Superhuman · AI product · Mid · 45 to 60 minutes

Where should AI fit in Superhuman email?

Superhuman is a premium email client built around helping people get through email much faster. Answer this product management case and get it graded line by line against a weighted rubric.

product sense
prioritization
AI product thinking
metric design
execution planning

The situation

You are the PM for a core product area at Superhuman. Leadership wants a recommendation on whether, and where, to add new AI capabilities to the email experience. Interest in generative AI is rising quickly, and there is pressure to respond while the moment is hot. At the same time, Superhuman users pay for speed, focus, and polish, not for novelty. A poorly chosen AI feature could slow common actions, create trust problems, or clutter a product that users value for feeling fast and crisp. The CEO asks you to propose a focused plan for the next build cycle. Assume you have one quarter and a small cross functional team. You need to decide what user problem to target first, how narrow or broad the initial scope should be, and how you would evaluate quality when AI output can be inconsistent. You also need to explain how the team should work through prompt design and iteration, since model behavior will shape the product experience as much as the UI does. Your recommendation should help the company move quickly without drifting away from its core promise.

What to produce

Write a 1 to 2 page product strategy memo.

Your answer must

Identify the primary user problem to solve and explain why it fits Superhuman's core promise better than other AI opportunities.
Recommend one initial AI use case or a very small set of closely related use cases, and explicitly state at least two attractive ideas you would not do first.
Define success metrics that cover speed, user value, and output quality or trust.
Outline an MVP scope, including where the feature appears in the email workflow and what the user experience should do when AI is slow, uncertain, or wrong.
Describe how the team will iterate after launch, including a practical approach to prompt testing, feedback collection, and decision making.

Constraints

The first release must fit within one quarter with a small team.
The product cannot make common email actions feel slower for core users.
Do not assume exclusive model access, a proprietary training dataset, or major changes to Superhuman's business model.
Stay within Superhuman's existing email product experience rather than proposing a separate new product.

Format

Use five sections: user problem, chosen use case, MVP scope, metrics, and launch plus iteration plan.

How Superhuman actually played it

Superhuman's CEO tells how the team built AI features for its email client in three months, with the aim of making an already fast product faster. The post covers the product choices and prompt work behind the launch. Useful for PMs adding AI to a speed-focused product.

Anchor AI work to the product's core promise
Move fast with a short build window and tight scope
Iterate on prompts as a product activity

Built from Blood, sweat, and prompts: How we built Superhuman AI by Rahul Vohra.