matters.
Add evidence that is difficult to imitate: a before-and-after example, a decision framework, a real constraint, a client quote with context or a clear opinion about what not to do. Specificity is a quiet form of credibility.
Check the answer, not only the ranking
Search your company, your people and the problems you solve in several AI tools. Look at the language used, the sources referenced and the competitors included. This is not a perfect measurement exercise. It is a useful mirror.
If the answer is wrong or empty, do not start by gaming the machine. Fix the underlying public evidence. Clear service pages, useful articles, consistent profiles, customer proof and sensible internal links still do the heavy lifting.
A practical 30-day move
Choose ten commercially useful questions. Publish or improve the best three answers. Add a named author, concrete proof and links to the next relevant question. Then show those pages to sales and customers. If humans find them vague, an answer engine will not rescue them.
The future of discovery is not a race to produce infinite text. It is a race to become unusually clear about something worth knowing.
One last thought
Keep what is useful.
AI may change where the answer appears. Your job remains pleasantly human: know something useful, explain it clearly and make it easy to trust.
