When a patient, a family or a buyer asks ChatGPT who to call, a few names come back. In your industry, a wrong name or a wrong fact is not a lost lead, it is a liability. Eidomai finds the questions you are losing, shows you the content that would earn them, and puts nothing on your site you did not verify. Every brief follows your industry’s advertising rules.
healthgrades.comPhysical Therapists near Encinitas, CA
movementlab.exampleMovement Lab: Physical Therapy for Athletes
a health publisherRunning injuries: causes and treatment
a directoryBest Physical Therapists in Encinitas
northshorept.comNorthshore Physical Therapy
a review siteEncinitas physical therapy: Top 10
a hospital systemRunner's knee treatment options
12345Next
Ten links. They pick one, or scroll, or go to page two.
→
Now: they ask
Best physical therapist for runners in Encinitas
AI
For a runner in Encinitas, I’d recommend Coastal Motion PT. They specialize in running injuries and gait analysis. Encinitas Sports Rehab and Movement Lab are also well regarded for sports rehab.
Northshore Physical Therapy is not mentioned. There is no page two to find it on.
Three names and a reason. They call one. Sample data
There is no page two.
For twenty years your customers typed a few words into a search box and sorted through a list of links. Now they ask an AI a full question, with their situation in it, and get back a few names. If you are not in the answer, your analytics will never tell you it happened. AI assembles that answer from pages it retrieved while answering. If no page of yours answers the question, you are not in the answer. Whoever published the best answer holds the customer. The engines do not pick a business. They pick a page.
How an AI answer gets built
Assembled from
Passages the engine retrieved while answering, often several sub-queries at once
Who gets cited
Pages that already rank. The AI does not discover you; it reads what Google already trusts
What earns it
A direct answer, your name stated plainly, facts the engine can verify somewhere else
What loses
A month of posts on autopilot. That is where most of the wrong claims in the library came from
Where AI Overview citations come from
Already in the top 2097%
Everything else3%
Semrush, 20,000 blog URLs, 2025. One piece of content has to earn the rank and the citation together.
how an AI answer gets built, and where the citations come from
What you get
The guesswork is gone. The next step is clear.
Getting recommended by AI is a job with a method. Eidomai is the method, the software and, if you want it, the people. Every month, in one place.
questions where an engine names you. A competitor holds 21. Nobody holds 7.
CHATGPT
5/6
PERPLEXITY
3/6
CLAUDE
2/6
GEMINI
3/6
GROK
1/6
AI OVERVIEWS
2/5
AI MODE
1/5
COPILOT
1/6
Lost question
“Best physical therapist for runners in Encinitas”
Perplexity · Sept 9 · answer names Coastal Motion PT and Encinitas Sports Rehab. Northshore not mentioned. Cited: yelp.com, coastalmotionpt.com/runners
Sample dataNorthshore Physical Therapy is fictional.
You see what's being done.
Every month the questions get asked and every answer is saved. Your map shows what shipped, what moved and what to write next.
One system, built for this.
Not five tools and a spreadsheet stitched together. Measuring, deciding, writing, checking and proving, in one place, in one order.
It learns your business.
Eleven questions at the start, and it keeps asking as things change. The map is yours, and it evolves with you.
AI knows what you do now.
Whether AI understands what you do now, not just whether it names you. Named as the business you used to be is its own problem, and it gets its own number.
A growth engine, not a dashboard.
Every lost question becomes a brief, every brief becomes content on your site, and next month's measurement grades it. The content, the facts and the questions you hold build on each other every month.
Good for search, too.
Every piece of content is a real page on your own site, written to answer the exact questions your customers ask, with the facts you confirmed. That's the kind of content search engines favor as well as AI engines. We don't promise rankings in either.
Visibility is not the outcome. A customer is. Every piece of content is measured all the way down to the inquiry in your analytics, reported as a floor. A name in passing is not a recommendation, so mentioned, cited and recommended are scored as three things, because only one of them gets the call. We test the questions that matter to your business and show you the answers we get. Next to them, we report the AI visits and outcomes your analytics can identify.
Illustrative example18 of 46customer questions where AI names you, sample business
Observed client result20%of answers recommend you first, one client, August 2026
Results describe specific measured businesses and are not a guarantee of similar outcomes. See a full sample result.
What each number means
Found
Named in the answer. Every buyer question, across eight engines, monthly. Held, lost, or nobody’s, with the saved answer behind every cell
Chosen
Recommended first, and cited. Mentioned, cited and recommended are three different scores
Understood
Whether the answer describes what you do now or the business you used to be. Scored apart from being named, because a mention in the old category sends the wrong buyer
Counted
Google Analytics joined to the map: AI referral sessions and the conversions they produced, next to the questions and content that drove them
Sampling
Every engine asked more than once per run. API answers approximate the consumer apps; we report trend and share of voice, never absolute truth
found, chosen, counted: what the map measures and how the numbers are scored
How it works
See. Answer. Prove.
Find out which questions you are losing. Publish the content that answers them. Ask again and watch it move.
Ledger · Aug to Sept 2026
Q: Northshore Physical Therapy · movement
+ "Physical therapy after ACL surgery Carlsbad" · Gemini · lost to held · confirmed 5/5 · content shipped Aug 14. + "Does Northshore PT take Medicare?" · Copilot · wrong claim corrected · directory updated Aug 20. = "Best physical therapist for runners in Encinitas" · Perplexity · still held by Coastal Motion PT · brief 14 in progress.
Sample ledger excerpt, fictional client. Every line names the query.
Sample data
01
See
Which customer questions name you, and which name a competitor.
02
Answer
Content on your site that answers the question you are losing.
03
Prove
The same questions, asked again, and what changed.
Run it with your own team on The Map, or have a person at Eidomai run it for you on Done for you. Either way, you approve every word that goes on your site.
See
The software
Asks the eight AI engines the questions your customers ask, every month, and saves every answer word for word.
Who does the rest
On The Map, you read the diagnosis and pick the question worth competing for first. On Done for you, a person at Eidomai reads it and proposes; you approve.
Answer
The software
Builds the brief from facts you confirmed, in your voice, shaped the way the engines lift from, and checks the draft against those facts before it ships.
Who does the rest
On The Map, you or your writer write from the brief and publish. On Done for you, Eidomai writes, checks, publishes and places it; you approve every word.
Prove
The software
Asks the same questions next month, every engine, and lines the answers up against last month.
Who does the rest
On The Map, you read what moved and pick the next question. On Done for you, a person at Eidomai reads it, tells you plainly, and picks the next one. Then around again.
Nobody controls what the engines say. What you control is what they find when they look: content that answers the question, built from checked facts. That is what moves the answer, and the proof is the same question asked again. A change after the work is evidence to read, not proof that one piece of content caused it.
Eidomai treats each buyer question worth working on as a piece of work: usually new content on your site, sometimes a fix to existing content or to the fact behind a wrong answer, built from facts you confirmed and checked before it ships (by the software on The Map, by a person as well on Done for you). The measurement says which question. The diagnosis says who holds it and why. The brief carries your verified facts, your voice and what the winning source said, in front of the draft, every time. The content goes live. Eight engines are asked again. Watch one question go through.
What goes in the brief
Your facts
The fact sheet you confirmed at intake. Nothing outside it gets claimed
The holder
What the page that beat you said, and the sources it holds the answer with
Your voice
How you sound, and what you never say
The shape
Direct answer first, every target question covered, your name as the entity
The check
Facts against the sheet, a reads-human score, a can-an-engine-lift-this score, then a person
01The question they ask
QBest physical therapist for runners in Encinitas
QPhysical therapy after ACL surgery, do I need a referral?
QDoes Northshore PT take Medicare?
QDry needling near Carlsbad
Twelve to twenty buyer questions, drafted from your market, confirmed by you, run across eight engines every month. The map says which ones you are losing, and to whom.
briefed
02The brief, built from your content
brief · Best physical therapist for runner…ready
Holds it todayCoastal Rehab, a 2021 blog post: "running injuries we treat"
From the mapWho holds this answer today, and which page taught the engines
From your fact fileVerified facts only, provenance stamped (3 for this piece)
From your voice rulesHow you sound, what you never claim
From how engines readDirect answer first, one question, self-contained, schema
An engine has no memory of your business. It starts from zero every time. So the brief carries everything: your facts, your voice, what the winning page said, the shape engines lift from. The words come out right because what went in was.
written
03The content that answers it
northshorept.com
Running injury rehab and gait analysis in Encinitas and Carlsbad
The answer in the first paragraph. Your name as the entity. Every claim traceable to a fact you confirmed.
Two clinicians certified in running gait analysis
Return-to-run protocol, staged by benchmark
Encinitas and Carlsbad locations, both stated
Facts3/3Reads human91ShippedSept 14
Written from the brief by you or by us, checked against the facts before it ships, published on your site. One piece of content built to earn the Google rank and the AI citation together, because the engines cite what already ranks.
measured
04Who names you now
ChatGPT
Perplexity
Gemini
Claude
AI Overviews
AI Mode
Copilot
Grok
Before1/8
→
After6/8
Banked for next month+3 facts in the fileone held question, one ledger line, one piece of content still working
Same question, same eight engines, next month. One piece of content can move several of them, because every engine reads the same map. The ledger records what changed after it went up, with the query named. Then the next question.
A tool resets every month. Published content does not. Each piece that ships adds a held question, facts the engines now repeat back, and a line in the ledger. The next brief starts from all of it, and it is all on your site.
01 MEASUREWhich questions you are losing
Your buyer questions, run across eight engines, several times, with every answer saved. The map shows who holds each answer.
02 DIAGNOSEWhy they hold it
Which page taught the engines, which facts about you are wrong or missing, and where your own descriptions contradict each other.
03 BRIEFThe brief, built from your content
What the winning page said, your verified facts, your voice, the shape engines lift from. Loaded in front of the draft, every time.
04 WRITE + PUBLISHThe content goes live
Written from the brief by you or by us, checked against the facts, approved, published. This is the stage that moves the answer.
05 RE-MEASUREThe ledger grades the content
Same questions, same engines, next month. What moved, with the query named. New facts banked. Then the next piece of content.
When published content shows up in the answers
60 to 90 days
Content published
Indexed
Flips start
Confirmed
Content published
Indexed
Flips start
Confirmed
Single-month moves are noise. Flips take 60 to 90 days after indexing, and a win only counts once it holds across the deeper sample. Trend, never truth.
All eight stages, published in full: How it works.
who does each step on each plan, the brief, one question going through the engine, and why it compounds
Built for compliance-sensitive industries
Content that knows your industry’s rules.
Health companies can’t promise cures. Therapists can’t promise outcomes. Wineries can’t make health claims. Advisers can’t show performance without the fine print. Tell us your industry, and every brief, prompt and draft check follows its advertising rules, plus any lines your own compliance team has drawn. Your reviewer still approves every piece of content. They just start from a draft that’s already inside the lines.
FTC truth in advertisingFTC and FDA health claimsHIPAAAPA, ACA, NASW, AAMFT codesTTBABA Model Rules 7.1 to 7.3SEC Marketing RuleFINRA Rule 2210Fair Housing Act
Draft checkRunning injury and gait program page
Facts3/3
Reads human91
LiftableYes
ShippedSept 14
Both scores are advisory. A person still decides.
Sample dataNorthshore Physical Therapy is fictional.
Your rules, in every brief
Each brief lists what the content can’t say, what needs proof, and what it must include. The prompt gives the same rules to whatever AI writes the draft.
Every sentence checked
Draft check flags the sentence, names the rule it hits and suggests a rewrite.
Your own rules come first
Paste the lines your compliance team never lets you say. They apply to your account and take priority over ours.
You approve every word
On every plan. Nothing goes live until your reviewer signs off, and each piece of content records who approved it, when, and which rules it was checked against.
No patient or client records
Eidomai reads public AI answers about your business. It never asks for patient or client records, and it stops you if one gets pasted in.
Every plan follows your industry’s advertising rules: health and medical; therapy and mental health; wine, beer and spirits; law; investment advice; securities brokerage; real estate; and hospice and home care. Every other business gets the FTC’s truth-in-advertising rules. Add your own, and they come first.
We save what each engine actually said about you, word for word.
Real competitors
See who gets recommended when you do not, and why.
Real work
For the questions worth working on, a next step built from facts you confirmed: improve existing content, add missing content, or correct the fact behind a wrong answer. Related questions are handled together, not one piece each.
Real proof
Same questions, same engines, next month. What moved, and what did not.
ChatGPT, Perplexity, Gemini, Claude, Copilot, Grok, AI Overviews and AI Mode were built by rivals on different data, and they still land on the same names, the same sources and the same mistakes. Your business is one region in one shared map, and every engine reads from approximately the same map. When a Nimbletoad client improved, it improved across engines together. When the engines got a client wrong, the error showed up everywhere too. So Eidomai measures engine by engine, because each retrieves differently, and fixes the map once. The inverse compounds too: a stale fact in a source the engines trust blurs you everywhere at once, which is why every claim an engine makes about you is audited monthly and corrected at the page that taught it (the accuracy layer).
BlurredThree descriptions that disagree, a stale fact in a source the engines trust, a category nobody can place. Five of eight engines reach past you for the competitor whose region is crisp.
SharpOne name, one set of facts, content that answers real questions in citable form, the same facts repeated by sources that are not you. Every engine reads the same map, so every engine reads the same facts about you.
What we saw across the client book
Observed
When a client improved, it improved across engines together. When the engines got a client wrong, the error showed up everywhere
So
Measure engine by engine, because each retrieves differently. Fix the map once
The inverse
A stale fact in a source the engines trust blurs you everywhere at once
Where it shows
The same sources, the same names and the same errors surface on every engine we ask, month after month
real answers, real competitors, real work, real proof, and why one piece of content moves eight engines
Your results open on this screen with your own business on it.
“Physical therapy after ACL surgery, North County”
ChatGPTYou are named. 8 of 8 engines name Northshore.
Northshore Physical Therapy in Encinitas runs a post-surgical ACL program with return-to-sport testing. Coastal Motion PT and Encinitas Sports Rehab also offer rehab...
“Best physical therapist for runners in Encinitas”
PerplexityCoastal Motion PT is named. 2 of 8 engines name Northshore.
For runners, Coastal Motion PT is frequently recommended for gait analysis and return-to-running plans. Movement Lab is another option in Carlsbad...
“PT for plantar fasciitis without a referral”
GeminiNobody is named consistently. Open territory.
In California you can see a physical therapist without a referral for a limited period. Look for a clinic that treats foot and ankle conditions...
Who each AI engine names for each of your buyer questions. The gauge up front is how many of the eight name you. In the cells, a green dot means you are named, a red dot names the competitor holding that answer today, and a hollow dot means nobody is named consistently yet.
Who holds the answers you are not winningCoastal Motion PT 9Movement Lab 6Encinitas Sports Rehab 4nobody yet 27
You
Buyer question
ChatGPT
Perplexity
Gemini
Claude
Google AIO
AI Mode
Copilot
Grok
8/8
Physical therapy after ACL surgery, North County
You
You
7/8You
You
You
You
You
You
7/8
Sports rehab clinic open Saturdays, North County
You
Coastal Motion
You
You
You
6/8You
You
You
5/8
Does Northshore PT take Medicare?
You
You
wrong claim
wrong claim
Open
You
You
You
5/8
Physical therapy for shoulder impingement Carlsbad
You
You
Open
You
Open
You
Movement Lab
You
2/8
Best physical therapist for runners in Encinitas
Coastal Motion
Coastal Motion
5/8Coastal Motion
You
Coastal Motion
Open
You
Open
2/8
Dry needling near Carlsbad
Movement Lab
Open
Movement Lab
You
4/8Movement Lab
Open
You
Movement Lab
1/8
Pelvic floor physical therapy Encinitas
Encinitas SR
Encinitas SR
Open
Open
Encinitas SR
You
Open
3/8Encinitas SR
1/8
Rotator cuff rehab Carlsbad
Coastal Motion
Coastal Motion
Open
Coastal Motion
Open
Coastal Motion
You
Open
1/8
Cash pay physical therapy North County, is it worth it?
Open
Movement Lab
Open
Open
Open
You
Open
Open
0/8
PT for plantar fasciitis without a referral
Open
Open
Open
Open
Open
Open
Open
Open
10 of 46 questions shown · sample data
The portal, on the fictional Northshore Physical Therapy set. Sample data.
Who each AI engine names for each of your buyer questions. The gauge up front is how many of the eight name you. In the cells, a green dot means you are named, a red dot names the competitor holding that answer today, and a hollow dot means nobody is named consistently yet.
Who holds the answers you are not winningCoastal Motion PT 9Movement Lab 6Encinitas Sports Rehab 4nobody yet 27
You
Buyer question
ChatGPT
Perplexity
Gemini
Claude
Google AIO
AI Mode
Copilot
Grok
8/8
Physical therapy after ACL surgery, North County
You
You
7/8You
You
You
You
You
You
7/8
Sports rehab clinic open Saturdays, North County
You
Coastal Motion
You
You
You
6/8You
You
You
5/8
Does Northshore PT take Medicare?
You
You
wrong claim
wrong claim
Open
You
You
You
5/8
Physical therapy for shoulder impingement Carlsbad
You
You
Open
You
Open
You
Movement Lab
You
2/8
Best physical therapist for runners in Encinitas
Coastal Motion
Coastal Motion
5/8Coastal Motion
You
Coastal Motion
Open
You
Open
2/8
Dry needling near Carlsbad
Movement Lab
Open
Movement Lab
You
4/8Movement Lab
Open
You
Movement Lab
1/8
Pelvic floor physical therapy Encinitas
Encinitas SR
Encinitas SR
Open
Open
Encinitas SR
You
Open
3/8Encinitas SR
1/8
Rotator cuff rehab Carlsbad
Coastal Motion
Coastal Motion
Open
Coastal Motion
Open
Coastal Motion
You
Open
1/8
Cash pay physical therapy North County, is it worth it?
Open
Movement Lab
Open
Open
Open
You
Open
Open
0/8
PT for plantar fasciitis without a referral
Open
Open
Open
Open
Open
Open
Open
Open
10 of 46 questions shown · sample data
The portal, on the fictional Northshore Physical Therapy set. Sample data.
The engines we ask, every month
ChatGPTPerplexityClaudeGeminiGrokGoogle AI OverviewsGoogle AI ModeCopilot
Free check results open on this screen, with the business's own data on it. Clients live in it: the map, the briefs, the draft check that scores content against the facts before it ships, the ledger, the analytics join.
What is on the portal
Battle map
Questions down the side, engines across the top, the saved answer one click from every cell
Accuracy
Every claim the engines make about you, judged against your fact sheet, traced to its source
Briefs
One per lost question: who holds it, why, and the content built to compete for it
Draft check
Facts, reads-human, liftable. Advisory scores; a person decides
Ledger
What moved after the work went live, confirmed or provisional. Evidence to read, not proof that one piece of content caused it
Methodology card
Panel version, market, sampling depth, scoring rule, date. On every map
what is on the portal
Works in Claude and ChatGPT
Ask Claude or ChatGPT what to write. It answers from your Eidomai account.
Connect once, and the AI you already use reads your work list, writes from your verified facts, and runs the draft check before anything ships. Part of The Map and Done for you.
Claude or ChatGPTEidomai connected
What should I write this month?
Reading your Eidomai work list
Start with “Best physical therapist for runners in Encinitas.” Two competitors hold it today and nothing on your site answers it.
I drafted the page from your verified facts and left one gap as [CLIENT TO CONFIRM: Saturday hours].
Draft check: blocked until you confirm the hours
Sample dataNorthshore Physical Therapy is fictional.
01
Connect once
Add Eidomai to Claude or ChatGPT and sign in. You choose read only, or read and write.
02
Ask what to write
It opens your work list and names the question worth winning next.
03
It drafts from your facts
Only what you have verified. Anything else comes back as a placeholder for you to confirm, never a guess.
04
The check clears it
The draft check runs before anything ships. A blocked draft stays blocked.
Month to month, or yearly with 2 months free. No contract.
The Scan
Want to start small?$79 $66 a month, billed yearly at $790 · save $158: your map, measured fresh every month, and one brief a month for a question you're losing: what to write, and the prompt to write it, built to your industry's rules. You write the content. We also flag it when an AI engine says something about you your industry doesn't allow. All eight engines. Move up to The Map any time. Start your free trial.
Done for you (We Cook): what is in it
01
A person at Eidomai who does the work for you
02
Twelve to twenty buyer questions, confirmed by you
03
Eight engines, every month, every answer saved
04
Accuracy watchdog weekly, corrections sent to the source
05
Entity accuracy and topic ownership: whether AI knows what you do now, and which topics you own
06
The prompt for every lost question, and a person who checks the draft
07
Six pieces of content a month: written, checked, published, placed
08
Citation outreach, the ledger, the portal
09
Connect to Claude and ChatGPT: ask what to write, and your AI answers from your account
The Map (You Cook): what is in it
01
The same twelve to twenty questions, confirmed by you
02
Eight engines, every month, every answer saved
03
Battle map: who holds each answer today
04
Monthly accuracy audit, wrong claims traced to the source
05
Entity accuracy and topic ownership: whether AI knows what you do now, and which topics you own
06
The context ledger: an thirteen-question interview that becomes your fact sheet
07
The prompt for every lost question: who holds it, what earns it, your verified facts, your voice. Copy it, write the content, mark it shipped
08
The ledger, the movement report, the portal
09
Connect to Claude and ChatGPT: ask what to write, and your AI answers from your account
Locked
A person at Eidomai who does the work for you
Locked
Check your draft: a person reads the content against the facts before it ships
Locked
Content written, published and placed
Locked
Weekly watchdog and corrections
The Scan: what is in it
01
The monthly scan on all eight engines
02
The battle map
03
Accuracy flags for your industry's rules
04
One brief and prompt a month for a question you're losing
Everything else opens on The Map.
Done for you is a person on your account, with the software underneath: we write, check and publish, and you approve every word. The Map is the same system, and you do the writing. The Scan is the place to start if you want to see your map first. The cookbook was never the hard part. The cook always was: The Cookbook Problem, on the Nimbletoad Insights.
everything in each plan, and what The Map leaves locked
Find out what AI says about you.
Free check on ChatGPT and Claude, results by email, no call required.
Free AI check →Free. No call required. You keep the results.
It starts with an eye. It ends with AI. Eidomai, by Nimbletoad.