Revora is a governed quote-to-order workspace for growing B2B teams. I designed the product around a central tension: sales needs speed, while finance and operations need every price, discount and customer document to remain trustworthy.
Role
Product strategy, UX, UI system and design QA
Team
Independent product with AI-assisted execution
Scope
Admin setup → quote creation → PDF → purchase order
Status
Working private-pilot product; outcomes are product evidence, not market claims
revora / commercial operations
2distinct role experiences
31product states audited
61trust and workflow issues mapped
1governed quote-to-PO journey validated
A quote is not just a PDF.
In many small and mid-sized B2B teams, a customer quote begins in a spreadsheet, borrows prices from another file, gets checked over chat, and ends as a manually edited document. The customer sees one PDF; the business carries the risk of every disconnected handoff behind it.
That reframed the problem from “make quote creation easier” to “create one controlled commercial record from price to purchase order.”
The fragmented workflow
Each handoff introduces another place for pricing, context or ownership to drift.
01Price listWhich file is current?
02Discount checkCan sales offer this?
03Quote documentAre totals and fields correct?
04Customer approvalWhich version was accepted?
05Purchase orderCan it be traced back?
I separated configuration from execution.
The system had two fundamentally different jobs. Administrators needed to define the commercial boundaries; sales representatives needed to create a valid quote without learning those boundaries every time. Combining both roles would either slow sales down or expose controls that should remain protected.
Admin · Finance · Operations
Configure the guardrails
Products and price models
Discount and tax rules
Templates and field visibility
Permissions and workspace defaults
Sales representative
Work inside the guardrails
Select an approved template
Add customer and products
Resolve inline pricing feedback
Preview and export the final quote
Admin architecture · Six connected governance journeys, mapped before screen-level execution.
Sales architecture · A focused path from template selection to export history.
01
Govern before the quote
Put products, prices, permissions and templates under administrative control before sales begins composing a document.
02
Guide, then intervene
Use sensible defaults and inline feedback for normal work. Reserve blocking states for decisions that create genuine commercial risk.
03
Preserve the record
Treat each exported quote as a versioned business record so later product or pricing changes cannot rewrite history.
04 · The central interaction decision
Make the system explain the rule at the moment it matters.
A pricing policy hidden in documentation will be ignored. A policy revealed only after export wastes work. I moved validation into the quote table where the representative is already making the decision.
Within policyContinue silently
Do not interrupt correct work.
Warning thresholdExplain and allow
Surface the consequence while preserving momentum.
Blocking thresholdStop and resolve
Protect the business when the action exceeds authority.
The blocked state connects the exact row, violated threshold, quote summary and disabled export action.
Global defaults, then explicit exceptions.
I organized pricing controls into four predictable domains—discount, tax, product pricing and export rules. Global thresholds provide a safe baseline; named rules can target specific products or categories. This keeps everyday setup approachable while still supporting real B2B edge cases.
Implemented pricing controls · Defaults, permissions, thresholds and targeted rules share one mental model.
A guided workflow built from approved inputs.
Sales begins with an approved template rather than a blank document. Customer details, eligible products, field visibility and commercial rules arrive with it. The builder then keeps totals, discounts and document readiness visible without forcing the representative through a long configuration wizard.
Template discovery reduces blank-page decisions and carries approved structure into the quote.
The quote builder keeps the editable work, live commercial total and readiness context in one view.
Preview exactly what the customer will receive.
The export step removes internal columns and protected pricing logic, validates required customer-facing information, and shows the final document before generation. A generated file is tied to an immutable quote version, allowing a later purchase order to point back to the exact commercial record the customer accepted.
Customer-ready preview · What users review is what the product exports.
Quote version→Validated export→Customer PDF→Purchase order
08 · The moment the project changed
The interface looked finished. The workflow was not trustworthy yet.
I audited 31 authenticated desktop, tablet and mobile states across four administrator personas. The audit exposed a gap between visual polish and product truth: hard-coded metrics, false-success messages, unpersisted configuration, missing role enforcement and incomplete responsive states.
“The next release should establish workspace identity and make one workflow fully persistent before polishing secondary UI.”
P0
Trust blockers
Identity, persistence, empty states, governance and mobile access.
P1
Workflow gaps
Templates, pricing, customers, permissions, versioning and auditability.
P2
System quality
Accessibility, responsive behaviour, error handling and complete state coverage.
Before the auditAdd more product surface
→
After the auditMake one journey real end to end
I treated code as a design material.
Not every screen moved through a conventional Figma-to-handoff path. I mapped complex flows, used AI to explore UI directions, and moved promising patterns into a coded product early. Working in the real system exposed permission, persistence, validation and edge-state problems that static happy-path screens would not reveal.
01
Model the work
Define actors, states, rules and the quote lifecycle.
02
Explore rapidly
Generate and critique multiple interface directions.
03
Prototype in code
Test the experience with real roles, data and constraints.
04
Audit the truth
Compare visible success with what the system actually saved and enforced.
05
Harden the journey
Prioritize trust, accessibility and end-to-end reliability.
What the current product proves.
Role integrity
Admin and Sales experiences remain isolated.
Sales can quote, but cannot access or mutate protected administration controls.
Pricing integrity
Warnings and blocks are evaluated in the workflow.
Out-of-policy discounts create visible feedback and prevent invalid export.
Document integrity
Exports are tied to a fixed quote version.
The generated PDF and later purchase-order record remain traceable to what was approved.
System integrity
The critical journey is tested end to end.
Setup, quoting, validation, preview, export and PO registration are exercised as one connected flow.
Polish earns attention. Trust earns adoption.
Revora began as an effort to simplify quote creation. Building the working product made the deeper challenge visible: commercial teams will only move away from spreadsheets when the replacement is not just easier, but more dependable.
If I continued the product, I would validate the warning-versus-blocking model with real pricing managers, measure time-to-valid-quote against the existing process, and test whether administrators can predict how a rule will affect sales before publishing it.
Next case study
Designing enterprise utility intelligence.
Smart CX · Multi-persona analytics and operational workflows
