OpenAI · AI product · Mid · 45 to 60 minutes

Where Should ChatGPT Improve First?

OpenAI builds AI systems and products, including ChatGPT, that help people generate and work with text, code, images, and more. Answer this product management case and get it graded line by line against a weighted rubric.

Segmentation
Product strategy
Journey mapping
Prioritization
Metrics

The situation

ChatGPT is used by many kinds of people, from students to professionals, but that breadth has made product planning harder rather than easier. A Director of Product at OpenAI asks you, a PM on the ChatGPT team, to recommend where the team should focus its next round of improvements. Leadership does not want a generic list of feature requests. They want a clear view on which user segment matters most, where in that segment's workflow ChatGPT creates or misses value today, and what product bets are worth making first. Why now: as ChatGPT matures, the team needs sharper prioritization for retention, user value, and willingness to pay. The hard part is that several segments can make a plausible case, and broad prompts like improve ChatGPT often lead to shallow answers. You need to narrow the problem, show strategic judgment, and still get concrete about product changes, success metrics, and trade-offs. Assume you have access to standard product analytics, recent user research, and support themes. Your recommendation should help the team decide what to build next in ChatGPT itself, not in a separate service.

What to produce

A 1 to 2 page product strategy memo.

Your answer must

Choose one primary user segment for ChatGPT and justify the choice with explicit selection criteria, such as frequency of use, pain intensity, and willingness to pay.
Map the chosen segment's end to end workflow or job to be done, and identify 2 to 3 moments where ChatGPT creates value, loses value, or could better capture value.
Recommend up to 3 product improvements for ChatGPT, prioritized in order, with rationale and key trade-offs.
Define success metrics for the top recommendation, including at least 1 leading indicator and 1 outcome metric.
Call out the main risks and unknowns, and explain how you would validate them before a broader rollout.

Constraints

Stay within ChatGPT as a product, do not rely on a separate services business or custom human support.
Do not assume a breakthrough in model quality, latency, or cost.
Choose one primary segment only, you may mention secondary segments briefly but do not optimize for everyone.

Format

Use headings for segment choice, workflow, opportunities, recommendation, metrics, and risks.

How OpenAI actually played it

Walks through a full answer to the broad question of improving ChatGPT. The author narrows to software engineers as the segment, maps their workflow end to end, and finds where value is created and captured before proposing features.

Broad prompts need a segment choice early, backed by usage frequency and willingness to pay
Map the whole workflow to find the highest value gap
Balance strategy talk with concrete product craft

Built from OpenAI PM Interview: How Would You Improve ChatGPT? by Ajitesh, Tough Tongue AI (PM Interview Prep Weekly).