See how leading AI answer engines are likely to represent your business. Below is what that means in practice — including the parts that are limitations rather than features. For the exact endpoint, model and settings per engine, see our methodology.
aisayso measures how visible your business is inside AI answers. We generate 50 buyer questions from your own website and ask them to three AI answer engines — ChatGPT, Google AI Overviews and Perplexity. For every answer we record whether your brand is mentioned, how prominently (from a passing reference to an outright recommendation), which competitors appear, and which sources the answer cites. See how leading AI answer engines are likely to represent your business.
Yes. The public scan is free and anonymous: no account, no card, no obligation. It is a real measurement, not a demo — we read your site, write your buyer questions and run 15 of them live across all three engines. Be aware of what free means here: there is no free plan and no trial. The free scan is the free part; everything after it needs a paid plan.
Before you pay you see your AI visibility index out of 100 with its stability band, how many of the tested buyer questions mention you, one concrete insight taken from your own measurement, which engines answered, and the headline of up to three recommended actions. Everything else stays on our servers. The dashboard you see blurred behind the pricing card is an empty skeleton — there is no hidden report underneath it to click away.
A paid plan unlocks the full report: all 50 buyer questions across the three engines, the literal answer per question with the sources it cited, which competitors won the questions you lost, the complete prioritized action list with ready-made material (draft pages, JSON-LD, review-site lists and outreach templates), weekly tracking with your history, the weekly email report, and monthly credits for on-demand work. Agency and Enterprise add white-label PDF reports under your own brand.
Automatically once a week. A weekly run takes 1 sample per buyer question per engine, always in the same clean setup, so week-over-week differences reflect the web and not our settings. Once a month a stability run repeats each question 3 times per engine to establish how noisy that question is. Your first measurement is a baseline: no trend, no delta, no arrow. From roughly 3 runs onwards we describe direction in words instead of inventing a precise change.
Yes, whenever you want — and it costs credits, because every recheck is a real paid API call. An instant recheck of a single question costs 15 credits and returns one sample, labelled indicative: it never overwrites your weekly score. A full re-scan of all questions costs 100 credits, and generating a full piece of content in Content Studio costs 120 credits. Every plan includes a monthly credit allowance that resets each month; top-up packs are available if you run out.
They are the questions your customers actually ask an AI assistant before they choose a supplier — 50 of them per site, in five categories: category questions, recommendation questions, comparison questions, use-case questions and brand questions. We write them from your own website, in the language of that site, and de-duplicate them by meaning so no two questions ask the same thing. Each one carries its service, location, intent and category, so you can see exactly why it is on the list.
Yes. You can edit the wording of any question, delete the ones that do not fit your business, and add your own, up to the 50-question limit that we enforce on the server. Your own questions are marked as yours, and regenerating the set never overwrites or silently drops them. Editing keeps the question's place in the list: the previous wording is archived with a timestamp, and older measurements stay attached to the version that was live when we took them — nothing is deleted or quietly relabelled.
No, and we will not pretend otherwise. We measure through the official APIs in a clean, consistent setup: never signed in, no personalisation, no memory, no chat history, no location history. A signed-in person can therefore see something different from what we record. That is a deliberate trade-off — the consistency is what makes a weekly trend meaningful at all. See how leading AI answer engines are likely to represent your business. We also do not scrape logged-in consumer apps: it breaks the providers' terms, it is technically fragile, and it is legally risky. Our methodology page lists the exact endpoint, model and settings per engine.
Three, today: ChatGPT through the OpenAI Responses API with the web_search tool enabled; Google AI Overviews through a SERP provider that returns the AI Overview Google itself shows for that search; and Perplexity through its Chat Completions API on the sonar model. When Google shows no AI Overview for a question, that counts as a valid measurement with an empty answer, not as a failure. If an engine cannot be measured at all, the answer is marked not measured and is left out of the score instead of counting as a zero.
By what a change would be worth, not by what is easiest to list. Each action is ranked on its reach across unique buyer intents, the severity of the gap, how strong the evidence is, how feasible it is for you, divided by the effort it takes. Related fixes are clustered into one campaign — for example, one pricing page that closes eleven missed questions — so you get one primary action, at most two alternatives, and the rest under Later. There is no weekly cap and no total to work through.
There are exactly three plans, all self-serve and all priced in USD: Solo is $99/month or $990/year and covers 1 site, 1 seat, 1,500 monthly credits and no white-label; Agency is $299/month or $2,990/year and covers 5 sites, 5 seats, 3,000 monthly credits and white-label reports; Enterprise is $499/month or $4,990/year and covers 10 sites, 15 seats, 10,000 monthly credits and white-label reports. Every plan includes the full report, all 50 buyer questions across the three engines, weekly measurement, the prioritized actions with their ready-made material, and the weekly email report. The subscription belongs to your account, so one plan covers every site you manage. There is no free plan and no trial: the free public scan is the way to try aisayso.
No. Nobody can promise a position in an AI answer, and we do not measure with a control group, so we never present a change as proof of cause. What we do is show you what changed and when: after you publish a fix we keep measuring the linked questions and report either Visibility change observed after publication or No measurable change yet. The numbers are indicative — they describe a clean, unpersonalised setup, and results in the wild vary per user and per day.
Yes. Agency covers 5 sites with 5 seats and Enterprise covers 10 sites with 15 seats; both include white-label reporting, so the PDF and the weekly email carry your own logo and accent colour. One subscription on your account covers every workspace you manage, site limits are enforced on the server, and client workspaces are strictly separated — data from one client is never visible from another.
A real measurement on 15 of your own buyer questions. No account, no card.
Scan my site free