ChatAI Pro Capability Discovery

A product analytics investigation into feature discovery, value realization, and subscription conversion in AI products.

What if users do not upgrade because they never discover the feature that would make the product valuable to them?

Fictional product: ChatAI Pro. Synthetic data only. This portfolio artifact demonstrates product analytics reasoning, not actual OpenAI, ChatGPT, or real AI product data.

Opportunity

Help users discover high-value AI workflows earlier

Method

Instrument → Analyze → Prototype → Experiment

Outcome

Product recommendations and experiment roadmap

This case study shows how I translated a product hypothesis into event taxonomy, funnel analysis, prototype concepts, and experiment recommendations. Each stage connects the product decision, the artifact created, and the PM skill demonstrated.

Decision

Frame the value-discovery problem as a measurable product bet.

Artifact

Event taxonomy, prototype screens, funnel readout, and roadmap.

Skill

Product strategy, analytics design, growth thinking, and AI product sense.

The value-discovery gap

As AI products add more capabilities, feature breadth can become invisible to users. A user may sign up, use basic chat, miss advanced capabilities, question the product's value, and fail to upgrade.

The core thesis: AI subscription products may lose users not because the product lacks value, but because users evaluate the subscription before discovering enough high-value use cases.

Decision / Artifact / Skill

Decision

Treat discovery as an activation and conversion problem, not just an onboarding surface.

Artifact

A product investigation focused on discovery cards, premium previews, and re-engagement moments.

Skill

Frames a fuzzy AI product challenge as a testable customer and business hypothesis.

Capability discovery should change activation, conversion, and retention

Users who discover and engage with multiple AI capabilities are more likely to convert, remain active, and retain their subscriptions.

Discovery cards

Surface high-value workflows earlier.

Contextual recommendations

Match capability suggestions to user intent.

Premium previews

Show paid value before the upgrade decision.

Save for later

Capture intent without interrupting the current task.

Proposed solution: Capability Discovery Engine.

Instrument the moments where value becomes visible

Event Taxonomy

account_created first_chat / chat_started discovery_card_viewed discovery_card_clicked premium_preview_viewed image_generated file_uploaded deep_research_attempted plus_upgrade subscription_cancelled

Activation

Discovery card clicked + preview viewed

Conversion

plus_upgrade rate

Engagement

Distinct capabilities used per user

Retention

Return usage after signup / upgrade

Guardrail

Cancellation + prompt dismissal

Prototype screens for earlier value realization

The prototype explores how ChatAI Pro could introduce high-value workflows at the right moment, preview premium value, and re-engage users before perceived value drops.

Prototype artifact: Figma screens showing discovery cards, premium preview, save-for-later, and Day-3 re-engagement.

ChatAI Pro capability discovery prototype overview showing selected Figma screens.

Explore the live prototype for screen-level interaction, or open the original slide-style analytics case study for the synthetic Amplitude findings and roadmap.

New Chat Discovery Cards

Activation: surface high-value workflows early.

Deep Research Premium Preview

Monetization: preview premium value before upgrade.

Save for Later

Intent capture: reduce pressure without losing interest.

Day-3 Re-Engagement

Retention: re-engage users before value perception drops.

Synthetic funnel signals point to discovery as the largest visible bottleneck

158

Synthetic users in latest funnel cohort

72

Users clicked a discovery card

45.6%

Discovery card click rate

61

Reached premium preview and completed Plus upgrade

Key Synthetic Findings

54.4% of users did not click a discovery card, making discovery engagement the largest visible bottleneck.

84.7% of card clickers reached premium preview, suggesting preview access may be a meaningful activation signal.

Once users reached the premium preview, conversion remained very strong in the simulated data. This should be treated as a validation signal, not a proven result.

Churn signals pointed more toward "not enough value" than price, suggesting onboarding and discovery may be better first levers than discounting.

Capability Engagement

Deep Research65 events
File Upload39 events
Image Generation30 events

Deep Research had the highest feature engagement signal and should anchor early value realization experiments.

Caveat: All data in this project is synthetic and intended to demonstrate product analytics reasoning, not actual user behavior from any real AI product. This is directional synthetic analysis, not causal proof.

Prioritize discovery before discounting

P0

Surface discovery in the first session

Place contextual discovery cards above the fold on the new chat screen and trigger within the first two minutes.

P0

Anchor premium value around Deep Research

Use the strongest engagement signal as the default premium preview and onboarding example.

P1

Create a Day-3 value intervention

Target active users who have not clicked a discovery card or attempted a high-value capability.

P1

Add save-for-later to reduce pressure

Let users save a suggestion without interrupting their current task and use it as a re-engagement signal.

P2

Personalize discovery cards

Use query type, device, acquisition source, and prior behavior to recommend the most relevant capability.

Validate discovery quality, not just clicks

Card placement

Above-fold vs inline vs modal

Metric: card click rate
Duration: 14 days

Deep Research intro

Guided first-session intro vs standard onboarding

Metric: Day-7 return + upgrade
Duration: 21 days

Day-3 re-engagement

Triggered value email or in-app prompt vs no intervention

Metric: cancellation rate
Duration: 30 days

Premium preview format

Inline preview vs modal preview

Metric: upgrade rate
Duration: 14 days

Personalized cards

Personalized vs generic card content

Metric: CTR + capability use
Duration: 21 days

Rollout Decision Rule

Ship broadly only if treatment improves upgrade conversion and Day-7 retention without increasing cancellation, prompt dismissal, or user complaints. If conversion rises but guardrails worsen, iterate on frequency caps and personalization before rollout.

Portfolio takeaway

This case study demonstrates product strategy, analytics design, growth thinking, experimentation, and AI product sense.

The strongest product opportunity is not merely adding more features. It is helping users discover the right capability at the right moment, then measuring whether that discovery changes activation, conversion, and retention behavior.

Kelly Connelly · PM portfolio artifact · Fictional product, synthetic data