This Meta project is password-protected. Enter the password to view it.
← Back to portfolioCross-platform concept testing across two markets that pressure-tested a subscription model before launch — driving design changes and shifting the roadmap from rigid subscriptions to flexible, usage-based plans.
This was evaluative design research and product strategy — testing a business model and its end-to-end user journey — that translated directly into UX and content-design changes. As co-lead, I structured the study around six journey stages, ran cross-platform interviews, and delivered stage-by-stage recommendations that prompted banner redesigns, clearer copy, and a fundamental rethink of the pricing model.
Meta was preparing to monetize its Business AI agent, which handles sales and support conversations on behalf of businesses. The team had designed a usage-based subscription model with tiered plans, but needed to validate whether real business users could:
I co-created the discussion guide and stimulus materials, framed the research around three constructs — comprehension, discoverability, and perceived value — and structured the study around six stages of the monetization journey, so findings would be immediately actionable for different stakeholders.
A note on confidentiality: due to NDA, I can't show the actual screens or any specifics of the subscription plans and pricing. The journey framework and findings here are shared at a level that protects confidential product details.
Monetizing an agent is itself a human–AI interaction problem: people have to understand what the AI does, what it costs, and what it will and won't do at the limit.
Specific findings that drove design changes: confusion over plan-tier differentiation → clearer tier naming; misunderstanding of rollover mechanics → copy improvements; unclear agent-behavior boundaries → contextual messaging; platform-specific payment expectations → payment flexibility; and a need for earlier billing visibility → repositioning the billing surface in the journey.
What worked: cross-platform synchronization yielded richer recommendations than isolated studies; AI-assisted translation made me self-sufficient; the journey-stage framework made findings immediately actionable; and evidence-based de-prioritization (showing what not to build) proved as valuable as confirming what to build.
What I'd do differently: push for earlier cross-team alignment on stimulus materials; make the most critical findings more scannable; and invest earlier in understanding a fast-moving domain.
What I learned: how to drive roadmap- and business-model-level influence through research; how to operate cross-platform with aligned methodology but flexible execution; and how to ramp into a complex domain quickly.