There's a new step in the purchasing process you can't see. Before a customer asks their AI assistant to find a lawyer, book a restaurant, order a part, or compare SaaS tools — the agent goes first. It scans your platform, tries to extract the information it needs, and either completes its task or gives up.

Most businesses score closer to "gave up" than they realize.

You've probably been on the other side of this. You've used ChatGPT to research a competitor. You hit a wall — pricing buried in a contact form, no spec data anywhere, a website that looked fine but told you nothing. You moved on. That experience you had as a researcher? That's what's happening when an agent goes looking for your business on behalf of a potential buyer.

This isn't a prediction about the future. It's an operational reality today for businesses that run AI agents in production — and we're one of them. The businesses that are legible to agents get the transaction. The rest don't. The frustrating part: this isn't a technology problem. It's a documentation and structure problem. And almost every business on the internet has it.

You run a business solo or with a small team — and you ARE the tribal knowledge.
Your customers could tell you things about your offer that aren't on your website. Every "well, it depends" answer you give on a sales call is a gap an agent can't bridge. The Surface Audit tells you exactly where those gaps are and how to close them. For most solopreneur and small business owners, the Quick Wins list alone covers the audit fee. If the audit reveals more than you can tackle alone, the Remediation Prompt Pack gives you 12 Claude prompts that produce implementation-ready fixes you can deploy yourself.

You have a dev team and you're actively deploying agents.
You know agents are reading business surfaces — you build workflows that depend on it. The Full Spectrum Audit gives your team a prioritized fix roadmap they can execute against immediately. The ARI Remediation Prompt Pack, included with Full Spectrum, functions as implementation specs as much as execution tools — each prompt produces a deployment-ready artifact your developers can review and ship.

Your Tribal Knowledge Gap is the worst of any segment.
Mid-market companies with siloed teams carry the heaviest tribal knowledge burden: pricing exceptions in one person's head, product selection guidance that lives only in sales decks, edge cases handled by whoever picks up the phone. If your company has multiple departments that all know different pieces of the customer-facing answer — and none of those pieces are published anywhere — the Full Spectrum Audit with the stakeholder session is built for this situation.

Not sure where you fall? The Surface Audit is the right starting point. It tells you what we can see from the public layer, scores you on five dimensions, and identifies the gaps. The report itself will show you whether the internal layer is where your most critical exposure lives.

We run your business through a structured, 5-dimension assessment and score it on the Agent Readability Index (ARI).

Not a theoretical framework. An empirical probe — we run actual AI agents against your platform using a standardized task battery, document every stall point, validate your structured data against published standards, and identify every piece of critical knowledge your business holds that doesn't exist anywhere an agent can find it.

The result is a scored report with a prioritized fix roadmap. Your developers can work from it directly.

1. Discoverability
Can an agent find your offer, understand what it is, and determine whether it fits a user's need — without calling you? We test schema markup completeness, navigation architecture, meta data quality, and how your business surfaces when an agent uses web search to research your category. The most common failure: product descriptions written for humans, with no machine-parseable specifications.

2. Transactability
Can an agent initiate or complete a transaction on behalf of a user without escalating to a human? We map your checkout flow step by step and count every decision point that requires inference, guessing, or information an agent can't reasonably possess. Phone-number-required checkout fields, booking systems that need a phone call to confirm, pricing that only appears after add-to-cart — for an agent, these are dead ends.

3. Schema Integrity
Is your structured data complete, accurate, and consistent across your platform? We run Google's Rich Results Test against your site, validate your product or service schema field by field, and check whether what your schema says matches what your pages actually show. Schema that contradicts visible content actively misleads agents.

4. Context Density
Does your platform contain enough machine-parseable information for an agent to make a confident purchasing recommendation? We check for comparison content, specification tables, policy accessibility, and social proof agents can actually extract. "Contact us for pricing" is fatal. Agents don't call.

5. Tribal Knowledge Gap
How much critical business knowledge exists only in your team's heads? This is the dimension most businesses don't think about — and the one where owner-operated businesses and mid-market teams are most exposed. The answer to "what's the question you answer on every sales call that isn't on your website" is always illuminating. Agents can't read obvious.

Step 1 — Live Agent Probe (before we talk to you)
We run a standardized task battery against your public platform using Claude before we've exchanged a word with you. This is intentional — we want empirical failure data before we know what you think we should find. Every stall point, every failed task, every inference the agent had to make — documented verbatim. This section of your report is reproducible. You can run every task yourself and compare.

Step 2 — Structured Intake Diagnostic (you complete, pre-engagement)
A 40-question diagnostic that maps your platform architecture, content inventory, transaction path, and what you already know about your own agent readability problems. The act of completing this reveals the tribal knowledge gap faster than any other method.

Step 3 — Document Harvest (you provide, pre-engagement)
Your product catalog export, FAQ documentation, and — if you have it — the internal onboarding document you give to new support staff. That last one is the tribal knowledge bible.

Step 4 — Shadow Session (optional, powerful)
We give you a prompt. You run it against your own platform using Claude or ChatGPT and record a 5-minute Loom. Clients who do this consistently say it was worth the price of the audit on its own. There's nothing like watching an AI agent break your business in real time.

Before we sold this to anyone, we ran the full methodology against sigmafoundry.io. We scored 41 out of 100. Agent-Partial.

Our Discoverability was the strongest dimension at 61. Schema markup was present but incomplete — price and availability fields were there; brand, SKU, and aggregate rating were absent. Transactability scored 44 — Gumroad checkout requires an account creation step agents can't complete autonomously. Schema Integrity was 38 — service pages had stale prices that didn't match current listings. Context Density scored 45 — product descriptions were readable but not structured. Tribal Knowledge Gap exposure was 67: most of what makes our agency distinctive doesn't exist in any format an agent can find and use.

We fixed the Schema Integrity issues in a single afternoon. ARI moved to approximately 52 after that one session. We're publishing the full self-audit findings on our blog.

No other vendor offering this type of service ran the same methodology on themselves first, published the score, and documented what changed. That's the credibility claim. It's verifiable.

30-day satisfaction guarantee — if the audit contains fewer than 5 actionable findings or a material methodological error we failed to correct, you receive a full refund. No runaround.

Score revision policy — submit your evidence within 30 days. If it supports revision, we fix it and reissue everything at no charge.

Prompt Validity Guarantee — if a remediation prompt produces incorrect output against its stated target, we fix it within 5 business days and push the update to all purchasers.

Confidentiality by default — everything you share is confidential. NDA available at no cost on request.

Annual re-audit discount — 40% off within 12 months. Surface: $299. Full Spectrum: $900.