Module 1
What we actually sell
One script, one endpoint, one analyst that knows the business.
After this module: You can explain what BusinessMCP does in thirty seconds, name who buys it, and say honestly who should not.
The one-liner, and why it is that one
Start here, because everything else is detail. BusinessMCP turns a company into a single, AI-ready data platform. A business connects its tools — analytics, CRM, ad platforms, Stripe — and installs one tracking script. We unify all of it into one hosted MCP server, a business-intelligence dashboard, and an AI business analyst sitting on top that can answer questions about the numbers.
That is the platform. But most deals do not start there, because "unified data platform" is not a problem anyone wakes up with. They start with something much more concrete: most of the companies visiting your website never identify themselves, and you are losing them.
Notice the shape of it: a problem they recognise, then what we do about it, then how little work it is. No mention of MCP, no mention of architecture. Those come later, when they ask — and a good prospect will ask.
Check yourself
Two ways to consume it
This trips people up, so be clear about it. There is one product and one codebase, but customers use it through two different front doors.
| The dashboard | Their own MCP endpoint | |
|---|---|---|
| Who uses it | The founder, marketer or sales lead, in a browser | Any external AI — Claude, ChatGPT, Gemini — or their own scripts |
| What they get | Reports, CRM, the AI business analyst, the outreach engine | The same unified data and the same toolset, over a standard interface |
| Typical buyer | Self-serve, marketing-led | More technical teams; often the enterprise conversation |
| What you demo | Almost always this | Only when they ask, or when the buyer is technical |
The second door is genuinely unusual and worth understanding, because it is what stops us being "another analytics dashboard". A customer gets a secure endpoint with a key. They paste it into Claude or ChatGPT, and that AI can now ask real questions about their business — not general knowledge, their actual numbers.
For most self-serve deals you will demo the dashboard and mention the endpoint in one sentence. For a technical buyer, the endpoint is often the thing that makes them lean forward — read the room.
Check yourself
Who buys, and who should not
The product works best in a narrow band, and knowing that band is most of qualification.
- Good fit: a B2B company, roughly 10 to 500 people, with real inbound traffic and no dedicated data or RevOps function. Marketing lead, founder or head of sales is the buyer.
- Good fit: a team already doing outbound whose reply rates are poor because the list is cold and generic.
- Poor fit: very low traffic. Identification is a percentage of visits — a few hundred visits a month produces almost nothing worth acting on.
- Poor fit: consumer businesses. Company identification works off business networks; consumer and mobile traffic mostly cannot be resolved.
- Poor fit: a large enterprise with an in-house data platform team. They will build it, and the buying cycle will exhaust you.
~98%
of visitors never fill in a form
~20–35%
of B2B traffic resolvable to a company
5–15%
reply rates on warm outreach, vs 1–3% cold
Those three numbers are the commercial case, and you should know them cold. They are also hedged ranges rather than promises, and that is deliberate — see module 4.
Check yourself
Worth remembering
- A business connects its tools and installs one script; we unify all of it into one hosted MCP server plus a dashboard and an AI business analyst.
- There are two ways to consume it: our dashboard, or the customer’s own MCP endpoint that any external AI can call.
- The commercial wedge is narrower than the platform: we identify the companies visiting a website and turn that into booked meetings.
- The buyer is almost always a B2B company with real inbound traffic and nobody whose job is joining the data up.
- A prospect too small to have traffic, or big enough to have a data team, is usually a bad fit — and saying so early wins more than it loses.
Words that came up
- First-party data
- Data a company collects about its own visitors on its own site. Ours is first-party, which is why it keeps working as third-party cookies disappear.
- ICP (ideal customer profile)
- The description of who a product is genuinely good for. Ours is B2B companies with meaningful inbound traffic and no dedicated data function.
- Warm outbound
- Contacting someone who already showed interest — visited the site, read pricing — rather than a cold stranger from a list.
Do this before moving on
Reading about a demo is not the same as having given one.
- 1Say the thirty-second version out loud, twice, without reading it. If you stumble, your version is too long.
- 2Open our own homepage and find the one sentence we lead with. Notice that it is not a feature.
- 3Write down three companies you know that fit the ICP and one that clearly does not, and be able to say why.