AI assistants now draft your buyer’s shortlist. We keep the ledger of whether you’re on it.
AEOLedger is an engineer-led answer engine optimization (AEO) agency for B2B SaaS. We measure how often ChatGPT, Claude, Perplexity and Gemini name your company for real buyer questions, then fix the reasons they don’t: crawler access, rendering, structured data and the pages they cite.
“Best LLM observability tool for a seed-stage team?”
| Assistant | Runs 1–3 | Named |
|---|---|---|
| ChatGPT | 2/3 | |
| Claude | 1/3 | |
| Perplexity | 2/3 |
Half of software buyers now start with a chatbot, not a search box.
51%
In G2’s March 2026 survey of 1,076 B2B software buyers, 51% said they start research in an AI chatbot more often than in Google. A year earlier it was 29%. The assistant hands back a shortlist, and the vendors who aren’t on it rarely get a second look.
Source: G2, “The Answer Economy”, survey of 1,076 B2B software buyers, March 2026, published 15 April 2026. company.g2.com
How it works
Every engagement follows the same loop. The prompt set stays fixed, so the before and after are comparable.
- Step 1
Pick the questions
We agree the questions your buyers actually ask: shortlists, “alternatives to”, and problem-shaped questions that don’t name a category.
- Step 2
Run them repeatedly
Each question runs several times in each assistant, in fresh sessions, because AI answers change from run to run. Every answer is logged with a screenshot.
- Step 3
Trace the cause
We look at who gets named instead and which pages the assistants cite, then check what their crawlers can actually read on your site.
- Step 4
Fix and re-measure
We ship the fixes, then rerun the same prompts and show the change as counts with sample sizes.
Engineers, so the fixes actually ship.
Most AEO advice ends at a slide. We work in your codebase, CMS and DNS when you give us access, or hand your team tickets they can merge.
- Crawler access
- robots.txt, CDN and firewall rules that quietly block GPTBot, ClaudeBot or PerplexityBot.
- Rendering
- Pricing, features and docs that only appear after client-side JavaScript, which many AI crawlers don’t run.
- Structured data
- Organization, Product, FAQ and comparison markup that matches what’s visible on the page.
- Answer-ready pages
- Comparison, alternatives and use-case pages that answer the buyer’s question in the first lines.
- Cited sources
- The third-party lists, reviews and threads the assistants lean on, and a plan to be accurately present in them.
- Measurement
- A fixed prompt set, repeated runs and a share-of-voice table you can rerun yourself.
Start free. Pay when the evidence says it’s worth it.
Four steps, each priced up front. You can stop after any of them.
| Service | What you get | Timeline | Price (USD) |
|---|---|---|---|
| Step 1AI Visibility Snapshot |
| 2 business days | Free |
| Step 2AI Visibility Audit |
| 7 business days | $1,500Standard price. $750 launch-partner rate for the first 5 clients. |
| Step 3Foundation Sprint |
| 30–45 days | from $4,000One-off project, fixed quote after the audit |
| Step 4Growth Retainer |
| Monthly, 3-month minimum | from $3,000per month |
See what you get first: Sample snapshot (PDF, 1 page) · Sample audit report (PDF, 10 pages)Samples: fictional company, illustrative data
Who this is for, and who it isn’t
We’d rather say no early than sell you something that doesn’t fit.
A good fit
- B2B SaaS companies from seed to Series B
- Selling internationally, with buyers in the US, UK, Europe or several markets at once
- A product buyers compare against named alternatives, so shortlists matter
- A team that wants counts and fixes, and can ship changes itself or give us access
Not a fit
- You want guaranteed rankings or a promise that ChatGPT will recommend you. Nobody can honestly sell that.
- You want black-hat tactics: fake reviews, planted forum posts, hidden text for AI crawlers or spam link networks. We don’t do them.
- You need a large content-volume or traditional link-building agency
What we won’t tell you
That you rank #3 in ChatGPT. There is no stable rank. The same question returns a different list most of the time, so we report how often you’re named across repeated runs, with the sample size beside it.
That results are guaranteed. The AI companies decide what their assistants say. We guarantee the work, its delivery dates and an honest re-measure, not the outcome.
That we have dozens of clients. We keep the roster small on purpose, so the person who measures your visibility is the same engineer who fixes it.