Skip to content
Ibid.

Answer engine optimization for professional practices

The answer names
one firm.
Make it yours.

Your patients and clients now ask an AI before they ask anyone else. Ibid builds practice websites that retrieval systems can actually read, then measures your share of the answers every month and reports it against booked revenue.

Request an Answer AuditSee the method

From $2,500, by practice type. Three weeks. Credited in full against any engagement you start within sixty days.

ChatGPT
Dental

The question

“

Sources

  • 1harborlightdental.com
  • 2ada.org
  • 3healthgrades.com

Answer

For full-arch implant cases in San Diego, patients most often see Harborlight Dental1 named, based on documented case volume and a verified provider profile.

Illustration

Ibid.

The footnote mark for the source cited again. Being named once is luck. Being named every time, across every engine, is a position, and it is the only thing worth paying for.

What the research actually says

This category runs on assertion. We would rather start from evidence, including the parts that are inconvenient for us.

46 x

Difference in how often brands are cited between the most generous and least generous answer engine, across 34,234 responses.

Multi-engine citation study, 2026

11 %

Of domains cited by ChatGPT are also cited by Perplexity. One optimization does not win every engine, and anyone selling you that is guessing.

Analysis of 680M citations

<5 %

Share of all citations held by the single most-cited domain on any platform. The remaining 95 percent is fragmented, which is exactly why a specific local practice can win.

Evertune, 200M prompts

0

Measurable citation lift attributable to llms.txt in four independent studies. We ship one anyway because it costs an hour. We will not bill it to you as a strategy.

Limy, OtterlyAI, ALLMO, SE Ranking

The incumbent problem

What you are probably paying for right now.

Most professional practices are on a retainer with an agency doing competent work against an obsolete definition of winning. None of this is malice. It is a 2015 playbook still running in 2026, on a monthly invoice.

01

Rankings you cannot spend

A monthly PDF showing position four for a phrase nobody types anymore. Position on a page that fewer people scroll is not a business outcome, and it was never meant to be reported as one.

02

A template forty practices share

The same vendor layout, the same stock operatory photo, the same three paragraphs with the city name swapped. Retrieval systems collapse near-identical pages. Being interchangeable is the one thing that guarantees you are never the answer.

03

Locked out at the door

Bot rules and firewall defaults that block GPTBot, ClaudeBot, and PerplexityBot without anyone noticing. We find this on roughly every other site we audit. You cannot be cited by a system that has never been allowed to read you.

The method

Indexed, cited, recommended.

Three stages, in order, because they depend on each other. You cannot be cited by a system that cannot read you, and you cannot hold a position you never measured.

01

Indexed

Make the practice legible.

Answer engines parse structured data to confirm who you are, what you do, who works there, and whether anyone credible agrees. Most practice sites give them prose where they need facts. We rebuild the site so every provider, procedure, location, and credential resolves as a real entity, and so the direct answer sits at the top of the page instead of below three paragraphs of welcome copy.

  • Entity graph in JSON-LD
  • Answer-shaped content
  • Crawler access engineered
  • Vitals passing at launch

02

Cited

Become the source.

Legibility gets you eligible. Citation is earned off-site. Web mentions are the strongest known predictor of whether a model names you, ahead of backlinks, so the program runs on coverage and corroboration: answer-shaped pages against the queries you do not yet cover, review volume that survives being parsed, and mentions in the places that feed the graph.

  • Answer coverage program
  • Review and reputation engine
  • Digital PR for mentions
  • Entity maintenance

03

Recommended

Hold the position and prove it paid.

Being named once is an accident. Being named consistently across five engines, three runs, every month, is a position. We measure it the same way every time, we show you the runs where you lost, and we tie the whole thing to booked revenue rather than to impressions.

  • Monthly Answer Share
  • Per-location tracking
  • Revenue attribution
  • Quarterly strategy
Read the full method

The instrument

One number, measured the same way every month.

Answer Share is the percentage of your golden query set where your practice is named or cited. Forty queries, agreed with you. Five engines. Three runs each, so you are reading a measurement rather than a lucky screenshot.

We report the runs you lost alongside the ones you won, and we show coverage gaps explicitly rather than quietly shrinking the denominator. A number that can only go up is not a measurement. It is marketing.

How we measure

Answer Share, twelve months

Share of a 40-query golden set where the practice is named or cited.

Sample
  • ChatGPT
    4% → 38%
  • Perplexity
    11% → 62%
  • Google AI Overviews
    7% → 40%
  • Gemini
    3% → 31%
  • Claude
    5% → 35%
Sample Answer Share by engine, month one to month twelve
EngineMonth 1Month 12
ChatGPT4%38%
Perplexity11%62%
Google AI Overviews7%40%
Gemini3%31%
Claude5%35%

Illustrative shape only. These are not client results. When Ibid has client results, they will appear here with the practice named and the query set published.

The arithmetic

Run the numbers before we ever meet.

Premium pricing is only defensible if the arithmetic works, so here is the arithmetic, with every assumption exposed and editable. Put your own case value and close rate in. If it does not clear, do not hire us.

Your numbers

Average case value$4,200
Inquiry to client conversion35%
Additional inquiries per month+8
Engagement

Modeled twelve-month return

+$123,120

Net of fees, at 7.8x return on the engagement. Each additional inquiry is worth $1,470 at your close rate.

Modeled twelve-month return
Revenue added$141,120
Investment$18,000
Net$123,120
Return multiple7.8x

Break-even

This engagement pays for itself at 1.0 additional inquiries per month. Everything above that is return.

This is a planning model, not a projection. It multiplies your own inputs and assumes nothing about what we will deliver. We publish it because the arithmetic should be yours to check before you spend anything.

Engagements

Published prices, in public.

Nobody in this category publishes pricing, which tells you the number moves based on what they think you can pay. Ours does not.

00 · Diagnostic

The Answer Audit

from $2,500one-time, by practice type3 weeks

Before anyone sells you a strategy, you should know exactly where you stand. We build your golden query set, run it across five engines, and hand you the baseline in writing. $2,500, $4,500, or $7,500 depending on practice type; the exact figure is on the price card below.

Best for

Any practice that wants the number before the pitch.

01 · The build

Indexed

Most start here
from $12,000one-time, by practice type8 to 10 weeks

A new site built the way retrieval systems actually read: clean entities, valid structured data, answer-shaped content, and instrumentation wired in from day one.

Best for

Practices whose current site was built for 2015 search.

02 · The program

Cited

from $595per month, by practice type12-month term

The ongoing work of becoming the source. Monthly measurement, entity maintenance, answer coverage, and the off-site mentions that actually move citation rates. Priced at 60 percent of Recommended for your practice type, without the exclusivity, and with first right to your slot if you upgrade.

Best for

Single-location practices in a competitive trade area.

03 · Full market position

Recommended

from $995per month, by practice type12-month term

Everything in Cited, plus the thing nobody else sells: exclusivity. One practice per territory and specialty, held in a ledger we honor. When you hold the slot, the answer we are optimizing for is yours alone.

Best for

Independent practices that want the market position held, not shared.

See what is included

Where we work

Four practices. Four different questions.

We work only where the buyer researches before they call and the case value makes the work worth doing. That is a narrower list than most agencies will admit to.

Dental and orthodontic practices

Dental

The implant consult starts in a chat window.

Single implant case
$3,000 - $6,000
Full-arch case
$25,000 - $50,000

Asked today

“who is the best dentist in san diego for dental implants”

Private practice and specialty medicine

Medical

Patients triage with an AI before they ever call an office.

New patient value
$1,200 - $4,000
Surgical episode
$8,000 - $40,000

Asked today

“best orthopedic surgeon in san diego for rotator cuff repair”

Government contractors and defense suppliers

GovCon

Primes and program offices research you before they call.

Teaming introduction
prime relationship
Task order
$250k - $5M

Asked today

“small business CMMC level 2 compliant software firms with SOCOM past performance”

Law firms and legal practices

Legal

The first consultation is now with a chatbot.

Matter value
$5,000 - $75,000
Contingency case
six figures

Asked today

“do i need a lawyer for a construction defect claim in california”

Questions

The ones that matter.

Answered directly, at the top, in plain language. Partly because that is how you should be treated, and partly because it is exactly how a page earns a citation.

What exactly is Answer Share?

The percentage of your golden query set where your practice is named or cited in the generated answer. We agree the query set with you at kickoff, run every query three times on each of five engines to account for variance, and report the share monthly. Same queries, same method, every month, so the number is comparable to itself. That is the whole point.

Why five engines and three runs?

Because a single run is a screenshot, not a measurement. Answer engines are non-deterministic, and citation behavior differs enormously between them. Any agency showing you one flattering ChatGPT screenshot is showing you noise. Three runs across five engines is the minimum that lets you distinguish a real move from variance.

Can you guarantee we will be cited?

No, and you should leave any meeting where someone does. We guarantee the method, the cadence, and the reporting. What we will commit to in writing: if your Answer Share has not improved by month six of a Cited or Recommended engagement, we continue working at no fee until it does or you leave with no termination obligation.

Do we need a whole new website?

Not always. The Answer Audit tells us. If your entity data is fixable and your content architecture is sound, we would rather remediate than sell you a rebuild. If the site is a shared template on a platform that will not let us control markup or crawler access, remediation is throwing money at a wall and we will say so.

How is this different from the SEO we already pay for?

Traditional SEO optimizes for a ranked list of links. Answer engines do not produce a ranked list of links. They retrieve, synthesize, and cite, which rewards entity clarity, answer-shaped content, and corroboration rather than keyword density and link volume. Much of good SEO still helps. The reporting layer built around it has stopped describing reality.

Why do you publish your prices?

Because almost nobody in this category does, and the reason is usually that the number changes based on what they think you can pay. Ours moves only with what a new client is worth in your field, and the entire card is public. It also happens to be good practice for what we sell: a page that states a concrete number is a page an answer engine can actually cite.

Something not covered here? Ask it on the audit request and it gets a real answer, whether or not you ever hire us.

Start here

Find out where you actually stand.

Every engagement opens with the Answer Audit. Forty queries, five engines, three runs each, and a written baseline. If the number is already good, we will tell you that and you will have spent the audit fee to stop worrying.

Request an Answer AuditCompare engagements