Node · VendorSage
Everything we've decided, researched and left open
Four documents, consolidated from the work-intelligence sessions. The plan, the competitive picture, the commercial shape, and the things we still don't know. Supersedes the earlier design note and the strategy memo.
01The thesis
Companies don't know how they work.
Revenue has a CRM. Money has a ledger. People have an HRIS. Code has a repository. Work itself — the thing all of those exist to support — lives in people's heads, stale wikis and habit.
That gap shows up as problems nobody diagnoses as the same problem:
- decisions made against an imagined version of the company
- budget following the loudest voice rather than measurable friction
- processes that walk out the door when someone resigns
- onboarding conducted as oral tradition
- AI initiatives aimed at work nobody has actually described
The shift from procurement
We started in procurement — renewals, spend, vendor decisions. Real business, not going away. But it answers what are we paying for, not how does this company actually work.
The second question is bigger, and procurement decisions quietly depend on it. You cannot judge whether a tool is worth renewing without knowing what work it powers, who relies on it, and whether the process it supports functions at all. So the centre of gravity moves to work intelligence, with procurement as the way in rather than the destination.
The inherited advantage. On day one we already hold a customer's tool stack, spend, seat counts and login data. Everyone else in this category arrives knowing nothing and has to ask.
Positioning line
What your tools do, plus what your people do, plus what your documents claim — reconciled, and kept continuously true.
The reconciliation is the value. When the documents say approvals sit with Finance, the people say it's really one person's assistant, and the system logs say most approvals skip both — that disagreement isn't a data-quality problem to clean up. It's the most valuable thing we can put in front of a leadership team.
02Hard constraints
Stated repeatedly across the sessions and not negotiable. Everything in the other three documents is downstream of these five.
- Product first
- No service-backed product. If a decision only works because a person does it for the customer, it is the wrong decision.
- Zero services
- Not "product-led with a services arm" — zero. This is the single biggest structural difference from Human Layer Lab and from the consultancies, and it is what makes the mid-market addressable at all.
- Self-serve
- The price floor is set by compute, not human time. That is the only way a 200-person company can afford to be mapped more than once.
- The customer never authors
- They correct, confirm and react. They never fill a blank page. Only ask users to do things they'd want to do anyway — arguing with a draft of their own team's process passes; authoring a taxonomy does not.
- Nothing leaves the workspace
- The workspace is the hard data boundary — EMEA, Enterprise, an acquired subsidiary. Everything scopes to it.