Design AI conversations that lead somewhere.
We help organisations create chat, voice, and AI-assisted experiences that understand user needs, guide people clearly, and connect them to the right outcome.
The Problem
Users experience frustration when their request is misunderstood; the answer does not solve the problem; the conversation reaches a dead end; human handoffs lose context; different channels provide different answers; or accessibility is treated as an afterthought.
Poor conversational experiences increase support demand, reduce trust, and create more work for internal teams.
We design the conversation as part of the whole service.
Understand user intent
Identify what people are trying to achieve—not just the words they type or say.
Design clearer conversations
Create simple flows, useful responses, and relevant next steps.
Improve error recovery
Help users move forward when the assistant cannot understand or complete the request.
Design human handoffs
Transfer the user and their context without forcing them to start again.
Connect service channels
Create consistent experiences across chat, voice, web, mobile, contact centres, and offline services.
Build in accessibility
Design for different impairments, communication needs, and levels of digital confidence.
Understand. Map. Design. Test.
01 — Understand
Review user needs, service goals, and existing conversational experiences.
02 — Map
Connect the conversation to the wider journey, teams, systems, and handoffs.
03 — Diagnose
Identify confusion, abandonment, failure points, and unmet needs.
04 — Design
Create improved conversation flows, responses, escalation routes, and prototypes.
05 — Test
Validate the experience with real users before scaling.
Depending on the engagement, your organisation may receive: user-intent mapping; conversation journey mapping; failure-point analysis; improved conversation flows; human-handoff design; tested prototypes; accessibility recommendations; and a prioritised improvement roadmap.
A well-designed conversational experience understands the need, guides the user clearly, completes the task, escalates with context, and improves through evidence. The goal is not simply to make an AI assistant sound more human. The goal is to make the service work better.
Service designers who understand AI—and the service around it.
Stray Theory brings together service design, user research, UX, conversational design, and practical AI experimentation. Our team’s experience includes work connected to: Britvic / Pepsi licensing · Peet’s Coffee · JDE.
Use client names and logos only where the relationship and usage permissions are confirmed.
Who This Is For
A Conversational UX Review may help if your chatbot has low completion or satisfaction rates; users frequently abandon conversations; support teams receive repeated questions; human agents receive incomplete context; chat, voice, and digital journeys feel disconnected; you are planning a new AI assistant; or accessibility has not been properly assessed.
Talk to Us
Make sure it helps people move forward rather than creating another layer of friction. No generic AI demonstration. No technology-first pitch.