Answera by Groundbase Request access

Become part of the answer.

Answera reads your page the way an AI system does and measures it on seventeen points — with, for every point, the evidence drawn from the page itself. You get the shortest set of changes that opens the earliest closed gate.

Invite-only · one page per report · by Groundbase

The problem

Search is moving from links to answers.

A growing share of queries is now resolved inside AI systems — ChatGPT, Google's AI Overviews and others — which read a page, decide what it is, and either use it or move on.

Gartner forecasts that traditional search-engine volume will fall by 25% by 2026 as users shift to AI chatbots and other virtual agents (Gartner, 19 February 2024). A page that never states plainly what it is, who it is for and what it claims does not exist for a language model — however well it ranks in a traditional search index. The work is no longer only to be found; it is to be understood, extracted and chosen. Most pages were never written for that, and there is no dashboard telling you why you were skipped.

This pattern is observable, and it can be corrected.

What is GEO?

What is Generative Engine Optimisation?

Generative Engine Optimisation is the practice of measuring and improving how generative AI systems find, mention, describe and recommend a brand, product or organisation — through on-site optimisation, off-site presence and external entity building. Optimisation does not determine the outcome; it increases the odds of presence and its quality.

Answera works on the on-site half of that practice, one page at a time: it measures whether a page is built so that an AI system can retrieve it, understand it, feature it and attribute it — and shows what is missing. Classic SEO still matters underneath: a page has to be reachable and indexed before anything else can happen.

What the bot sees

You browse your site with Chrome. AI systems don't.

Answera announces itself as a crawler and renders like a browser — so it sees what AI systems are given, and that can differ from what you see.

We have measured pages that returned an error to every bot user-agent and a normal page to the browser: invisible to the owner, who visits his own site with Chrome, and decisive for every gate that follows. A page that blocks the bot gets no score and no advice — we do not judge content we never saw.

How it works

An AI answer is assembled through four gates, in order.

Each gate is a question the system must answer “yes” to before the next one matters. Answera checks them in sequence and reports on the earliest one that is closed.

gate 01

Access

Can the system reach your page and process it? If a crawler is blocked, the markup is unreadable, or the content only appears after scripts run, nothing downstream can happen.

gate 02

Retrieval

For a given question, is the right passage on your page found at all? A page can be reachable and still never surface, because nothing on it clearly matches how people ask.

gate 03

Extraction

Once found, is there a self-contained, quotable answer? A sentence that stands on its own — without the paragraph around it — is what a model can lift into a response.

gate 04

Selection

Among the candidates, does your answer win? Clarity, specificity and a defensible claim decide which source the system uses when several could serve.

Failing at an early gate makes the rest irrelevant, so the report is ordered the same way: it opens the earliest closed gate first, then re-checks what follows.

What you get

You receive one report per page.

It diagnoses the page across three pillars, then gives the shortest action list that opens its weakest gate. Every action names the endpoint that tells you it is finished.

Pillar 01

SEO

The technical and structural conditions for a page to be reachable, parsed and indexed at all.

Pillar 02

GEO

Whether the page offers self-contained, extractable answers that AI systems can retrieve and quote.

Pillar 03

Language & sentiment

Whether the wording is clear, specific and defensible enough to be chosen over the alternatives.

Each recommended action carries a plain endpoint, so there is no ambiguity about when it is done:

Done when · the page states, in one sentence, what it is and who it is for.
Excerpt from a real Answera report: score card and synthesis.
Excerpt from a real Answera report: a prioritised action list item with its done-when endpoint.

From a real report — our own site, scanned July 2026.

No invented facts

The report never fabricates data about your page or your market. If something cannot be established from the page, it says “not assessable” — in both directions: we do not fill gaps in your favour either.

Potential, not status

Answera measures what is in the page, not what the page currently earns in AI answers. No citation counts, no predictions, no guarantees — a potential measurement, honestly labelled.

Facts plainly, judgements as judgements

Measurable findings are stated directly; opinions are labelled as opinions, so you can weigh them yourself.

Two scans, same answer

Run the same page twice against the same version and the report says the same thing — and every report carries the version block that proves which yardstick produced it.

A blocked page gets no score

If the crawler is refused, you get the finding and the next step — not a score over content we never saw, and no advice about it either.

Shortest path first

The action list is deliberately brief: the fewest changes that open the weakest gate, in the order that matters.

Evidence

A real page, before and after

Our first public example is our own site, groundbase.be. We ran the audit, carried out the shortest action list it produced, and ran it again — same measurement, no special treatment. This is what one focused afternoon of changes did.

Before
63/ 100

17 findings, including: the page never states what the platform is or who is behind it; and no self-contained answer near the top.

After
72/ 100

11 findings, now refinement rather than fundamentals.

“A strong page — what remains is polish, not foundation.”

Change 01

A citable definition, first

One sentence directly under the headline stating what the platform is — the answer an AI can lift.

 Done when: the opening states what this is and who it’s for.
Change 02

A named author

The person behind the platform now appears in the visible text and in the structured data — not just implied.

 Done when: the author is named, visibly and machine-readably.
Change 03

An FAQ that covers the follow-ups

Five questions, each answered in the first sentence, marked up as structured data.

 Done when: every question has a self-contained answer.

Two recommendations we deliberately chose not to follow: the audit is a shortest path, not a checklist — you decide which gates you want to open. That distinction is the point of the tool.

Both runs: fresh scans of the same URL, identical settings, July 2026.

For whom

Answera is for marketers and site owners who want their pages to work in AI answers.

If a page carries your positioning, your product or your expertise, it is a candidate. Answera works with any content management system, because it reads the page as an AI system does — from the rendered result, not from your stack.

Answera is built by Jan Van Hove at Groundbase, the Belgian consultancy for GEO and AI visibility. Jan brings over twenty years in digital marketing, with a specialisation in Generative Engine Optimisation.

01
Marketing and content teams
02
Founders and site owners
03
SEO and GEO specialists
04
Any CMS — no integration required

FAQ

Frequently asked questions

What is Generative Engine Optimisation?

Generative Engine Optimisation is the practice of measuring and improving how generative AI systems find, mention, describe and recommend a brand, product or organisation — through on-site optimisation, off-site presence and external entity building. Optimisation does not determine the outcome; it increases the odds of presence and its quality. It shifts the emphasis from keywords and backlinks towards clear, self-contained, verifiable statements that a language model can lift with confidence.

Who is Answera for?

Answera is for marketers, content teams and site owners who want their pages to work in AI answers. If a page carries your positioning, your product or your expertise, it is a candidate — on any CMS, with no integration required.

Which AI systems does the audit cover?

The audit assesses the mechanisms shared by systems such as ChatGPT, Google’s AI Overviews, Claude, Gemini and Perplexity: access, retrieval, extraction and selection. It measures the page against those mechanisms rather than against any single provider.

How long does a report take?

A report is typically prepared within a few working days of your request. Each report covers one page, diagnosed across three pillars, with the shortest action list to open its weakest gate.

How much does Answera cost?

Answera is invite-only: reports are prepared on request and invoiced. Request access and we will confirm the details for the page you choose.

Next step

Become part of the answer.

Answera is invite-only. Request access and we will prepare a report for one page you choose.

Request access

hello@answera.app