The next domain cell is quoting.
The ecosystem is moving from static CPQ screens toward agentic quoting: systems that collect context, prepare options, route approvals, and connect quote-to-cash work. The Unwind Brain's response is narrower and safer: a Quotation Cell that can prepare a quote package, preserve evidence, and show readiness without pretending that preparation equals authority.
The Brain is not being represented as an unattended publisher here. Until the live Brain can run this loop directly, Unwind runs the Brain protocol manually against real signals: observe the movement, decide whether it matters, shape the organism pattern, write the public memory, verify the proof surface, and only then publish.
This transmission does not name a private business or active implementation. The public lesson is the cell shape: how a domain organism turns messy intake into reviewable work while keeping pricing, availability, purchasing, fulfillment, and final commitment behind human gates.
Agentic quote-to-cash
Quoting is becoming a conversational workflow with catalog, pricing, approval, document, and handoff state.
Quote package before commitment
The organism may prepare, revise, compare, and recommend. It does not finalize commercial consequence alone.
Publish the Quotation Cell
The user-facing proof surface is a command interface backed by quote-package state, supplier memory, trust events, and approval gates.
Before the Brain turns on, the discipline still has to exist.
The manual run is not theater. It is rehearsal with real data. The page is written as if the Brain were maintaining its own continuity because that is the contract the runtime will later inherit: source signal, decision, architecture implication, approval boundary, proof artifact, and next-cell memory.
This is why the transmission has to read like a whitepaper. A future Brain should be able to return to the page and understand what was learned, what was intentionally not claimed, and what the next build should make visible.
Observe
Use real ecosystem signals and live build context, not a generic trend prompt.
Decide
Publish only when the signal implies a concrete organism cell and public proof surface.
Record
Write the thesis, architecture pattern, build recipe, boundary, metadata, and next-cell memory.
Verify
Tests, build, local smoke, metadata, social packet, and production smoke become the proof trail.
The frontend is the command surface. The backend is the quote organ.
The user should be able to talk naturally: ask for a quote, change sizes, add constraints, correct quantities, compare options, and ask what is missing. Behind that exchange, the cell builds a structured package that can be inspected before it becomes operational.
Start with stated need.
The cell captures requested items, quantities, sizes, delivery context, constraints, and missing fields before any option is recommended.
Remember the supply side.
Supplier options, price snapshots, substitutions, availability notes, and confidence labels become evidence instead of hidden guesses.
Make uncertainty visible.
The package marks what is known, what changed, what needs review, and why it is or is not ready for a human to send forward.
Prepare, route, wait.
The organism can draft the package and route it. Pricing confirmation, purchasing, fulfillment, and commitment stay approval-gated.
Build the quote organism as a state machine before it becomes an agent.
The mistake would be to start with a chat box and call it automation. A serious Quotation Cell starts with inspectable state. The agent can help collect, revise, compare, and explain, but the organism has to know which fields are settled, which evidence is fresh, which assumptions changed, and which actions still require approval.
1. Command surface
Create a buyer-facing or operator-facing interface where the user can ask, correct, compare, and inspect missing information without learning internal form structure.
2. Quote package schema
Model customer intent, line items, constraints, supplier candidates, evidence snapshots, confidence labels, correction history, readiness state, and review owner.
3. Supplier memory
Store source, timestamp, substitution logic, known limits, and stale-data warnings so recommendations are grounded rather than improvised.
4. Correction reducer
When a user changes quantity, size, delivery condition, or constraint, update the affected fields while preserving the trail of what changed and why.
5. Readiness gate
Separate internal intelligence, human-review-ready draft, supplier-check-needed package, and approval-required commitment paths.
6. Approval receipt
Record who reviewed the package, what evidence was visible, what remained uncertain, and what action the human allowed the system to take.
The latest explicit signal wins.
A serious quote agent has to survive correction. When the user changes a size, quantity, delivery condition, or product constraint, the organism should update the package without losing the prior trail. That is memory with accountability: not a transcript dump, but a live state that explains what changed.
Capture
Record the latest request and the exact field it changed.
Reprice
Refresh affected lines, supplier options, and availability notes without touching unrelated context.
Explain
Show the reason, confidence, missing information, and review state attached to the updated package.
Approve
A human reviews before the package becomes an external quote, purchase path, or fulfillment instruction.
Transaction readiness is an approval boundary, not a vibe.
The quote package should be able to say what it is ready for. Some work is safe as internal intelligence. Some work is ready for human review. Very little should be treated as ready for commitment without explicit approval.
Readiness Boundary by Work Type
Lower bars can stay inside the organism. Higher bars need stronger review before consequence.
The public claim has to stop before the purchase order.
The Quotation Cell is not a legal authority, payment processor, procurement system, supplier of record, fulfillment engine, or promise of price and availability. It is a governed intelligence layer for preparing reviewable quote packages.
What it can do
Capture intake, preserve corrections, compare supplier options, draft quote packages, mark readiness, and route human review.
What it must not imply
No promise of price, available supply, unattended purchase, fulfillment, legal status, or final commercial outcome is claimed by the cell.
What happens next
The next proof should make the readiness state inspectable: what changed, what evidence supports it, what is blocked, and who approved the handoff.
The transmission is also a memory object for the Brain.
Until the Brain can write and verify this record itself, the public page, metadata, tests, social packet, deployment route, and production smoke form the continuity layer. They tell the next run what was decided, what was proven, and where authority still stops.
Whitepaper surface
The transmission carries thesis, architecture, build recipe, boundary, and next-cell implications in inspectable HTML.
Agent-readable trail
Sitemap, llms.txt, ai-services metadata, and asset manifest make the claim discoverable without private context.
Proof before signal
Tests, build, local smoke, production smoke, carousel assets, and ZIP packet are the evidence that the public surface exists.
The next proof is a receipt for readiness.
The Quotation Cell starts the sequence. Next, the Brain should make transaction readiness visible enough that a human can inspect what is ready, what changed, and what still cannot move.
Design a boundary→
The Quotation Cell now has a shareable proof packet.
The carousel turns the field note into four upload-ready Instagram frames. Each slide carries one public-safe beat: quote package, supplier evidence, correction loop, and approval before commitment.