Questions, answered in full

Search and reach optimisation (SRO) is the practice of making a business findable, quotable and correctly described across both traditional search engines and AI assistants such as ChatGPT, Perplexity, Gemini, Claude and Copilot. It combines tactical keyword and prompt-level work with strategic positioning across the wider online channels that AI systems draw on.

It is our own term, and we should be transparent about that. We use it because "reach" captures something the other labels miss: appearing in an answer is only half the outcome, and being described accurately when you appear is the other half.

SEO optimises for a ranked list of links; search and reach optimisation optimises for a written answer. The mechanics differ: AI systems synthesise from sources they trust rather than ranking pages, they weigh brand presence and third-party mentions more heavily than backlinks, and they favour recently updated, clearly structured content.

Traditional SEO still matters — the two compound rather than compete — but strong rankings no longer guarantee you appear in AI answers. The overlap between the pages Google ranks highly and the sources assistants actually cite has fallen dramatically.

They are synonyms, and arguing about which is correct wastes time better spent on the work. GEO (generative engine optimisation) has the most academic grounding, having been formalised in peer-reviewed research in 2024. AEO frames the same activity around the output; LLMO frames it around the technology.

Pick one, use it consistently across your organisation, and make sure everyone agrees on the outcome being measured rather than the label.

A mention is when an AI answer names your brand. A citation is when your content is used to build the answer — sometimes without your name appearing at all. The two come apart more often than people expect.

The worst version is surprisingly common: your content is good enough that the model uses it to answer the question, but your brand is not strong enough in its sources for you to be the one recommended. You are doing the work and a competitor is getting the deal. It is a different problem from simple invisibility and it needs a different fix, which is why we track both.

ChatGPT, Perplexity, Google's AI Overviews and AI Mode, Gemini, Claude and Microsoft Copilot as standard. Coverage changes as the market does — we add platforms when they start influencing real buying decisions in your category, rather than when they launch.

Less traffic, better traffic. Volume from AI assistants remains small relative to organic search, and anyone telling you otherwise is selling something. But those visitors arrive far later in the buying process — often already shortlisted, with budget — and convert at several times the rate of ordinary organic visitors.

The larger effect is the one you cannot see in analytics. A buyer asks an assistant who they should consider, gets three names, and starts their evaluation from that list. That shortlist forms before any trackable visit exists.

The terminology probably is; the underlying shift is not. Whatever the practice ends up being called, buyers asking a machine for a recommendation instead of scanning a list of links is not a trend that reverses.

The honest caveat is that the tactics will change, possibly quite fast. Anything we tell you about how the models weigh sources today is a snapshot. What does not change is the underlying requirement: be clearly described, be present in the sources that matter, and keep both current.

No, and mid-sized companies often have the bigger opportunity. AI assistants weigh how clearly and consistently a business is described more heavily than how large it is, which means a focused specialist can be recommended ahead of a household name for a specific buying question. Narrow, well-defined categories are where we see the fastest gains.

Yes, and it is usually where the biggest gaps are. Assistants frequently name different brands depending on the language of the question, because they draw on different sources: national review sites, local directories, domestic trade press, country-specific regulatory context.

A company can be well represented in English and effectively absent in German — and almost nobody checks. We run prompt sets in German, French, Italian, Spanish and English as standard, and add languages where your buyers are.

Yes, and usually harder. In health, finance, legal and other areas where a bad answer causes real harm, assistants are noticeably more conservative: they hedge, they recommend consulting a professional, and they lean much more heavily on institutional and editorial sources rather than company websites.

That means the work shifts away from your own domain and towards the sources those models trust in your sector. It takes longer and the ceiling is lower — but so is the competition, because most firms in regulated categories have not started.

Sometimes, yes — and we would rather say so than talk you into something. If you have a content team with spare capacity, a technical resource who can restructure your site, and someone willing to spend months chasing third-party source corrections, then a monitoring tool plus internal execution is a perfectly reasonable plan. Tools start at around a hundred pounds a month.

What we see is that most companies have the tool and none of the other three. The dashboard tells them they are invisible, produces a list of things to fix, and then sits unused from the second month onwards because nobody owns the work. If that sounds familiar, that is the gap we fill.

Level 1 gets you into the answers. It is tactical, measurable and moves within a quarter: mapping the questions your buyers ask, establishing where you stand, and doing the content and structural work to get you cited.

Level 2 shapes what those answers say. It is strategic and slower, covering your entire online footprint — reviews, directories, editorial coverage, comparison content, the sources you do not own — and aligning all of it with the position you want to hold. Most clients start at Level 1; the findings usually make the case for Level 2 better than we can.

We write it. Comparison pages, definitions, answer-first explainers, structured FAQs and the markup underneath them. You approve everything before it publishes, and it goes out in your voice rather than ours.

We go and get them changed. Outdated directory listings corrected, review profiles refreshed, comparison articles updated where they have you wrong, and presence built in the places your category is genuinely discussed.

This is the slowest and least glamorous part of the job, it cannot be automated, and it is usually where the biggest gains sit — because it is the part everyone else leaves on a recommendations slide.

Not directly — nobody can edit an AI's answer. What can be changed is the evidence it draws on. When the underlying sources are corrected, updated and outweighed by accurate ones, the description shifts. It takes longer than anyone would like, and it works.

No, and be sceptical of anyone who does. AI assistants generate answers dynamically from sources they weigh in ways no one outside those companies controls.

What we commit to is doing the work that demonstrably shifts the odds, showing you the movement transparently, and telling you early if your category is one where progress will be slow.

First movement typically appears within six to ten weeks, with meaningful change over a quarter. Repositioning work at Level 2 takes two to four quarters.

The variable is how much of the work sits outside your own domain. Changes to your site register relatively quickly; changes that depend on a third party updating their page move on their schedule, not ours. We tell you which of those your situation is before you commit.

A written brief at the start, access to your website and analytics, one person empowered to approve content, and replies within a couple of days. That is genuinely it. No workshops, no discovery days, no standing meeting.

Three things, which are becoming the category standard:

Visibility — the share of answers where you are named. Position — whether you are the first name or the afterthought, because being mentioned third is not the same as being recommended. Sentiment — how you are described, which is often where the commercial damage sits.

We also track which sources the models are drawing on, split by language and market, and separate brand mentions from source citations.

No. Keep it, and keep using it to check our work — we would rather you did. Our platform is where you steer the engagement, not a replacement for your analytics stack, and we are happy to work from the data you already have.

If anything, running a third-party tool alongside us is the cleanest way to verify that what we report is real.

Because it is, and because most of the effect never becomes traffic at all. Judging AI visibility by referral sessions is like judging a trade show by how many people walked past your stand — it misses the conversations that happened.

Look at conversion rate on the traffic that does arrive, and at your visibility and position trend against named competitors. Those are the numbers that reflect what is actually changing.

Honestly, attribution here is imperfect and we will not pretend otherwise. What we can show is a clean before-and-after on a fixed prompt set, run identically each month, with your competitors as a control group. If your visibility moves and theirs does not, that is about as clear a signal as this discipline currently offers.

Rarely, and never as the default. Work moves in writing, which means decisions happen in an afternoon rather than waiting a fortnight for a slot that suits six calendars, and every recommendation and rationale stays documented and searchable.

When something genuinely benefits from a conversation — a difficult decision, a sensitive finding, a new stakeholder — we get on a call. We just will not put thirty minutes in your diary to tell you something that fits in a paragraph.

Usually better than clients expect — but we would rather be honest about the trade-off than talk you into a bad fit. If your organisation makes decisions in rooms, and a written recommendation will not carry weight without someone presenting it, we are probably not the right partner. We will tell you that at the scorecard stage rather than three months in.

The people you are introduced to at the start. We do not run a pitch team and a delivery team, and there is no hand-off to a junior in month two.

Because the difference between a good result and an exceptional one is in details that do not survive being spread across fifty accounts. It does mean availability is limited, and occasionally it means telling a company we would like to work with that we cannot start until next quarter.

[Level 1 starts at £X per month on a [N]-month minimum; Level 2 starts at £X.] Pricing depends on the number of markets and languages, how much of the work sits on third-party sources, and how much content production is involved. You will get a fixed figure in writing with the proposal — no day rates, no scope creep.

English, German, French, Italian and Spanish as standard, with others added where your buyers are. Because we are distributed across Europe, this is a native capability rather than a translation exercise — which matters, because a prompt set translated word-for-word does not reflect how people actually ask questions in another language.

Because our clients ask us to keep them that way, and their visibility strategy is commercially sensitive. Several work in sectors where discretion is a condition of the engagement.

We will not publish one client's results to win the next one — which is precisely the discretion you would get from us.

Yes, once a conversation is serious. Where a client has agreed in advance, we will introduce you directly and step out of the way. We will not do it at first enquiry, because their time is not a marketing resource.

The clearest is a health sector client that came to us in early 2025 with effectively no presence in AI-generated answers for its category, and by early 2026 had become the most frequently mentioned brand in that category across all the major assistants.

We are happy to walk through the method, the prompt set and the measurement behind that at the appropriate point in a conversation.

Check us the same way you would check a client. Open an AI assistant and ask which companies do AI search visibility work, or ask what it knows about Ergovero. If we are not in the answer, that is a fair thing to hold against us.

It is an uncomfortable test to invite, and it is the only one in this category that a buyer can run themselves in thirty seconds.