BT Wholesale, Message Now Chatbot
AI-powered chatbot for BT Wholesale, designed to deflect common queries and support users contextually. Reducing pressure on service teams and streamlining self-serve journeys.
Problem: After speaking with BT Account Managers and the Service Help Desk, we found that a large chunk of their time was spent solving issues users could easily self-serve — like resetting passwords or locating the right documentation. This not only delayed responses to more critical problems like outages but also put users through frustrating, multi-call journeys.
Goal: Design and deliver a chatbot that could sit across the BT Wholesale site, surfacing common support questions contextually based on the user’s location — and escalate to the right support teams when necessary.
Outcome: The chatbot launched as a smart layer across BT Wholesale’s support site, capable of addressing common user questions and freeing up service teams to focus on high-value problems. The structured flow and content planning approach laid a foundation for future BAU enhancements and AI model training.
Skills used: Leading workshops | Wireframing | Prototyping | CUI design | Team leadership | 3 Amigos | Requirement gathering | Stakeholder management | Sprint planning | Content strategy | Conversation flows
The Approach
Collaborative discovery
Partnered with a Business Analyst, Content Strategist, and UI Designer to explore how a chatbot could reduce pressure on service teams and improve user experience.
Sprint planning:
Took a lead role in shaping the work — breaking it down into epics, features, and tasks across UX, UI, content, dev, and QA. This helped the Scrum Master and PO structure quarterly planning and gave stakeholders clear visibility into delivery.
Conversation workshops
Built on early BA requirements by running FigJam workshops to surface key use cases, draft components like intro prompts, navigation flows, FAQs, and escalation logic.
Scripting and tone:
Led quick-turnaround scripting exercises with the team to test tone of voice, user inputs, and response flows.
Flow modelling
Worked with the UI Designer and Content Strategist to build complete end-to-end conversation flows tailored to different stages of the user journey — from homepage FAQs to live-agent escalation during transactions.
Documentation for delivery
Formalised the conversation structure to guide development and copywriting, shared regularly in sprint demos and Amigo sessions with developers to align build.
Content sourcing
Co-led workshops with four internal support teams to surface the top issues users faced, categorised into: reseller self-serve, service desk needs, and inter-team escalations.
Content synthesis & training:
Consolidated all chatbot content ideas into a prioritised spreadsheet, reviewed with support leads, then used to generate over 20 intent variants — enabling smarter AI training and response mapping.
Build review
Reviewed the initial chatbot build and led refinement sprints to optimise any flows or logic gaps.
