A competitor is. The answer only names three.
Your marketing does its job — someone becomes interested and goes to check you out. That check increasingly happens inside an AI answer, and Bain puts around 60% of searches as ending without the user reaching any site at all.[5] If the answer does not name you, you have paid to create demand that someone else converts.
Getting named takes technical work and earned media together. We do both.
Simplicity is a digital specialist inside a traditional advertising agency — London, Toronto and Halifax. Answer engines need schema, entities and crawler access on one side, and press, reviews and directories on the other. Most suppliers only do one of those.
Trusted by brands across Canada, the UK and EMEA
Nothing about your advertising has stopped working. What changed is the step immediately after it — the quiet moment where an interested buyer verifies you before they get in touch. That used to be a search and a few clicks. Now it is often a single question to an assistant, and a short answer naming two or three companies.
You can buy a CTV impression, an OOH face or a podcast mid-roll. You cannot buy a sentence in a ChatGPT answer. That does not make it unbuyable — it makes it earned, and earned inventory is won by the agency that plans for it first.
The click it replaces is measurable. Across 68,879 Google searches from 900 US adults, users clicked a traditional result on 8% of visits where an AI summary appeared, against 15% where none did — and clicked a link inside the summary itself on 1%.[1] Roughly half the onward journey, gone from the surface you were planning against.
Across 68,879 Google searches from 900 US adults, people clicked a traditional result on 8% of visits where an AI summary was shown, against 15% where none was — and clicked a link inside the summary itself on just 1%.[1] Roughly half the onward journey disappears. What the summary said about you is doing the work the click used to do.
Outbound referral traffic from ChatGPT to the rest of the web grew 206% year on year to January 2026, measured across more than a billion lines of US clickstream data.[7] The people who do click through are arriving from a recommendation rather than a list — which is why they behave differently once they land.
Averaged over the year to May 2026, up 70% in twelve months across 655 million unique visitors.[6] This is not a niche behaviour among early adopters any more. It is where a large share of consideration now happens, including for the buyers your existing advertising is already reaching.
Absence from an AI answer is not random and it is not a mystery. It comes down to whether the engines can read you, place you, or verify you — and the fix for each is a different piece of work. Two of them are technical. The third is not, which is where most suppliers stop.
The first two causes are technical. Either the site blocks AI crawlers or renders in a way they cannot parse, or the entity is ambiguous — inconsistent naming, missing schema, thin structured signals, nothing telling a machine unambiguously what you are and what you do. Ranking well does not rescue either. BrightEdge's own 16-month finding is convergence, with citation-to-ranking overlap rising from 32.3% to 54.5% by September 2025, but only 16.7% of citations come from the organic top ten and most cited pages sit between positions 21 and 100.[2] A category leader can own the blue links and still be missing from the answer above them.
The third cause is the common one, and it is not on your website. Omniscient Digital analyzed 23,387 unique sources cited across 240 branded queries in ChatGPT, Perplexity, Gemini, AI Mode and AI Overviews. Editorial, forums, review sites and directories account for nearly half of every citation, against 23% for the brand's own site. On customer review queries earned media reaches 82%.[9] If the sources an engine trusts have nothing to say about you, no amount of on-site work will put you in the answer.
This is the awkward shape of the problem, and the reason it so often gets half-solved. A technical SEO supplier can fix crawler access, schema and entity structure, and will make real progress — the 2023 generative engine optimization paper from Princeton, IIT Delhi, Georgia Tech and the Allen Institute for AI found structured methods, principally adding citations, quotations and statistics, lifted source visibility by up to 40% on the authors' own benchmark.[10] But that supplier cannot place you in trade press or get you reviewed. A PR agency can do the second and will not touch your schema. Simplicity is a digital specialist that sits inside a traditional advertising agency, so both halves are in the same building, working off the same prompt set and reported in the same monthly document. That is the whole reason we can sell this credibly.
A mention inside an answer is closer to a recommendation than to a listing, and the numbers behave accordingly — both in how many people come, and in what they do when they arrive.
Seer Interactive tracked 53 brands across 5.47 million queries and 2.43 billion organic impressions from January 2025 to February 2026. Brands cited in an AI Overview averaged 2.07% organic click-through; brands not cited averaged 0.94%.[3] Seer are explicit that this is an association rather than proof of cause — cited brands may simply be better known. Worth reading as a strong signal, not a guaranteed multiplier.
Adobe Digital Insights tracked AI-sourced traffic to US retail sites from October 2024 to May 2026. In May 2026 these visitors were 36% less likely to bounce, viewed 23% more pages per visit and spent 53% more time on site — before you get to the 54% higher conversion rate and 53% more revenue per visit.[4] Scope it honestly: that is US retail in a single month, and non-retail verticals are materially weaker — financial services shows no conversion uplift at all.
Similarweb found the share of US ChatGPT prompts returning any cited source rose from 1.6% to 6.8% between June 2025 and May 2026 — that is the rate at which answers cite anything at all, not a brand citation rate. It varies sharply by sector: travel and hospitality carry a citation about 23% of the time, automotive roughly 20%, professional services under 4%.[6] Read it as a map of how far along each category is. Where citation rates are already high the answers are contested and entry costs more. Where they are low the habit is still forming, and the sources that get quoted early tend to keep getting quoted — which makes now the cheapest this will be.
None of this is a proprietary trick, and we will tell you exactly what we are doing at every stage. The difficulty is that it needs two skill sets running in parallel against one fixed measurement set — which is the part that usually does not happen.
Nothing starts until we know what the engines say about you today, in their own words, and which of the three causes is holding you back.
Make the site mechanically easy to quote and your identity unambiguous to a machine. This is engineering work, and we do it in-house.
48% of citations come from property you do not own.[9] This is press and relationships, and it is the half a technical supplier cannot do for you.
One number, tracked from the baseline: how often you are named. Monthly, per engine, against competitors you name yourself.
In 2025 we asked the engines who the best CTV and out-of-home agencies in Canada were, and we were not in the answer. So before selling this to anyone, we spent six months fixing it on our own site — both halves, technical and earned, exactly as described above.
We fixed a 50-prompt set covering how Canadian advertisers ask about CTV, out-of-home and digital audio buying, recorded a baseline of zero, and ran all four workstreams against it. Structured service pages and a case-study library written for extraction, full entity and schema build, then trade press and directory placement. No paid media supported any of it, so the movement below came from the work alone. Reported monthly against a named competitive set — the same document a client gets.
Canadian dollars, excluding applicable GST, HST and QST. No setup fee, and no tool subscriptions marked up and passed back as a line item.
Find out exactly what the engines say about you, and which of the three causes is yours. No commitment beyond it.
The technical half, run continuously — structure, schema and entity — with monthly reporting.
Adds the earned half — press, reviews and directories, where 48% of citations actually come from.
All figures are Canadian dollars, excluding applicable GST/HST/QST. They sit inside the published North American mid-market band of roughly US$3,000–$10,000 per month for a full AEO program.[11][12][13] Regulated sectors — financial services, legal, healthcare — are quoted at an additional 35% premium over the equivalent unregulated retainer, reflecting compliance review load. We will quote that explicitly if it applies to you.
*Digital PR has tiered options for external placement depending on product and industry, these are priced separately.
Three things have to be true, and they are worth checking in order. ChatGPT has to be able to reach and parse your site, so crawler access and rendering come first. It has to be able to place you unambiguously as an entity, which is schema, consistent naming and structured signals. And it has to find corroboration in the sources it trusts — press, reviews, directories, communities — because 48% of the sources cited in AI answers are earned media and only 23% are the brand's own site. Most brands have already done some version of the first two through normal SEO work and have done nothing at all about the third, which is usually why they are still absent. [9]
Almost always one of the three above, and it is diagnosable rather than mysterious. The engines cannot read you, cannot place you, or cannot verify you. The third is the most common and the most frustrating, because nothing you do on your own website fixes it — if the publications and review platforms an engine trusts have no substantive mention of you, you will not appear no matter how good your pages are. Our CA$4,500 Diagnose engagement tells you which of the three applies to you and hands you the plan, whether or not you go further with us.
Perplexity leans harder on live retrieval and shows its sources openly, which makes it the most legible of the engines to work on — you can see precisely which pages are being pulled for a given question. In practice that rewards two things: pages that answer a specific question cleanly near the top rather than burying it under preamble, and presence in the specific publications it keeps returning to for your category. We baseline it alongside ChatGPT, Gemini, Claude and Google AI Overviews, because the source sets differ and a brand can hold position on one and be invisible on another.
Ranking helps, but it does not settle it. BrightEdge's own 16-month conclusion is convergence — overlap between AI Overview citations and organically ranking pages rose from 32.3% to 54.5% by September 2025 — yet only 16.7% of citations come from the organic top ten, and most cited pages sit between positions 21 and 100. You can be first on Google and absent from the answer above your own listing. What moves it is content structured for extraction, an unambiguous entity, and being present in the sources being quoted. One common misconception worth clearing up: Google-Extended does not control AI Overviews eligibility. It governs Gemini training and grounding. AI Overviews is a Search surface fed by Googlebot and governed by the normal snippet controls. [2]
It overlaps, and anyone claiming it shares nothing with SEO is overselling. Crawlability, clean information architecture and structured data matter to both. The difference is what you are competing for and where the work happens. SEO competes for a position in a list of links, and almost all of the work is on property you control. AEO competes for inclusion in a generated answer that may name only two or three companies, and roughly half the work is on property you do not control. That second half is press and relationships rather than engineering, which is the honest reason a technical SEO supplier struggles to deliver it alone.
Because the problem needs two skill sets that rarely sit together. Simplicity is a digital specialist inside a traditional advertising agency: the technical team does schema, entity resolution, crawler access and answer-structured content, and the same building has the press and media relationships needed to get you into the third-party sources engines actually quote. A pure technical supplier does the first half well and cannot do the second. A PR agency does the second and will not touch your schema. Running both against one prompt set, on one monthly report, is the entire proposition.
Frequently, and it works in your favour. For market-specific questions the sources engines lean on skew local — Canadian trade press, Canadian directories, Canadian review platforms — which is a much smaller and more winnable pool than the global one. It is also the part of this work that benefits most from an agency that already has those relationships in market. We baseline Canadian-phrased prompts separately from generic ones for exactly this reason, because a brand can hold share on one and none on the other. [9]
Published North American benchmarks put a one-off audit at roughly US$1,500 to $5,000, mid-market monthly retainers at US$3,000 to $10,000, and enterprise scopes at US$15,000 and above. In Canadian dollars the picture tightens: Toronto agencies quote CA$1,500 to $5,000 per month for single-channel retainers, and a US$8,000 retainer lands near CA$11,000 once FX and tax are applied, which is why we quote in CAD and publish the numbers. Diagnose is CA$4,500 one-off, Presence is CA$6,000 per month on a six-month minimum, and Share of Answer is from CA$13,000 per month on a twelve-month minimum. On the same benchmarks, anything advertised under about CA$2,000 per month is usually rebranded SEO. [11][12][13]
The technical half moves first. Crawler access, schema and entity work can change how you are described within weeks, because the engines re-retrieve continuously. The earned half is slower and honest about it — placement takes time to commission, publish and then be picked up, so meaningful movement in how often you are named typically appears between month three and month six. That is why Presence runs on a six-month minimum. If share of answer is not moving by month four, you will hear it from us before you ask.
No, and anyone who does is either misleading you or does not understand the mechanism. Answer engines are ranked retrieval systems with undisclosed weightings that change without notice, and none of them sell placement — there is no rate card and no insertion order. What we commit to is a measured baseline before any work starts, the work itself done in full, monthly reporting per engine against competitors you name, and telling you plainly when a surface moves against us.
If we are going to argue that citation is the currency, we should show ours. All figures below link to the originating publisher, not to a summary of it.