I turned scattered agent research into an in-product onboarding that helps travel support agents trust Acai, learn to prompt it, and get to their first "aha" moment fast.

Client: Acai Travel, an AI copilot for travel customer-support agents (OTAs and TMCs) My role: UX research synthesis, onboarding strategy, flow design and build Methods: Affinity mapping, user journey mapping, persona-based opportunity mapping Built with: Userflow (in-app checklists, tooltips, nudges) Deliverables: Research synthesis, design principles, onboarding scope, live onboarding flows

Acai sits inside the agent's CRM and answers questions about fares, baggage, GDS commands and airline rules. Agents had access to it, but many treated it as a nice-to-have. They opened it, got a vague answer, and went back to their old tools.

The goal was not a feature tour. It was to prove value fast, inside the agent's real workflow, so adoption grows on its own. That meant making three benefits visible: faster answers, fewer errors, easier communication with customers.

Friction came down to two things: trust and AI literacy. I synthesised agent interviews and feedback into an affinity map with five clusters.

The big picture: adoption depends not only on agents, but on manager priorities and the existing tool ecosystem. Every time agents stepped outside Acai to double-check, its perceived value dropped.

The affinity map showed what the problems were; the journey map showed where they hit. The emotional low point is case investigation, with the customer waiting and five tabs open.

Instead of one long product tour, I built six small flows that appear at the right moment and build on each other. Each one only shows up once the agent has seen the one before it, and only for clients enrolled in onboarding.

The onboarding was designed as a system, not a one-off tour:

Behavioural segmentation so tooltips and nudges reach the right agents at the right moment

A/B testing of wording, timing and placement of nudges

Feature announcements to keep agents up to date as Acai evolves

Data-driven optimisation based on the success metrics below

Multilingual content that adapts to the agent's language