AI-Augmented Plan Assembly for Wireless Sales Reps

AI Strategy Lead

Held a recurring strategic-advisory cadence with the carrier's department director, working through AI use-case prioritisation across the department's portfolio. This engagement emerged from that cadence: a cognitive-load problem on the sales floor was identified as a candidate for AI augmentation, and I was assigned to lead the design and originate the application. The carrier had no in-house AI design proficiency. Ran strategy sessions with the director throughout the engagement, translating executive intent into the design strategy that shaped the pilot. As the work evolved, drove the team's shift toward a more collaborative, agile way of working with engineering and the product owner.

Principal Designer

Sole principal designer on the engagement. Owned conversation and UX design, built the prototype, adapted the existing design system for the AI use case, and saw the work through to production-ready UI. The questionnaire-as-conversation interaction model at the heart of the application was originated here.

AI-Augmented Design Practice

An early "design with AI" pilot inside the enterprise. Initial prototyping was done entirely in Figma Make, which served as a fast executive-alignment artefact. As the work moved toward production, the design process shifted to a conventional one. The lessons from that transition (on where AI-driven prototyping accelerates the work, and where it stops being sufficient for production craft) were socialised across the consultancy and are being adapted into broader practice.

Challenge

Building wireless plans is high-cognitive-load work for sales reps. In-person retail and contact-centre reps spend significant time and effort manually assembling the right combination of plan features, promos, discounts, and internal product codes for each customer. They do this while holding a live conversation with the customer: listening for what the customer wants, mentally translating between casual descriptions of needs and the technical product catalogue, looking up the relevant SKUs, and assembling a quote that the rep can defend on the spot.

The result is a slow, error-prone, mentally exhausting workflow. It costs the carrier in throughput (longer calls and visits), training (long ramp times for new reps), and consistency (the plan a customer ends up with depends heavily on which rep they happen to talk to). Competitive research suggests that other Tier-1 carriers have similar projects in early R&D. This is where the sector is heading.

Goals

Reduce

the cognitive load on sales reps building wireless plans during live customer conversations

Embed

AI capability inside the sales workflow itself, where reps already work, without forcing them to context-switch

Move

from a manual lookup-and-assembly process to AI-assisted plan generation

Preserve

rep judgement and customer relationship as the centrepiece of the interaction

Why This Approach Was Different:

The substance of this engagement was reducing the cognitive load on the rep while still putting AI inside the workflow.

Most executives who imagine AI for a sales floor reach immediately for a chatbot. The reasoning is intuitive: if AI can hold a conversation, point it at the rep and let it help. The problem with that intuition is straightforward when you watch a rep work. Reps are already holding one demanding conversation with the customer. Adding a second conversation with an AI assistant, one that the rep has to actively manage in real time, multiplies the cognitive load. Discovery and rep observation made this clear early.

The interaction model that emerged was different. AI runs in the background, gleaning the relevant plan details from the natural flow of the rep's conversation with the customer. The rep's surface is a flexible questionnaire-and-notepad UI that affords free-form note-taking rather than a unidirectional line of questions. Reps can converse naturally and capture the customer's needs in their own words. The AI does the work of mapping those notes to the technical catalogue, assembling the candidate plan, and presenting it back to the rep with tools to tweak and adjust. The rep stays in the conversation. The AI handles the assembly the rep used to do in their head.

The questionnaire layer turned out to do double duty. As well as serving as the rep's input surface, it gives the business a structured place to update plan parameters, promotional logic, and discount structures as those things change. The application is configurable without redesign each time the business shifts.

Strategic Deliverables

Discovery and rep-observation findings that informed the interaction model
Conversation-and-questionnaire interaction model conceived and validated for production
Prototype built with AI tools; used as the executive-alignment artefact and rapid-iteration tool
Production-ready UX and UI, including adaptation of the existing design system for AI-augmented use
Findings on AI-augmented design practice, socialised across the org

Outcomes

First-of-its-kind AI-integrated sales tool for this carrier,

now in pilot. Addresses a workflow that competitors are building toward in early R&D.

Reps using the application stay in the customer conversation.

AI handles plan assembly in the background; the rep refines the result.

Configurable architecture

allows the business to update plan parameters, promotional logic, and discount structures without redesign.

Foundation laid for the next stage:

extracting plan requirements directly from the live conversation, removing the questionnaire intermediary.

Internal findings on AI-augmented design practice socialised

across the org and being adapted into broader practice.

About

A Tier-1 Canadian wireless carrier
2026 (in pilot)

Impact at a Glance:

First-of-its-kind AI-integrated sales tool for this carrier. AI assembles wireless plans inside the rep-customer conversation. Currently in pilot.