Strategy

From SEO to AEO: why answer engines need their own strategy

Search rankings don’t tell you what AI assistants actually say about you. Here’s how to bridge the gap between traditional search and the new AI reality.


For the last 20 years, “being found” has meant one thing: getting to page one of Google.

We built content strategies around keywords, obsessively watched position tracking tools and celebrated when an important term nudged from #6 to #3. SEO became a discipline, a budget line, a career.

That world hasn’t disappeared. But something important has changed.

Your buyers are increasingly skipping the SERP altogether. They’re opening ChatGPT, Perplexity, Gemini or Copilot and asking:

  • “Who are the leading providers of [category] for [ICP]?”
  • “How do I solve [job-to-be-done]?”
  • “Which tools are best for [very specific use case]?”

They’re not scanning 10 blue links. They’re reading a single, synthesised answer – and perhaps clicking one or two recommendations the assistant offers.

Welcome to the era of Answer Engine Optimisation (AEO).

Search vs answers: what actually changed?

Traditional search and AI assistants both start from a query box, but they operate – and influence buyers – in very different ways.

1. From lists to answers

  • Search engines return a list of pages ranked by relevance and authority.
  • Answer engines (LLMs and AI assistants) return a narrative answer that blends explanation, opinion and recommendations.

In the search world, visibility is a rank on a page. In the answer world, visibility is whether you’re:

  • mentioned at all
  • recommended explicitly
  • or quietly ignored.

2. From “clicks” to “trust”

SEO reports tell you about:

  • impressions
  • clicks
  • click-through rate
  • time on page.

They say almost nothing about how the buyer’s understanding of the problem and market shifts while they read.

Answer engines, on the other hand, shape the story directly:

  • They define the category.
  • They decide which options are “top” or “recommended”.
  • They frame trade-offs and risks.

The buyer’s first serious impression of your category – and whether you belong in it – may now come from an AI answer, not your homepage.

3. From keyword lists to research journeys

Traditional SEO tends to start with keywords and search volumes:

  • “What do people type into Google?”
  • “Which keywords can we realistically rank for?”

Answer engines, however, are fed questions that look more like natural language:

  • “We’re a mid-market UK bank, how should we approach [X regulation]?”
  • “What’s the best way to…?”

These are research journey questions, not tidy keyword stems. They reflect where someone is in their decision process, not just what they’re trying to type into a search bar.

If you want to influence AI answers, you can’t just think in keywords. You have to think in ICPs, needs and journeys.

Why SEO alone won’t cut it anymore

SEO isn’t dead – but it is partial.

Here’s what it still does brilliantly:

  • Helps you understand and capture intent that still goes through Google
  • Ensures your site is technically crawlable and fast
  • Gives you a framework for content structure and internal linking
  • Delivers a steady stream of organic traffic in many categories.

But it doesn’t tell you:

  • What ChatGPT says when someone asks about your category
  • Whether you appear in Perplexity’s top three recommendations
  • How you’re framed vs competitors in Gemini’s answers
  • Which of your assets are influencing those answers (if any).

That’s a problem if:

  • Your category is complex or high-value
  • Your buyers are time-poor senior stakeholders
  • Your competitors are already experimenting with AI search.

You can be winning in SEO and losing in AI, and your dashboards will look perfectly healthy while it’s happening.

What is Answer Engine Optimisation (AEO)?

Answer Engine Optimisation is about deliberately influencing the answers your buyers see from AI assistants, not just the search results they might have seen from Google.

At its core, AEO means:

  1. Understanding the questions your ICP actually asks across their research journey.
  2. Seeing what AI assistants currently say in response – including which brands they mention and recommend.
  3. Creating and structuring content so that, when those assistants look for evidence, your material is:
    • easy to find
    • easy to interpret correctly
    • strong enough to be used in the answer.
  4. Re-testing over time to see how those answers shift as you and your competitors publish.

AEO isn’t a replacement for SEO. It’s an additional layer that sits on top of your existing search and content work.

How to bridge the gap: a practical framework

You don’t have to throw away your SEO playbook. You do have to add some new pages to it.

Here’s a structured way to move from “we know our rankings” to “we know what AI says about us”.

Step 1 – Map your ICP and research journey

Start with who and how, not keywords.

  • Define your Ideal Customer Profile (ICP): Sector, size, geography, role, triggers (what makes them seek a solution now?)
  • Map the research steps they go through, for example:
    1. Early exploration – “What is this? Why should we care?”
    2. Problem framing – “What does this mean for us?”
    3. Solution discovery – “What options exist?”
    4. Selection – “Can we trust this vendor?”

For each step, collect real questions your ICP might ask an AI assistant (from sales calls, interviews, SMEs, etc.). You’re building a question universe that reflects buyer behaviour, not a keyword spreadsheet.

Step 2 – Run those questions through AI assistants

Pick a small set of AI tools your buyers are likely to use (ChatGPT, Perplexity, Gemini, Copilot).

For each question, capture:

  • The full answer
  • Whether your brand is mentioned
  • Whether any of your specific pages or assets are cited or linked
  • Which competitors are mentioned
  • How the answer frames the problem and category.

At this stage you’re not optimising – you’re benchmarking. You’re building an “answer snapshot” of how the market looks today in AI.

Step 3 – Score your AI visibility

To make this actionable, you need more than anecdotes and screenshots. Create a simple scoring model, for example:

  • Brand visibility:
    0 = not mentioned | 1 = mentioned, not recommended | 2 = recommended / highlighted
  • Evidence use:
    0 = no assets used | 1 = brand mentioned with no concrete evidence | 2 = answer cites/links to your pages
  • Competitive position:
    0 = competitors only | 1 = you + competitors | 2 = you clearly differentiated / preferred

Roll those up by Theme, Need, Journey step, and AI engine. You now have an AI Visibility Scorecard.

Step 4 – Connect AI answers to your content

Next question: why are you invisible or weak?

Audit your existing content through the same lens. For each Theme / Need / Step, ask:

  • Do we have one or more strong assets?
  • Are they open, crawlable and easy to interpret?
  • Do they offer evidence (data, examples, process detail), not just high-level claims?

You’ll usually find three types of gap:

  1. No content – you simply haven’t written anything that would help an AI answer the query.
  2. Weak content – you have a blog post, but it’s thin, vague or generic.
  3. Invisible content – you have great material, but it’s buried, hard to parse, or blocked by technical issues.

Step 5 – Build a 90-day AEO roadmap

Keep it focused. You don’t need to optimise everything at once.

For the next 90 days, pick:

  • 2–3 priority Themes
  • 3–5 high-value Needs / jobs
  • 1–2 key journey steps where being visible would really change the pipeline.

For each, define flagship assets to create or refresh (guides, case studies, comparisons) and structural improvements (headings, schema, internal linking).

And importantly, define a measurement plan: when you’ll re-run the AI questions, how you’ll tag AI-driven traffic, and which business metrics you’ll watch.

What about tools and platforms?

Today, most organisations are doing this manually: copy-pasting prompts, taking screenshots, sharing anecdotes.

That’s a start. But it’s very hard to maintain consistency, compare runs over time, and tie insights back to your content and analytics.

That’s exactly the gap platforms like Odyssiant are emerging to fill: turning Answer Engine Optimisation from a one-off experiment into a repeatable, measurable discipline.

Regardless of whether you use a tool or a spreadsheet, the principle is the same: Treat AI answers as a channel you can measure and influence, not a black box you hope for the best with.

What to do next

If you do nothing else after reading this, do this one exercise:

  1. Pick one of your most important ICPs.
  2. Write down 10 questions they might ask ChatGPT when they’re trying to understand your category or shortlisting vendors.
  3. Actually ask those questions in an AI assistant and note: Do you appear? Who else does? How are you described?

That’s your first AI visibility baseline.

From there, you can decide whether AEO is a minor curiosity or a serious strategic gap.

Either way, the days of saying “we rank #1 for [keyword], so we’re fine” are numbered. Your buyers have moved. It’s time your visibility strategy caught up with them.

Want to see your AI Visibility Score?

Odyssiant can run this audit for you automatically across multiple LLMs.

Related reading