Roadmap
What we’re building next
What we are working on, what we are weighing up, and what we have decided not to build. There are no dates on this page, deliberately — we would rather tell you what is true than what sounds committed.
In progress
See what an engine knows about you without searching
In progressEvery engine indextr measures runs with live web search on, because that reproduces what a real person gets. But models also carry knowledge from training, and the two can disagree sharply. We tested this: asked with search switched off, Gemini knew one business cold while ChatGPT had never heard of it. Reporting the split explains results that otherwise look contradictory — being named often but cited never usually means the model remembers your industry, not your website. It is a diagnostic rather than a lever, and we will say so: nothing you do this quarter changes what a model already learned.
Run a scan before creating an account
In progressToday you sign up, then see your score. It should be the other way round — enter your business, watch the scan, and decide afterwards whether to keep it.
Up next
More detail on each competitor
Up nextRight now a tracked competitor shows a mention rate. We already record which engines named them, how they were described, and where they appeared in the answer — it just isn't shown yet. Seeing that a rival owns ChatGPT while you own Google AI Overview is a different problem from losing everywhere.
The actual words each engine used
Up nextYou see that sentiment was negative. You cannot yet see the sentence. The raw response is stored for every search and engine, and reading it is far more useful than a label — it usually shows exactly which page or review the engine drew from.
A record of what you changed, and when
Up nextMark a recommendation done and it should appear on your trend chart, so a score movement can be read against the work that preceded it. Plus free-text notes, for the things we never recommended — claiming a listing, asking customers for reviews.
Choose how often we email you
Up nextOne weekly report suits most people and is too much for some. A setting for every scan, a monthly summary, or nothing at all.
Considering
Measured time-to-impact, from real results
ConsideringEvery timing estimate in indextr — check back in two to four weeks, this is a slow change — comes from how often crawlers revisit pages. None of it is measured, and we say so wherever it appears. But we hold both halves of the data: when a customer marks work done, and how their score moved afterwards. Across enough customers that becomes evidence rather than an estimate. It needs real volume before it would be honest, so it waits.
One account, several businesses
ConsideringFor agencies, and for owners running more than one brand. Today each account tracks one business, which means separate logins for people who should have one.
Separating where you are from where you sell
ConsideringAn installer in a small town sells to the whole metro, and buyers search the metro. But a Google Business Profile is registered at the physical address. One field is currently doing both jobs, which is fine until the two differ.
Decided against, and why
Things you might reasonably expect, that we have chosen not to do. The reasons matter more than the decisions.
Microsoft Copilot, for now
We tracked Copilot until July 2026 and then removed it, because the code labelled Copilot was calling an OpenAI model. It had never been measuring Copilot at all, and a number attributed to the wrong engine is worse than no number. We have since confirmed a way to read genuine Bing Copilot answers, so this is a question of when rather than whether. Two things have to be settled first: telling a Copilot answer apart from an ordinary Bing search result, and knowing enough about real usage to weight it honestly against the other five.
Claiming to cover the whole AI market
We track five engines. The percentages you see beside them are their weights relative to each other, not shares of every AI assistant in existence — Copilot is not in them, and neither is Grok or anything newer. We would rather tell you exactly what was measured than round it up to everything.
Recommending schema markup for AI visibility
We checked all five engines in August 2026 and not one documents a requirement for it. Google states outright that there is no special structured data you need to add. So we don't recommend adding it — and we never recommend removing what you already have, because it does real work for Google's rich results. A lower priority than it once was, not a mistake to undo.
Guaranteeing how long a change takes to show up
No engine publishes a recrawl schedule, and no vendor will tell you when your page is next read. Our estimates are labelled as estimates everywhere they appear, and they will stay that way until we can replace them with measurements.
Tell us what you need
Most of the list above came from someone asking. If indextr is missing something you would use, this reaches us directly.