Consumer AI Assistant Gallery: Scaling PLG and Enterprise Discovery

Soul Machines · Group Product Manager

Turned a consumer assistant gallery into a market-validation engine, lifting free-to-paid conversion 48% and ARPV 209% in six months while proving there was a real consumer market for empathetic AI.

Walkthrough: a guided tour of the Consumer AI Assistant Gallery.

Consumer AI Assistant Gallery homepage showcasing diverse digital people and use cases — A consumer storefront where anyone can find a digital person and start in seconds.

Gallery options showing different AI assistants available for various use cases — Browse assistants by the job to be done, not by the technology.

Detailed view of AI assistant capabilities and interaction options — Each assistant earns trust upfront by being clear about what it helps with.

Interview Coach session demonstrating the rehearsal use case in action — Practise a real interview with zero judgement, as often as you like.

Context

Consumer-facing product to prove real value for digital people, designed to help individuals pick an assistant and start a useful dialogue fast (not novelty). We needed a place to demonstrate digital people in action with real use cases: enterprises required this proof, but we had nothing concrete to show them. The Consumer AI Assistant Gallery became both our market validation experiment and discovery funnel, generating thousands of sign-ups and paid customers monthly while teaching us what actually resonated with users.

Challenges

First-party consumer validation: building on our 2023 desk research, we needed real, first-party data on what use cases would drive actual demand for digital people.
Understanding consumer behaviour with AI assistants: activation, trust, willingness to engage on sensitive topics.
Rapid iteration at low engineering lift: to keep pace with market shifts.
Monetisation efficiency at consumer price points: driving ARPV without sacrificing conversion.

What I did

Owned consumer product vision and roadmap: defined a Gen-Z-oriented, no-judgement assistant catalogue focused on useful help (practice, rehearse, get unstuck).
Ran contextual inquiry, usability tests, and cohort analysis: converted insights into scenario-led UX and pricing/packaging experiments.
Built a lightweight experiment system: feature gating, cohorting, rapid copy/placement tests to learn without queueing on engineering.
Established weekly business review: tracking traffic to registration to activation to free-to-paid to ARPV, turning findings into weekly product and lifecycle changes.
Privacy and safety cues tuned for Gen-Z onboarding: tone, tiles, 1-tap start for faster selection and first dialogue.
North Star was conversations delivered, not MAU: optimised minutes-to-first-dialogue and depth of session.
Insights from consumer usage: informed our enterprise positioning and demos.

Extended detail panel showing conversation starters and customisation options — Conversation starters get people into a useful dialogue faster.

AI Gallery persona research and user segmentation — Persona research grounded every design choice in a real user.

Pricing research and packaging experiments — Pricing experiments lifted revenue per user without hurting conversion.

Design thinking session: user persona mapping — Mapping personas together aligned the team before building.

Results

Traffic: 9.7k visitors in first month; approximately 46k visitors in H1 2024.
Conversion: raised free-to-paid conversion from 27% to 40% (+48%).
ARPV: increased from $21 to $65 (+209%) within six months.
Paid customers: stabilised at approximately 90/month even as traffic declined from 9.7k to 2.4k.
Use case validation: identified three core value areas: onboarding/engagement (getting people to take action), training (interactive content delivery), and rehearsal (practising vulnerable scenarios).
Empathy insights: discovered users valued digital people for discussing sensitive topics they weren't ready to discuss with humans, validating our "no-judgement assistance" bet.
Enterprise proof: Gallery learnings became the foundation for enterprise sales conversations and product positioning.

Why it matters

Proved a consumer market for empathetic AI when framed around useful outcomes, and established a repeatable research-to-experiment-to-ship-to-learn loop that lifted activation, conversion, and revenue quality.

Consumer AI Assistant Gallery: Scaling PLG and Enterprise Discovery

Open to senior product and technology leadership roles.