The first visit now belongs to an agent.

On your website, your Agent Concierge answers the buyer's agent and lets it complete the request with the person's go-ahead. Inside your product, the upgrade moment reaches your customers where they already work.

The story — including the reversal

We started by making products legible to AI agents — detection, structured content, the intent graph. It worked, and then the platforms proved the point for us: agentic discovery and checkout became free platform defaults. When you publish into someone else's protocol, you sit inside their walled garden — and lose the signal that was always the real product.

So we reversed, deliberately and in the open, toward SaaS products consumed through their own MCP servers, where a paying user hits a wall mid-task and the funnel can't see it. The ecommerce estate stays live. For software, we built the revenue layer: detection to decision to checkout to the resumed task, on the vendor's own Stripe.

Then the same shift reached the website. Buyers began sending an agent to research a business, and more and more to take the next step for them, often with nobody watching a browser. That conversation is the only record of what the buyer needed. So we built a website that can talk to agents: an Agent Concierge that answers their agent with proof, lets it complete the request with the person's go-ahead, and keeps the record in your workspace. We took our own website live on it first.

"The pricing page is disappearing into the tool call."

"The contact form is disappearing into the tool call."

As buying moves into agents, the research, the evaluation and, more and more, the next step happen in conversations your analytics never see.

Their agent does the research. You win the customer.

Four principles that shape the product.

Honest by construction

No offer without a recorded block — enforced in the schema, tested in CI. We claim live capability, never customer proof we don't have. The honesty engine is the culture, not a feature.

Agent-native thinking

We don't retrofit dashboards for AI. We build for where buying and using now happen: in the conversation between a buyer's agent and your business, and inside the tool call.

Every yes on the record

Nothing reaches you without the person's go-ahead, and every lead says how it was given. An agent can complete a request, but it can never opt anyone into your marketing.

Privacy-first architecture

Contact details in an agent's question are masked when it is recorded, and end-user identity in your product is pseudonymous by contract. Card data never touches us — Stripe owns everything sensitive.

How we got here.

  1. 2024 Q4
    Idea born: AI agents are reshaping how products get discovered
  2. 2025 Q1
    First prototype: edge-based agent detection on Cloudflare Workers
  3. 2025 Q2
    The ecommerce estate ships: llms.txt, structured data and the commerce protocol files
  4. 2025 Q4
    The intent graph ships
  5. 2026 Q1
    The platforms commoditize agentic checkout — and validate the thesis
  6. 2026 Q2
    The pivot: the company reorients around MCP Revenue Orchestration for SaaS
  7. 2026 Q3
    Revenue moments ship: expansion and free/trial conversion, proven end to end in Stripe test mode
  8. 2026 Q3
    Websites that talk to AI agents: their agent, your Agent Concierge, and the conversations recorded in your workspace
  9. 2026 Q3
    We take our own website live on the platform, by asking Claude

Nearly half of consumers now ask AI to recommend a local business. Every business needs a concierge.

45% of US adults used an AI tool for local business recommendations in the past year (BrightLocal, 2026). See what their agent can read and do on your site today.

Inception Agents