Strategy question

Deepgram bundles STT+LLM+TTS at $0.08/min. How would you decide on this bundled pricing?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Tests pricing strategy for a bundled product, weighing simplicity for customers against margin risk from variable underlying costs.

Identify why bundling makes sense here: a customer building a voice agent needs all three components together, and pricing them separately adds integration friction and cost uncertainty.
Check the bundle's cost structure: STT, the LLM call, and TTS each have different and independently variable costs, so a flat per minute price must be built from a realistic blended average, not a guess.
Stress test the blend against usage variance: a customer with longer LLM generations per minute of audio costs more to serve than one with short responses, so pricing must not lose money on the heavy tail.
Compare to unbundled pricing: keep a usage based unbundled option available for sophisticated customers who want to swap in their own LLM, protecting margin on that segment separately.

Explicitly models the three cost components separately before blending them into one price, rather than picking a flat number without justification.
Flags the heavy tail risk directly: customers whose usage skews toward long LLM responses per minute of audio could erode margin under a flat blended rate.
Proposes keeping an unbundled, usage based option alongside the flat bundle, protecting margin from customers who would otherwise select against the flat price.

Treating 0.08 dollars per minute as fixed and given rather than reasoning about the cost components that must justify it.
Ignoring adverse selection, where the heaviest LLM users are drawn to the flat bundle while light users opt out.

How would you detect that certain customers are unprofitable under this flat price?
Would you offer a separate tier for customers who want to bring their own LLM?