You can see you lost
Plenty of tools count how often a competitor gets named. Few show what the assistant read that made it pick them.
GEO for agencies
Verdict puts your clients' buyer questions to ChatGPT, Claude, Gemini, Grok, Perplexity and Google AI Overviews every week. You see who gets named instead, the pages each answer quoted, and a fix list ranked by what will change the answer.
14-day free trial after your demo. No card needed.
Northwind CRMHealthy
northwindcrm.com · United States
38.2%+7.5 pts
Named in 73 of 192 answers from 6 assistants. 2nd of 5 tracked companies on share of voice, usually named second.
Share of AI answers to the tracked prompts that name this client. Change is against the previous period of the same length.
One line per assistant. Dashed markers are fixes marked done.
What moved in the answers since the check before.
Most likely to get this client named more often.
Share of voice among tracked companies.
Prompts whose visibility changed most since the previous check.
Buyers ask an AI assistant before they ask you. To change its answer you need to know what it read, which of that you can influence, and whether your changes worked.
Plenty of tools count how often a competitor gets named. Few show what the assistant read that made it pick them.
Another comparison page won't help if the answer was swayed by a review badge, a forum thread or a video.
Assistants lean on review sites, Reddit, Wikipedia, YouTube and roundups. You don't own those pages, but you can change what they say.
A screenshot of one ChatGPT answer proves nothing. A weekly trend across six assistants, with every fix marked on it, does.
The evidence
Tracked question
Named first instead: CareLoop
ChatGPT
#3
AI Overviews
#1
Page the answer quoted
g2.com
Quoted 4 times. Backs CareLoop, not your client.
Page the answer quoted
reddit.com
Two practitioner threads. Back CareLoop and MedDesk.
Page the answer quoted
northwindcrm.com
Your client's compliance page. Quoted once, and it backs them.
Share of voice
Next fix
Get your client reviewed on g2.com, and ask happy customers for fresh reviews
Off-site · high impact · 3 prompts
We set up your first client with you on the demo call. After that, checks run every week on their own.
Their website, the competitors that matter, and the questions their buyers ask. Verdict suggests questions to start with.
Every question goes to six assistants each week, from the client's country. We record who's named, in what order, and which pages were quoted.
Gaps become fixes, on the website and off it, worst first. Start one, mark it done, leave a note for the team.
Visibility over time with a marker for every fix you shipped, and a report your client can open from a link.
ChatGPT, Claude, Gemini, Grok, Perplexity and Google AI Overviews, each with web search on.
Checks run for the client's market, in 16 countries, so a London firm isn't judged on American answers.
Answers wobble from one asking to the next. We confirm a change before we report it, so alerts mean something.
A company counts once per answer, at the position a reader first meets it.
Every plan has all six assistants and every feature. Plans differ in how many clients and prompts you track.
$149a month
3 clients · 25 tracked prompts · 2 seats
$399a month
15 clients · 150 tracked prompts · 5 seats
$999a month
50 clients · 500 tracked prompts · Unlimited seats
14-day free trial after your demo. No card needed. Two months free on annual plans.
GEO, generative engine optimization, is the work of getting a brand named when people ask AI assistants for recommendations. It's SEO's sibling: instead of ranking a page in a list of links, you're trying to be one of the three or four names in the answer.
Website changes usually show up within a few weeks. Anything that depends on other people, like reviews, forum threads and third-party articles, takes one to two months. Mark a fix done and the trend shows what happened after.
No, and that's the point. AI answers lean on review sites, Reddit, YouTube, Wikipedia and industry roundups. Verdict separates what you can fix on the site from what lives elsewhere, because the second group often decides the answer.
AI answers vary from one run to the next. A weekly check across six assistants, with every change confirmed before we report it, gives a steadier trend than a noisy daily one, and it's what keeps six assistants affordable on every plan.
Yes, 14 days, no card. We set it up with you after the demo, with your first client and their questions. Fourteen days means two weekly checks, so you see a trend, not a single snapshot.
Bring one client and the questions their buyers ask. On a 30-minute call we'll show you what six AI assistants say about them, and what we'd fix first.