AI Quoting -
Sales rep assistant pricing

Designing and validating a new key feature part of SalesHub and CustomerHub

Company

Choco - Digitialising food industry

Role

Lead Product Designer: owned end-to-end design for AI Quoting, SalesHub's and CustomerHub's quote-builder across web and app, partnering with Engineering (Backend, Frontend, ML), Product, and GTM/Sales.

Duration

May - Sept 2026 · Shipped three sequential releases in iterative cycles

Challenge

Food distributor sales reps built quotes for their products manually across spreadsheets and email, often with no visibility into margins and no consistent approval step before a price went to a customer.

Pricing setups varied wildly by distributor including different ERPs, currencies, gross profit margin rules, and legal constraints.

Approach

User problems

Sales reps had no fast, accurate way to build customer quotes with margin guidance.
Sales and pricing managers had no consistent way to review pricing before it reached a customer.

Goals

1. Enable reps to build fast, accurate, margin-aware quotes for their leads and customers by scanning a menu or document and matching products with AI reducing manual errors and time to create a quote.

2. Give managers and pricing teams visibility and control over margins through an approval flow that doesn't slow reps down.

3. Support complex pricing setups by allowing to import margin data replacing fixed price sheets and enabling syncing quoted prices back into the ERP to avoid manual overhead.

Process

Led discovery with Product together with different sized distributors to understand current processes and pain points.
Designed close to Engineering to figure out edge-cases and implemented in tight-knit cycles.
Validated flows directly with early-adopter distributors surfacing real edge cases: e.g. currency mismatches, minimum vs target margin logic gaps, ERP price-sync needs.
Ran frequent bug bashes with engineering ahead of each release, prioritising issues with Product before rollout.
Identified EU legal restrictions on Document scan and adapted the product so European distributors could quote via manual entry and menu scan only without losing quoting
Worked with the ML team on the underlying menu-scan and product-matching model from user’s expectations perspective.

Design principles

Transparent: when reviewing the quote created from a menu scan, it should be clear which product has been matched and what the guardrails are for adjusting a price following target and/or minimum margins
Flexible: sales reps should be able to generate a quote with automatically approved prices without unnecessary back and forth with the pricing team
Trustworthy: sales and pricing managers should be able to trust the tool following their company pricing guidelines and have the ability to view automatically approved prices if needed

Key design decisions

Surface margin health inline using a traffic lights system so reps are aware where they can adjust pricing, when they’re below target and which prices will need manual approval.
Avoid unnecessary back and forth between pricing team and sales reps by adding comments.
Give the ability for sales reps to export approved prices in a PDF without gating the whole process due to an approval.
Allow syncing back approved prices to the ERP to avoid manual overhead.
Hide margin data from quote builder for instances where sales rep quotes prices in front of the customer protecting sensitive profit data.

Results

Met all implementation deadlines and gathered contextual feedback by scoping product in 3 sequential releases, each one focusing on solving a different problem:

1. Product matching

2. GP% margins

3. Approval process & comments

A live demo to a ~£95M-revenue Distributor drew direct feedback that the tool was "market leading."

A customer indicated Quoting could unlock over £10K in incremental annual contract value.

Iterated on an ERP price-sync path, so quoted prices now flow back into a distributor's own systems automatically after receiving customer feedback.