Artificial Intelligence

How to use AI for RICE and Assumption Testing?

AI for Product Management - prioritization

BLOG 4: Prioritization Without “Whoever Shouts Loudest”

“My feature is the most urgent.”
“Our customer is asking for this.”
“Sales says this has to go first.”
“Management wants this exact thing.”

A Product Manager rarely suffers from a lack of priorities. They suffer from having too many reasons why everything is a priority.

That’s why frameworks like RICE can help. But only if we use them as a tool to structure the conversation, rather than as a decision-making machine.

RICE Is a Framework, Not a Verdict

The formula is simple: RICE = (Reach × Impact × Confidence) / Effort

AI can very quickly:

  • calculate the score;
  • rank the initiatives;
  • explain the difference;
  • run a scenario analysis

But there is one thing it cannot do: it cannot turn bad estimates into objective truth.

If you got Reach, Impact, Confidence, or Effort wrong – you’ll get a very precise wrong answer.

That’s why a much more interesting question is: How stable is our priority?

Proposed prompt to get started

Role: You are an analytical Product Manager helping a team structure prioritization decisions. Use RICE as a consistent way to compare initiatives. Do not make the final product decision.
Formula: RICE = (Reach × Impact × Confidence) / Effort
Use the same unit for Effort across all initiatives. Do not invent missing values.
For each initiative, I have:

  • Name;
  • Reach;
  • Impact;
  • Confidence;
  • Effort;
  • basis for the estimate.

Calculate RICE and rank the initiatives.
Then:

  • explain the top 2–3 initiatives;
  • highlight the biggest assumptions;
  • assess how stable the ranking is;
  • show what would change if key estimates shifted;
  • list up to 3 factors that RICE doesn’t see;
  • list up to 3 questions the Product Manager needs to clarify.

Do not present RICE as an objective or final decision.

The Question Worth More Than the RICE Score Itself

Suppose Initiative A comes out on top. Don’t stop there. Ask AI: “What would have to be wrong in our assumptions for Initiative B to actually be the better choice?” This is much more useful.

Suddenly, the conversation is no longer: “A has a RICE score of 184, B has 142.”
Instead, it becomes: “If our Reach for A is overestimated, does the priority change?”

That is a real product conversation.

What Doesn’t RICE See?

RICE can be helpful, but it doesn’t capture everything. For example:

  • strategic goals;
  • regulatory requirements;
  • customer commitments;
  • revenue/risk factors;
  • dependencies on other initiatives;
  • cost of delay.

That’s why the framework should be used as one input into the decision, not as the decision itself.

Try it today

Take three real initiatives. Enter your estimates. Calculate RICE.

Then do something more interesting:
Ask AI which single change in an assumption would be most likely to change the ranking.

If the order shifts easily – you might not have enough certainty to act as though the priority is obvious.

BLOG 3: How to Use AI for Acceptance Criteria

Once your Epic has been broken down into meaningful vertical slices, the next challenge is often even more concrete: is each individual User Story clear enough for the team to understand, estimate and refine without endless questions mid-sprint? This is where AI can act as another pair of eyes, and that’s what we cover in the next part of the series.

BLOG 5: How to Use AI for Better Communication

Once you make a product decision, the job isn’t done. You have to explain it to different people. The development team needs one context. Management needs another. The user needs a third.
Yet the core information remains the same.