Pre-seed · 2026

The work-ownership
layer for AI agents.

Contractual gives organizations a rigorous way to transfer a class of consequential work—and gives agents a path to earn more.

The categoryOWNfrom model capability to organizational responsibility

The missing layer

Agent stacks can execute. They cannot hold the job.

WHAT EXISTSModelsToolsMemoryOrchestrationCapability reaches the system of record.
THE GAP?

What may the agent own, what proves completion, and who expands its responsibility?

WHAT BUSINESS NEEDSOwned outcomeOne accountable operator for a defined class of work.

Why now

Capability is compounding.
Responsibility is stuck.

MODEL CAPABILITYBroadening every release
SYSTEM ACCESSAgents can already act
WORK OWNERSHIPStill trapped behind human completion

The next infrastructure market is not more intelligence. It is the transfer of responsibility.

The product

One contract carries work from evidence to ownership.

01Definethe job and outcome
02Replayagainst completed cases
03Grantexplicit authority
04Executein the external system
05Verifyread-back and receipt
06Promotethe next work class
THE OPERATING CONTRACT BINDS

Outcome · evidence · authority · actions · postconditions · economics

Horizontal architecture

The kernel stays fixed. The job changes.

CONTRACTUAL KERNELContractReplayGrantReceiptPromotion
REVENUEInbound opportunity response
FINANCEVendor exception resolution
OPERATIONSDocument package completion
NEW DOMAINForm from the job, not a template
Domain contracts and adapters vary. Ownership semantics do not.

The entry motion

One buyer. One queue. Two weeks to the first contract.

DESIGN PARTNEROperations owner

Active AI mandate · existing agent initiative · direct production authority

  1. DAY 0Select

    A frequent, measurable, consequential queue.

  2. WEEK 1Form

    50–100 resolved cases become a temporal Replay.

  3. WEEK 2Contract

    Price the first ownership-ready class.

  4. NEXTActivate

    Countersign the grant and execute live work.

Product status

A working kernel. A clear external proof boundary.

SHIPPED

Versioned operating contracts and amendments

Temporal Replay with future-field exclusion

Signed grants and multi-party activation

External execution, read-back, and postconditions

Content-addressed receipts and public verification

Ownership reporting, promotion, and recall

NEXT PROOF01

Paid formation on a real queue

02

Countersigned live ownership

03

Externally verified completion

04

Promotion and second-domain transfer

Business model

Revenue rises only when ownership rises.

FORMATION$15k

Work Ownership Review

OPERATION$90k

Initial annual commitment

EXPANSION

More verified work classes

Escalations and reversals never bill. The unit is verified, receipt-bearing work.

Market expansion

Land on a queue. Become the ownership layer.

01One queueprove one owned class
02One functionexpand across adjacent classes
03One enterprisestandardize contracts, evidence, and economics
04Across agent stackssystem of record for transferred responsibility

Defensibility

The evidence compounds at the work layer.

COMPOUNDING ASSETWhich contract reliably owns which work?
CONTRACT GRAPHREPLAY → LIVE CALIBRATIONRECEIPT + OUTCOME HISTORYCROSS-DOMAIN TRANSFER

Models commoditize. Contractual learns how responsibility moves.

The raise

$400k to convert the kernel
into market evidence.

USE OF FUNDS
45%

Founding engineering

35%

Forward-deployed partner work

20%

Infrastructure and operations

THE FINANCING EARNS

01 Three paid formations

02 First live receipt

03 First ownership promotion

04 Second-domain transfer

The agent economy needs an ownership layer

Give AI
a real job.

CONTRACTUAL.TECH / DECKS / PRE-SEED