Capsule: AI Expense Tracker

The customer problem

Capsule's official website describes receipt capture, grouping and export. Editorial interpretation: the central problem is recovering a useful record after a purchase, when the paper receipt might otherwise disappear. Study the whole handoff rather than only the camera screen. A person may successfully scan a receipt and still struggle to find it when an employer asks for an explanation. A practical interview should follow a recent expense from payment through reconciliation and identify the first point where information was lost.

Core workflow

1. Capture a receipt and review the extracted merchant, amount and category against the image.
2. Group the checked record with its trip, project or client so it can be retrieved later.
3. Export the selected records for an employer, bookkeeper or personal review, checking totals before submission.

What customers pay for

Editorial analysis: measure correctly completed expense records that users can retrieve and explain later. Installation counts and the number of images uploaded are weaker signals because neither guarantees a usable export. A reasonable trial would ask participants to capture a small set of ordinary receipts, correct uncertain fields, and assemble a report a week later. Compare that process with their existing folder or spreadsheet, recording total effort and unresolved discrepancies. For monetization, distinguish occasional users with a single trip from people who submit expenses repeatedly. The historical launch receipt is an observation of initial purchasing, while renewal requires a separate cohort and a later observation.

Distribution signals

Observed channels and entry points; their performance is not inferred.

• Personal network and word of mouth

  The founder's replies attribute initial downloads to his network and the first organic customer to word of mouth. That is a launch account, not measured channel attribution.

  Evidence ↗