Every client's AI reputation, in one triage list.
You own dozens of retainers, and AI is quietly reshaping how each client is recommended. Open one screen that answers a single question in five seconds: which account do I touch today, and what do I say to them.
Or start free and add clients as you go.
Asked the way your audience asks
Diagnose
one root cause
Correct
from your facts
Prove
a crawler fetched it
The problem
A grid of scores doesn't tell an account manager what to do.
AI search is the new ORM, and it now runs against every client you hold — the recommendation, the price, the founder's name, the damaging post, all decided in one AI answer none of you can see. A number of 61 is not an action. ‘Their comparison table is JavaScript-only’ is. Across a portfolio, you need the causes ranked by consequence, the wins queued for approval, and a way to prove what you moved — without stitching five tools together or logging into each client.
In the wild
Moments this earns its keep
The pitch you walk in already winning
Run a prospect domain the night before and open with what the AI engines actually say about them today, ranked by what it costs. Instead of a capabilities slide, you hand them a diagnosed list of fixable losses.
The client who calls at 9am
A founder forwards a screenshot of an AI assistant recommending a competitor and wants answers by lunch. You pull their triage list, see the root cause already ranked, and reply with the reason and a queued correction — not a promise to look into it.
The monthly report that writes itself
Reporting week no longer means exporting from five tools. Each account already carries what moved, which corrections shipped, and the before-and-after in the answers — so the retainer renews on evidence, not activity.
The junior who sounds senior
A new hire opens the triage list and sees the top item already stated as a cause — a comparison table rendered only in JavaScript — with a queued fix. They send a credible, specific update on day one.
The client whose CEO name AI gets wrong
A client forwards an AI assistant naming a founder who left, or quoting a price they retired. You run a fact-accuracy check, see which of the six engines is wrong and why, and hand them a dated correction pack to publish — a billable deliverable, proven on re-measure.
The AI answer quoting a hit piece
A client wants a damaging claim gone by Friday. You open their triage list, see the exact source already flagged, and reply with both levers ready — a white-label counter and a report-request draft — as a billable reputation-defense line, proven on re-measure.
The mandate a brand-only tool can't take
The prospect isn't a company — it's a sitting minister, or a state department. On a brand-only GEO tool you'd have to pass. Here you open the same book in defend-the-record or accuracy-and-audit mode, and win a retainer your competitors literally can't service.
How it works for you
Find the one cause. Fix it from your facts. Prove it landed.
- 1
Triage
A consequence-ranked list of three to five items per client, each stating the cause — not just the symptom. At least one row will say ‘do nothing’, which is what makes the rest trustworthy.
- 2
Diagnose
Each item carries its root cause and evidence, so an account manager knows the fix on sight and can justify the order to the client.
- 3
Approve
One approval inbox spanning every client, each with its own tone and sign-off trail. Nothing publishes without a human.
- 4
Prove
Forecast and actual on the same axes, per client — including where the forecast was wrong. That honesty is what keeps the account.
The substance
What the portfolio view actually does
Portfolio triage ranked by consequence, not alphabet
One screen, every client, ranked by what each issue is costing the brand — not by score or name. A buying-intent prompt losing to a competitor outranks a cosmetic tone issue automatically, so you spend billable hours where an hour moves the most, and can defend the call to anyone who asks.
A root cause you can say out loud to a client
A grid of scores tells a client they're losing without telling them why — every review turns defensive. We isolate the one reason an answer is wrong (say, a comparison table rendered only in JavaScript no crawler reads) — a sentence you can put in an email and an engineer can action.
One approval inbox, tuned per client
Corrections across every account queue into one inbox — you review, edit and approve before anything ships. Each client keeps its own voice and sign-off, and nothing goes live without a human release, so you hold editorial control across the whole book without living inside each account.
White-label so the work looks like yours
Dashboards, exports and client reports carry your identity, not ours. The prospect in your pitch and the founder reading your monthly see your agency as the source — you resell the outcome as a first-party service line, not a tool marked up.
Pitch mode on a domain that isn't yet a client
Point it at any prospect domain and get the same diagnosed triage list you give paying clients — before a contract exists. You walk in with their real exposure across the engines and a short list of losses you can fix, which reframes the meeting from selling hours to closing a gap.
Forecast against actual, honestly
When a correction ships, we project the expected lift, then measure what actually moved in the answers — and show both. You're never asked to trust a number that only points up, which is what lets you put a forecast in front of a client and keep your credibility at the next review.
A new billable line, not another cost centre
AI-search reputation is a service most clients now want and few agencies can price with confidence. Consequence-ranked work each day plus forecast-vs-actual lets you scope hours to real volume and defend the retainer with before-and-after proof in the answers themselves — margin, not overhead.
Fact-accuracy across every client, as a portfolio capability
Beyond share-of-answer, run a fact-accuracy check on each client — price, founder, CEO, HQ, funding, product names — graded across all six engines as correct, outdated or wrong. The worst-fidelity accounts surface at the top of the same triage list, so the client whose price or CEO AI states wrong gets worked first — the same measure-diagnose-correct-prove loop, extended to 'is AI getting their facts right'.
A deliverable you can name, price and prove
"We make sure AI states your facts correctly" is a service line clients understand and will pay for. For each misstated fact we compute one root cause and generate a dated correction pack — JSON-LD, an HTML fact block, an llms.txt entry — that you publish under your white-label, then re-measure to prove the fact now states right. A billable outcome with evidence attached, not an audit you give away.
Capabilities
Built for how you work
Portfolio triage home
Which account to touch today and why, ranked by consequence across every retainer.
Every kind of client, one book
Not just brands. Track corporates, public figures and leaders, ministries and public bodies in the same portfolio — each with its own purpose-built product: visibility and share-of-answer for a brand, defend-the-record for a leader, accuracy-and-audit for an institution. No other GEO tool lets one practice serve all of them.
White-label
Your domain, your logo — nothing the client sees mentions us.
Pitch mode
Run a full panel on a prospect's domain before the first meeting, and walk in with the diagnosis.
One approval inbox
Every client's queue in one place, each with its own sign-off trail.
Forecast versus actual, per client
Project the lift a correction should produce, then show what actually moved in the answers — including where the forecast missed, which is what keeps a renewal honest.
Six engines, no client logins
ChatGPT, Gemini, Claude, Perplexity, Grok and Google AI Overviews watched daily for every account, rolled into one view without logging into a single client tool.
Fact-accuracy across the portfolio
Run a fact-accuracy check on every client brand — price, leadership, HQ, funding, product names — graded correct, outdated or wrong across all six engines, so you see which clients AI is describing incorrectly, not just who's under-recommended.
A new billable deliverable
"We make sure AI states your facts correctly" — a service line you can price, with a computed root cause and a copy-paste correction pack per wrong fact handed to the client as evidence, then re-measured to prove it landed.
Worst-fidelity clients first
The portfolio triage surfaces the accounts with the most misstated facts at the top, so you fix the client whose price or CEO AI is getting wrong before the one with a cosmetic gap.
Reputation defense across the portfolio
For every client, surface the damaging claims and fake reviews AI is repeating and the exact sources feeding them, then queue two levers per finding into the same approval inbox — a counter drafted from the client's own record and a report/takedown-request draft. A reputation-defense line you can price and prove, not just monitoring.
Inside the team
Who on the team lives in it
Account manager
Opens the triage list each morning, works the top of the queue across their book, and answers client questions with a cause and a queued fix instead of a promise to investigate.
Agency owner or principal
Reads the whole portfolio in one glance to see which retainers are at risk, which are proving value, and where to point capacity, then defends the service line in renewals with evidence.
Strategist
Uses root causes and forecast-versus-actual to decide which corrections are worth the effort per account and to shape the quarter's roadmap around the losses that actually cost the client.
New-business lead
Runs prospect domains before pitches to walk in with diagnosed exposure, turning cold meetings into fixable-problem conversations that close on proof rather than credentials.
Content and production team
Takes the named root cause and the generated correction — comparison page, FAQ schema, llms.txt, or a dated fact-block for a wrong price or CEO name — and ships it to the client's site, spending hours on the fixes that move an answer instead of on speculative content.
The takeaway
What it changes for the agency
FAQ
Questions, answered
How do we package this as a retainer?
Most agencies fold it into an existing monthly retainer as a distinct AI-search reputation line, or sell it as a standalone GEO retainer priced on the number of client brands and the pace of corrections. Because the platform surfaces consequence-ranked work each day, you can scope hours to the triage volume rather than guessing, and the forecast-versus-actual record gives you the renewal justification.
Can we sell 'make sure AI states our facts correctly' as its own deliverable?
Yes, and it's a strong one. For every client you define their key facts — price, founder and CEO, HQ, funding, product names — as a record they approve, then measure how correctly all six engines state each one. Each wrong fact carries a computed root cause and a copy-paste correction pack (JSON-LD, HTML fact block, llms.txt, dated), which you publish under your white-label and re-measure to prove. It sits alongside share-of-answer in the same triage list, so it's a priced, provable line item rather than a favour, and the portfolio view floats the worst-fidelity clients to the top.
How does fact-accuracy work across dozens of clients at once?
It rolls into the same portfolio triage. Each client's misstated facts are ranked by consequence next to their visibility issues, so a client whose current price or CEO an engine is getting wrong outranks one with a cosmetic gap. You point the platform at a client's domain, it fetches the page a crawler's way to compute why a fact is wrong — not machine-readable, not indexed, outranked by a G2 or Wikipedia page, or just stale in one engine — and you triage the whole book from one screen without logging into a single client tool.
Is it genuinely white-label?
Yes. Client-facing dashboards, exports, and reports carry your agency's identity, and the prospect or founder never sees our brand as the source. You present the intelligence as a first-party capability of your agency.
How many client brands can we manage at once?
The portfolio view is built for dozens of retainers on one screen, and there is no practical ceiling that forces you into separate logins. Accounts roll up into a single consequence-ranked triage list, so adding a client adds a row, not another tool to check.
What does onboarding a new client look like?
You point the platform at the client's domain and it begins measuring how the AI engines describe and recommend them across the major engines, then surfaces the first ranked root causes without a lengthy setup. There is no tag to install or data pipeline to wire before you see value, which is why the same flow doubles as your pitch tool on prospect domains.
Do we keep control over what gets published?
Always. Every correction queues into an approval inbox and nothing ships until a human on your team releases it, with per-client tone and sign-off preserved. You retain full editorial control across the whole portfolio while still moving faster than logging into each account by hand.
Can we mix our own brands with client brands in one portfolio?
Yes. A brand is a brand whether it is a client or your own agency, so you can track your own AI reputation in the same triage list you use for clients. Many agencies run their own domain through it first to learn the loop, then add client accounts as rows without changing tools.
What do we hand a client who asks to see the raw evidence?
Every finding carries the verbatim engine answers, the sources each engine cited, and the before-and-after once a correction lands, all under your white-label. You can share the diagnosed list and the proof directly, so a sceptical client sees exactly what the AI assistants said and exactly what changed rather than a summary they have to trust.
Can we manage public figures and government clients, not just brands?
Yes — and this is the difference. A politician isn't a brand and a ministry isn't a company, so each client gets its own purpose-built product rather than a relabelled brand dashboard. A public figure gets defend-the-record — correct the facts from their own statements, catch fabricated audio, video and claims, and act one decision at a time. An institution gets accuracy-and-audit — a verified record, per-language accuracy checks, dated machine-readable corrections and an audit trail. A brand gets visibility and share-of-answer. They all live in the same white-label portfolio and one approval inbox, so a single comms practice can run a FMCG brand, a sitting minister and a government department from one book — a book no brand-only GEO tool can service.
Every client's AI reputation, in one triage list.
Or start free and add clients as you go.
Not sure which fits? See every kind of reputation we work on.