AEO for Doctors | Progression Agency

Updated September 2026 · Written and maintained by the Progression Agency strategy team

Clinical questions are filtered toward institutional sources by design. The administrative questions patients ask constantly are not — and almost nobody answers them properly.

On this page · 9 sections

Medical questions are filtered before an answer is assembled
The second entity record, and why it is the weaker one
Administrative questions, answered properly
Measuring this without pretending it is precise
Where medical answers come from, by question type
Multi-location and multi-clinician practices
What we would not sell you
What an engagement looks like from your side
Everything else we have written on search, AI and getting found

The short answerMedical questions are filtered before an answer is assembled: the source set narrows toward government health bodies and major medical publishers, and a consult-a-professional line is attached regardless. A single practice will not be cited there. What is open is the administrative layer — accepting new patients, named insurance networks, referral requirements, wait times, telehealth, hospital affiliation. Alongside that sits a second entity record you did not write: your entry in every health-plan directory and hospital roster, structured, easy for systems to consume, and consistently the less accurate of the two.

Filter behaviour reflects our own observation of assistant responses to health queries and is not a published specification. Anything touching clinical claims belongs in your own compliance review.

What it costs: AEO for Doctors

Answer engine optimization is priced like SEO, because most of the work overlaps with it. For AEO for Doctors, the ranges below are the published US bands we quote against; the number for a specific site is driven by rendering, template count and how much of the work already sits inside an SEO engagement.

AEO planning ranges (US figures)
Engagement	Typical range	What it covers
AEO audit	$1,000–$4,000 one-off	Eight checks, roughly 10–20 hours, with a prioritized fix list
AEO added to an existing SEO retainer	A few hours of setup, roughly $400	Schema, extractable answers, entity clean-up on pages already being worked
Standalone AEO retainer	$1,500–$20,000 / month	Content restructuring, schema, citations and monitoring across a site
Technical remediation for extractability	$1,500–$6,000 one-off	Rendering, template and structured-data fixes

Ranges are US planning figures, not quotes. Every engagement is priced after a written scope, and the planning range tells you which tier the conversation starts in.

Medical questions are filtered before an answer is assembled

A safety posture narrows the source set for clinical questions before your material is considered at all, and no amount of content routes around it.

The filter is the fact everything else follows from

When someone asks an assistant about a symptom, a medication or a treatment, the answer is not assembled the way an answer about accounting software is. A safety posture engages: the source set narrows sharply toward government health bodies, large hospital systems and established medical reference publishers, and a consult-a-professional line is attached regardless of what was cited. This is deliberate and it is not going to be relaxed. Any plan that depends on a single practice being cited for clinical information is working against that design rather than with it.

What that leaves, and why it is better than it sounds

The filter applies to clinical content, not to facts about your practice. Whether you are accepting new patients, which insurance networks you participate in, whether a referral is required, how long the wait is, whether you offer telehealth, which hospital you admit to — none of that is safety-sensitive, and all of it is asked constantly. It is also answered badly almost everywhere, which makes it the rare category that is both high-intent and genuinely open. A practice that answers those six clearly is competing where the competition is weakest.

Why we lead with this rather than with content

Most marketing advice for medical practices begins with publishing, because publishing is what marketing services sell. The work with the highest return in this sector is administrative, and it is work we do as part of an engagement precisely because nobody enjoys doing it. Pulling every directory record, reconciling it field by field and filing corrections with each plan is slow, procedural and unglamorous. It also moves the number that matters.

The second entity record, and why it is the weaker one

Every physician exists twice — in the record they publish and in the provider directories they did not write, which are structured, easy to consume and consistently less accurate.

Every physician exists twice

You have the record you publish — your own site, your own profile, your own stated hours and networks. You also have a record you did not write: your entry in each health plan’s provider directory, your line on a hospital roster, your national provider entry and everything downstream of it. Both are reachable. The second is structured, which makes it easy to consume, and it is maintained by people who have never met you, which makes it consistently wrong.

The fields that drift

Address is the most common, usually because a practice moved and one plan never caught up. New-patient status is the most consequential, because a directory that says you are closed to new patients removes you from consideration entirely. Insurance participation drifts in both directions — listed where you are not, missing where you are. Specialty and hospital affiliation drift more slowly but stay wrong for longer once they do.

What a correction actually involves

Each plan has its own process, its own timeline and its own definition of done. A correction filed today may appear in four weeks or may quietly fail. That is why the sequence starts with the pull, not the publish: a correction you never verify is indistinguishable from one you never filed. We re-pull at six weeks as a matter of course and treat anything unchanged as still open.

Why this outranks a content programme

A wrong network listing does not cost a click. It removes you from the consideration set of someone who has already decided to seek care and is checking whether they can afford to see you. No amount of published content reaches that person, because they never got far enough to read anything.

It removes you from the consideration set of someone who has already decided to seek care and is checking whether they can afford to see you.

Administrative questions, answered properly

These six questions sit outside the safety filter entirely, are asked constantly, and are answered badly almost everywhere.

Accepting new patients

State it plainly, state it dated, and state it in the same words on your site and in every directory you appear in. If the answer differs by location or by clinician, say so explicitly rather than leaving one blanket sentence that is true for part of the practice. This is the single most-asked administrative question and it is the one most often answered by silence.

Insurance networks

Name the plans. ‘We accept most major insurance’ answers nothing, cannot be extracted and cannot be checked. A named list — plan by plan, with the date it was last reviewed — is specific, verifiable and directly useful. Where participation is pending or in negotiation, saying so is better than omission.

Referral requirements and first visits

Whether a referral is needed, what to bring, what happens at a first appointment and how long it usually takes. These are procedural facts about your practice that carry no clinical weight and therefore pass straight through the filter. They also reduce no-shows, which is a second return on the same page.

Wait times and telehealth

Typical time to a first appointment is almost never published, is asked frequently, and is answerable with a range and a date. Telehealth availability — what is offered, for whom, under what conditions, in which states — is rising demand met by thin supply. Both are straightforward pages that very few practices have written.

Measuring this without pretending it is precise

Three questions in two assistants, monthly, recorded verbatim — and accuracy of description treated as the primary metric rather than frequency of mention.

What to ask

Three questions, two assistants, monthly, recorded verbatim with the date: one new-patient question, one insurance question and one practice-identity question naming you directly. That set covers the administrative surface and tests whether you are described correctly when you are described at all.

What counts as a result

In this sector, accuracy is the primary metric and presence is secondary. A practice named with a wrong network list is in a worse position than one not named, because the wrong list actively misinforms someone making a care decision. Track both, and treat an accuracy failure as more urgent than an absence.

What not to read into it

Assistant answers vary between sessions and change without notice. A single run is an observation, not a measurement, and a month of runs is a trend rather than a rank. We report it that way and we would be suspicious of anyone reporting it more confidently.

Where this fits with the rest of search

None of this replaces ordinary search work. The relationship between AEO and SEO is complementary rather than competitive, and a practice that is invisible in conventional search will not be rescued by assistant visibility.

The relationship between AEO and SEO is complementary rather than competitive, and a practice that is invisible in conventional search will not be rescued by assistant visibility.

Where medical answers come from, by question type

The filter, made explicit
Patient question	Typical source	Filtered?	Can a practice appear?
Are you accepting new patients	Practice site and directories	No	Yes
Do you take [plan]	Practice site and plan directory	No	Yes
Is a referral required	Practice site	No	Yes
How long is the wait	Practice site	No	Yes
Do you offer telehealth	Practice site	No	Yes
What hospital are you affiliated with	Hospital roster	No	Yes
What are the symptoms of [condition]	Institutional health publishers	Yes	Rarely
Is [medication] safe with [medication]	Institutional health publishers	Yes	No
Should I worry about [symptom]	Institutional health publishers	Yes	No

The two records, field by field
Field	Who maintains the second record	Common failure	Consequence
Address	Health plans and hospital rosters	Former location retained	Patient goes to the wrong place
Accepting new patients	Health plans	Marked closed in error	Removed from consideration
Insurance participation	Health plans	Listed where not participating	Surprise cost at the visit
Phone	Health plans	Old front-desk line	Unreachable
Specialty	National provider record	Outdated subspecialty	Wrong referrals
Hospital affiliation	Hospital roster	Never updated after a change	Misleading on capability
Clinician list	Health plans	Departed clinicians retained	Appointments booked with nobody

What we do, in order, and what each costs
When	Action	Cost	Why it is in this position
Week 1	Pull every directory record	Free	Slowest propagation, so it starts first
Week 1	Reconcile field by field against your own records	Free	Finds the disagreements
Week 2	File corrections with each plan	Free	Multi-week lead time
Week 2	Publish new-patient status and named network list	Low	Highest-intent open questions
Week 4	Publish referral, first-visit, wait-time and telehealth pages	Low	Passes the filter, reduces friction
Week 6	Re-pull directories and confirm	Free	Corrections silently fail
Monthly	Three questions, two assistants, recorded	Low	Accuracy first, presence second

Content types ranked by what they return here
Content type	Return	Why	Who usually sells it
Named insurance network list	High	Specific, checkable, constantly asked	Nobody
New-patient status, dated	High	Most-asked, most often silent	Nobody
Referral and first-visit process	Moderate	Procedural, unfiltered, reduces no-shows	Rarely
Wait-time ranges	Moderate	Almost never published	Nobody
Telehealth rules by state	Moderate	Rising demand, thin supply	Rarely
Condition explainers	Low	Filtered to institutional sources	Most content packages
Treatment comparisons	Low	Guideline bodies own these	Most content packages

Want this done for your site?We build and maintain the search, content and paid programmes described on this page.

Get a free proposal

Multi-location and multi-clinician practices

Any fact that varies by location or clinician has to be stated per location and per clinician, because one blanket sentence is the most reliable way to produce a confidently wrong answer.

The blanket sentence is the failure mode

One statement that is true for part of a practice is the most reliable way to produce a confidently wrong answer. If one location accepts new patients and another does not, if one clinician participates in a plan and another does not, that has to be stated per location and per clinician. Systems extract the sentence they find; they do not infer the exception you left unwritten.

Clinician-level records need their own attention

Each clinician has their own directory entries, and departed clinicians are among the most persistent errors in provider data. A roster that still lists someone who left two years ago produces appointment requests for a person who cannot see them, and it undermines confidence in every other fact on the same record.

What to publish per location

Address, phone, new-patient status, the clinicians who practise there, the plans accepted there and the hospital affiliations relevant to it — each stated on its own page rather than in a combined list. Combined lists read well to a person scanning them and extract badly.

What we would not sell you

Three things a practice is commonly sold in this category that we would decline to provide, and the reasoning behind each.

A clinical content programme aimed at assistant citation

It sits inside the filter. We would rather spend the same budget on directory reconciliation and administrative pages, where the same money moves something measurable.

A guaranteed position in AI answers

Nobody can offer this honestly. Assistant responses vary between sessions and change without notice, and there is no ranking surface to buy into. What can be promised is that your facts are correct, consistent and extractable everywhere a system might look.

A monthly report dressed up as a rank tracker

We report what the assistants actually said, verbatim, with dates, alongside what changed in your directory records. That is less tidy than a number going up and it is the honest shape of the evidence.

We report what the assistants actually said, verbatim, with dates, alongside what changed in your directory records.

What an engagement looks like from your side

What we need, what we hand back, and how long the first cycle takes — bounded by how slowly plan directories process corrections.

What we need from you

A list of every plan you participate in, every location, every clinician and every hospital affiliation, plus access to file corrections or a named person who can. The audit itself we run. The corrections need your authority, which is the one part that cannot be outsourced.

What we hand back

A field-by-field reconciliation of every external record against your own, the corrections filed with dates, the administrative pages written and published, and a monthly record of what two assistants said when asked three questions about your practice.

How long the first cycle takes

Roughly six weeks to a first verified result, driven entirely by how long plans take to process corrections. Site-side work is faster and starts in parallel, so the administrative pages are live well before the directory work completes.

Where this sits next to the rest of your marketing

It is foundational rather than promotional. It does not replace referral relationships, reputation work or conventional search. It makes sure that when any of those sends someone to check a fact about you, the fact they find is correct.

Talk to us about your practice’s records

We start with a free field-by-field audit of every external record we can find for your practice, and show you the disagreements before anything else is discussed.

/contact

Everything else we have written on search, AI and getting found

Getting found in search

AI, AEO and what is changing

Paid media and lead generation

Websites and design

Choosing and working with an agency

By industry and by situation

Not sure which of these applies to you?Tell us the situation and we will say plainly what we would do first, and what we would not.

Talk it through

Getting found in search

AI, AEO and what is changing

Paid media and lead generation

Websites and design

Choosing and working with an agency

Social, content and brand

By industry and by situation

What happens to a medical question before an answer is assembled — The filter is the single most important fact about medical AEO. Content quality does not route around it, which is why a clinical blogging programme aimed at AI citation is the wrong instrument.

Which questions a medical practice can realistically appear in — Six administrative, four clinical. The administrative six are asked constantly, answered badly almost everywhere, and sit outside the safety filter entirely.

A physician has two entity records, and they disagree — The second record is the one assistants can reach through structured feeds, and it is consistently the less accurate of the two. Correcting it is unglamorous administrative work with more effect than a year of blogging.

Where a medical practice should spend effort first — Note the inversion against ordinary content advice: the top three are database corrections, not publishing.

Medical query types plotted against what governs them — The diagonal is close to perfect. The more clinical a question is, the more heavily it is filtered, and the less a single practice can do about it.

The order the work actually happens in — Directory corrections are the only item here with a multi-week propagation delay, which is why they start on day one rather than after the content work.

Medical AEO, in numbers — The most valuable work in this sector is free, administrative, and nobody sells it as a marketing service.

Frequently asked questions

Does AEO work for medical practices?

For administrative questions, yes and unusually well. For clinical questions, no — those are filtered toward institutional health sources by design, and a single practice will not be cited there regardless of content quality.

What is the safety filter?

When a health question arrives, assistants narrow their source set toward government health bodies, hospital systems and major medical reference publishers, and append a consult-a-professional line. It is deliberate, and content quality does not route around it.

Which questions can a practice actually appear in?

New-patient acceptance, insurance networks, referral requirements, wait times, telehealth availability and hospital affiliation. All six are administrative, all six are asked constantly, and all six are answered badly nearly everywhere.

Why are provider directories such a big deal?

Because every physician has a second entity record maintained by people who have never met them. It is structured, easy for systems to consume, and consistently less accurate than the record you publish yourself.

Which directory field causes the most harm when wrong?

New-patient status. A directory that says you are closed removes you from consideration entirely, before anyone reads a word you have written.

How long do directory corrections take?

Weeks, and each plan has its own process and its own definition of done. Some corrections quietly fail. That is why we re-pull at six weeks and treat anything unchanged as still open.

Should we publish condition and symptom content?

Not as a route to assistant citation. It sits squarely inside the filter. If it serves your existing patients for other reasons, that is a separate and legitimate decision.

How should we write an insurance page?

Name the plans, plan by plan, with the date the list was last reviewed. ‘Most major insurance’ cannot be extracted, cannot be checked, and answers nothing.

Is wait time really worth publishing?

It is asked frequently and published almost nowhere, which is the definition of an open question. A range plus a date is enough — precision is not required and would not be credible.

What about telehealth?

Availability, eligibility, conditions and which states you can treat in. Demand for this has risen faster than anyone has published clear answers, so the ground is unusually open.

How do we measure this?

Three questions in two assistants monthly, recorded verbatim with dates: one new-patient, one insurance, one naming your practice directly. Accuracy is the primary metric here; presence is secondary.

Why is accuracy the primary metric?

Because a practice named with a wrong network list is worse off than one not named at all. The wrong list misinforms someone making a care decision, which is a different category of failure from a missed click.

Do reviews matter?

The text does. A substantive review describing the administrative experience — getting an appointment, front-desk responsiveness, insurance handling — is retrievable evidence. A star rating alone is not.

Does this apply to specialists as well as primary care?

Yes, and referral requirements matter more for specialists. The referral question is among the highest-intent administrative questions a specialist practice receives and is rarely answered on the site.

What about multi-location practices?

Every fact that varies by location has to be stated per location. One blanket sentence that is true for part of the practice is the most common cause of a confidently wrong answer.

Can we be cited for clinical content ever?

Occasionally, where a practice publishes something genuinely distinctive and specific to its own procedures. It is not a reliable basis for a programme, which is why we do not build one on it.

How does this compare to ordinary SEO?

Complementary. A practice invisible in conventional search will not be rescued by assistant visibility, and the administrative pages described here serve both.

How long before anything changes?

Site-side administrative pages show in extraction within weeks. Directory corrections run on a multi-week cycle with no guarantee on the first attempt. Nothing here is instant and anyone promising otherwise is guessing.

Is there compliance risk in any of this?

The administrative content described here carries none. Anything touching clinical claims or patient information does, and is handled under your own compliance review rather than ours.

What is the cheapest thing to do first?

Pull your own listing from every health-plan directory you appear in and read it. It is free, it takes an afternoon, and it usually produces the most surprising finding in the whole engagement.

Do we need a page per insurance plan?

Not usually. One clearly structured, named and dated list is sufficient and easier to keep current, which matters more than page count here.

What if our practice is already at capacity?

The accuracy work still matters — being described incorrectly has consequences regardless of your book. The content work returns less when there is no capacity to fill.

Want this done for your site?We build and maintain the search, content and paid programmes described on this page.

Get a free proposal

Get a free marketing proposal

Tell us what you are trying to grow and we will come back with a plan, not a pitch deck. Same-day reply on weekdays.

AEO for Doctors: Working With the Safety Filter, Not Against It

What it costs: AEO for Doctors

Medical questions are filtered before an answer is assembled

The filter is the fact everything else follows from

What that leaves, and why it is better than it sounds

Why we lead with this rather than with content

The second entity record, and why it is the weaker one

Every physician exists twice

The fields that drift

What a correction actually involves

Why this outranks a content programme

Administrative questions, answered properly

Accepting new patients

Insurance networks

Referral requirements and first visits

Wait times and telehealth

Measuring this without pretending it is precise

What to ask

What counts as a result

What not to read into it

Where this fits with the rest of search

Where medical answers come from, by question type

Multi-location and multi-clinician practices

The blanket sentence is the failure mode

Clinician-level records need their own attention

What to publish per location

What we would not sell you

A clinical content programme aimed at assistant citation

A guaranteed position in AI answers

A monthly report dressed up as a rank tracker

What an engagement looks like from your side

What we need from you

What we hand back

How long the first cycle takes

Where this sits next to the rest of your marketing

Talk to us about your practice’s records

Everything else we have written on search, AI and getting found

Getting found in search

AI, AEO and what is changing

Paid media and lead generation

Websites and design

Choosing and working with an agency

Social, content and brand

By industry and by situation

Getting found in search

AI, AEO and what is changing

Paid media and lead generation

Websites and design

Choosing and working with an agency

Social, content and brand

By industry and by situation

Frequently asked questions

Get a free marketing proposal