Service description (plain text)

Also served at /llms.txt.

# AI Search Visibility

## What this is
Generative Engine Optimization (GEO) consulting: getting a company cited,
recommended and correctly described when a customer asks an AI assistant a
question in its category.

## Operator
Isaac Hevi — computer engineer, sole operator. No agency, no account
manager, nothing subcontracted. The prompt execution, citation parsing, entity
checks and month-over-month tracking are all custom code, not a reseller
dashboard.

## How the engagement runs
Entirely over email, including scoping and reporting. There are no calls and
nothing to book. The only entry point is an email to
isaachevi@aisearchvisibility.site
The site asks for two things — name and company website URL — and opens the
visitor's own mail client with the message already written. Nothing is posted
to a third party and nothing is stored.

## Who it is for
B2B SaaS companies in considered-purchase categories, whose customers compare
vendors before committing, and who already have at least ten pages of real
content.

## Who it is not for
Local services, transactional businesses with no research cycle, and
pre-product-market-fit companies still defining their customer.

## Engines measured
ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews.

## Why it matters
- 900M people use ChatGPT every week (OpenAI, February 2026)
- 94% of B2B buyers used generative AI in their purchase process (Forrester Buyers’ Journey Survey 2026, ~18,000 buyers)
- 55% use it specifically to compare vendors against each other (Forrester, January 2026)
- 80% of the time, the vendor a buyer already favours before contacting sales goes on to win the deal (6sense, 2025)

## Method — The Citation Loop
Four stages, running on a fixed cycle. Each pass compounds on the last.
1. MEASURE — a prompt matrix built from the language the category's customers
   actually use, spanning informational, comparative and commercial intent,
   executed across five engines with repeated sampling per prompt, because
   these systems are non-deterministic and one run is noise. Every response
   logged, every cited URL captured.
2. STRUCTURE — pages rebuilt for passage-level extraction: heading hierarchy,
   self-contained answer blocks, claim density, internal consistency. Then the
   structured data layer, wired into one entity graph with identity signals
   consistent across every source a model reads.
3. SOURCE — placement on the specific third-party pages the citation map proves
   are being pulled from in the category. Targeted by evidence, not by domain
   rating.
4. RE-MEASURE — the identical prompt matrix, re-run on a fixed interval and
   scored against baseline. Share of voice, citation count and how the models
   describe the company, tracked as a delta rather than a snapshot.

## Deliverables
- Visibility Baseline: A prompt matrix built from the language your customers actually use, spanning informational, comparative and commercial intent, executed across five engines with repeated sampling per prompt. Every response logged and scored: were you named, in what position, and in what framing.
- Citation Source Map: Every URL the engines cited, deduplicated by domain, resolved to its owner and ranked by citation frequency across your query set. Split three ways: sources you control, sources you can earn, sources you can’t touch. That split is what makes the plan actionable instead of aspirational.
- Retrieval Engineering: Your highest-intent pages rebuilt so a model can lift a clean, correct answer out of them — heading hierarchy, self-contained answer blocks, claim density, internal consistency. Plus the structured data layer: Organization, Product, FAQ and comparison schema wired into a single entity graph with consistent identity signals across every profile a model reads.
- Off-Site Source Placement: Placement work aimed only at sources your citation map has already proven are being pulled from in your category — review platforms, category listicles, directories, community threads. Every target justified by evidence from your own data, not by a domain rating.
- Measurement & Crawl Monitoring: The same prompt matrix re-run on a fixed interval so month-over-month deltas actually mean something. Plus server-log monitoring of AI crawler activity — GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended — so you can see retrieval demand, not just outcomes.

## Pricing
One offer, one scope. Only the billing period changes; the work does not.
Month one includes the full build — it is not an extra charge on top of the
first month. From month two onward the rate drops to $1,200 per month, less on a
longer commitment.
- Monthly: $1,500 charged today (then $1,200/month · cancel any time). Month one $1,500 and includes the full build; $1,200 per month after that.
- 3 months: $3,510 charged today (covers months 1–3 · you save $390). Month one $1,350 and includes the full build; $1,080 per month after that.
- Yearly: $11,760 charged today (covers 12 months · you save $2,940). Month one $1,200 and includes the full build; $960 per month after that.
Larger scope quoted by email.

## Timeline
Days 1-14: baseline measurement pass and citation source map. Days 15-30:
retrieval work shipped — entity graph, schema layer, rewritten pages,
comparison architecture. Days 30-60: source placement executed. Day 30 and
monthly thereafter: re-measurement against the identical prompt matrix.
Citation movement is typically measurable in 60 to 90 days.

## Guarantees
None on placement — AI engines are closed systems that change retrieval
behaviour without notice, and they are non-deterministic. What is guaranteed:
measurement against the same prompt matrix every month, and work shipped rather
than recommended.

## Relationship to SEO
Complementary, not a replacement. SEO ranks a link; this gets the brand named
inside the answer. AI systems reference well-ranked content heavily, so strong
organic presence accelerates the work.

## Contact
The entry point is email, and there is nothing else to fill in:
isaachevi@aisearchvisibility.site
Site: https://aisearchvisibility.site