Marketing in the Age of the Answer Engine

For twenty-five years, marketing has been organised around a single behavioural assumption: when people want something, they search, they scan a list of results, and they click. Entire industries — SEO, paid search, conversion optimisation — exist to win that click.

The assumption is now breaking. A growing share of buying research happens inside AI assistants that read the web so your customer doesn't have to. The customer asks a question; the machine gives an answer. No list, no scanning, and very often, no click.

We've spent the past year watching this shift play out across our clients' analytics — and helping them respond. Here's what we've learnt.

What actually changes

The panic version of this story says "search traffic is dying". The truth is more specific and more useful: the top of the funnel is being re-intermediated. Three consequences matter.

Comparison happens without you. When a buyer asks an assistant to compare five vendors, the comparison is assembled from whatever the machine can find and verify. Your carefully staged comparison page is one input among many — alongside review sites, forum threads, documentation and journalists. You no longer control the shortlist narrative; you can only feed it.

Citation replaces ranking. Assistants prefer sources they can quote cleanly: specific claims, real numbers, named authors, clear structure. Vague thought-leadership soup — the kind written to rank rather than to inform — gets read and discarded. The irony is delicious: after years of content designed for algorithms, the machines now reward content designed for humans.

Brand becomes the query. When the middle of the funnel disappears into a chat window, the battle moves to the two ends: being the name the customer types into the question ("is Prisma any good for brand strategy?"), and being the answer that comes back. Unprompted awareness — the most old-fashioned metric in marketing — is suddenly the most modern.

The new playbook

None of this requires abandoning what works. It requires re-weighting. Four moves we're making with clients now:

1. Publish quotable facts. Original research, benchmark data, named case results with numbers. Machines cite specifics, and so do journalists and analysts. One genuine data point outperforms fifty opinion posts.
2. Win the places machines trust. Reviews, community discussions, trade press, technical documentation. These third-party surfaces now carry more weight than your own blog, because assistants treat independence as a signal of truth. This is PR's revenge decade.
3. Write for extraction. Clear claims, clean structure, answers stated before they're justified. If a well-read human can skim your page and repeat your argument accurately, so can a machine. If they can't, neither can it.
4. Invest in memory, not just discovery. Podcasts, events, communities, distinctive campaigns — the channels that put your name in a buyer's head before they ever ask a question. The best answer-engine strategy is being the brand the question is about.

If you want the mechanics rather than the argument, we have written practical guides to answer engine optimisation and generative engine optimisation. Much of it comes back to earned visibility — the discipline behind our PR and visibility work.

The uncomfortable conclusion

The answer-engine era punishes exactly the marketing that the search era let brands get away with: derivative content, invisible authors, claims without evidence, brands with rankings but no reputation.

Everything it rewards — original thinking, genuine authority, third-party trust, memorable brand — is what strong marketing looked like before we all became algorithm farmers. The machines, of all things, are pushing marketing back towards substance.

We'd call that a trend worth getting ahead of.