AI on your organisation's reality
Ask your organisation.
Get a path, not a paragraph.
Nodwise turns how your organisation actually works into a typed, live graph — so your AI agents answer by traversing it, and show you every edge the answer was made of.
Pre-launch. One click with your work account — no form, no sales call.
If Victory Bikes fails, which customers are affected?
- Victory BikesExternal entity · Partner & supplierhas party ←
- Supply of Touring Tire TubeGovernance & compliance · Agreementis subject to ←
- Touring Tire TubePhysical assets · Inventory & materialdepends on ←
- SO70282Business · Commercial interactionis placed by →
- Dalton AdamsExternal entity · Customer
Answer: Dalton Adams. A second supplier holds a contract for the same part — reroute, not stop.
The formula
Metamodel + instance graph + AI.
The grammar
What can exist, how it may relate, which states it can be in. Closed on purpose.
The facts
Systems, processes, people, data and obligations — each a typed node with an owner.
The consumer
Agents stop summarising documents and start traversing your organisation.
See it work
Real screens. No mock-ups.
Every clip below is a recording of the demo tenant you can open yourself.
Questions answered as paths.
Ask in plain language. The agent traverses typed relationships and returns the path it took — not three plausible paragraphs.
- Every answer carries its edges
- Crosses functions nobody owned end to end
- Nuance kept: second suppliers, superseded states
A closed grammar for your organisation.
What kinds of things exist, how they may relate, which states they can be in — readable end to end.
- Node types on organisational and functional planes
- Relations with allowed sides and cardinality
- Lifecycle states per type
Start anywhere. Keep expanding.
Find one node and open its neighbourhood as a diagram. Every + walks one more typed edge — from a product category to each bike to the parts it is built from.
- 2D diagram and 3D scene of the same graph
- Expand one hop at a time, along typed relations
- Filter the diagram by relation type
Every change has an owner and a decision.
Anyone can propose a change. It reaches the accountable person as a task with the exact diff, and lands in the model only when they approve it — already translated.
- Proposals as tasks, with the diff to approve
- Approve, reject or request improvement
- Machine translation into every supported language
Loading, checked
Strict on the way in. Fast anyway.
24 min
to load 2.1 million nodes from empty
6.3 million property values, 4.2 million relations. Every node checked.
3%
of that time spent checking
47 of 1,416 seconds validate against the grammar.
28 min
to read it all again
Short enough for every night.
Agents that watch
It sees before you do. It doesn't act alone.
Every node carries owner, lifecycle and lineage. Agents identify, report and escalate — they never change your organisation on their own.
- A process lost its ownerEscalated to the accountable role — the model knows who that is.
- A control drifted from its policyThe edge still exists; the states no longer agree.
- A critical task rests on one personNobody wrote it down. The graph shows it anyway.
The long read
Why your AI implementation failed — and why better retrieval won't fix it.
Retrieval finds the closest paragraph. You needed the actual relationship. The structural argument, the impact example, and the failure mode nobody catches.
See it
A fully modelled organisation, open now.
One click with your work account. No form, no call.
Open the demoBuild it with us
Design Partner Program
Your own environment, your data in the model, a solutions architect modelling your first dimension with your team. A limited group.
Apply to the program