Product Metrics with AI: the HEART Framework

Product Metrics with AI: the HEART Framework

Key answer

The HEART framework measures product health across five dimensions, Happiness, Engagement, Adoption, Retention, and Task success, and AI scores each, explains the movement, and refreshes it on a 30/60/90 cadence. AI instruments and narrates; the PM picks the few signals that matter and decides the response.

The HEART framework measures product health across five dimensions, Happiness, Engagement, Adoption, Retention, and Task success, and AI scores each, explains the movement, and refreshes it on a 30/60/90 cadence. AI instruments and narrates; the PM picks the few signals that matter and decides the response. The point is a balanced read on health, not a single vanity number.

Your HEART scorecard#

Score each dimension, read the colours, and refresh after a release.

Your HEART scorecard, live

Happiness (CSAT)
72/100
Engagement
64/100
Adoption
81/100
Retention
49/100
Task success
76/100
Referral (NPS)
58/100
Green  Amber  Red
Each HEART dimension scored Red, Amber, Green. Refresh to model a release; red means dig in before you scale.

Red means a dimension is below its healthy range and needs a look before you scale, amber means watch, green means hold. The discipline is to pick the few dimensions that matter for this product, not to chase all five equally.

Why product reporting is automating#

of descriptive and diagnostic analytics will be automated by 2027, product reporting included

90% of descriptive and diagnostic analyticswill be automated by 2027, product Gartner, Autonomous Finance

Gartner expects 90% of descriptive and diagnostic analytics to be automated by 2027, product reporting included, and 75% of new analytics content to be contextualised through GenAI by 2027. Benchmarks give a red score meaning: median product retention is about 39% at month one and 30% by month three, and 7% of users returning on day 7 puts a product in the top quartile. The PM moves from building dashboards to deciding what a moved metric means. The wider lifecycle is in the GenAI in Product Management guide.

What good product retention looks like

39%Month onemedian user retention one month after first use.30%Month threemedian retention by month three (top 10% near1.9x).7%Day-7 rulereturning on day 7 puts a product in the topquartile.

Benchmark a HEART Retention score before calling it red. Sources: Pendo (2025); Amplitude.

The five HEART dimensions#

The five HEART dimensions

HappinessSatisfaction and perceivedvalue.EngagementDepth and frequency of use.AdoptionNew users reaching first value.RetentionUsers who come back and stay.

What each one actually measures.

Happiness, Engagement, Adoption, Retention, and Task success each measure a different facet of health. Pick the few that fit your product and stage; an early product weights Adoption, a mature one Retention.

Where AI helps, and where you decide#

Where AI helps, and where you decide

InstrumentWire the metric to the data.ExplainNarrate why a dimension moved.DiagnoseSurface the likely driver.DecideThe PM picks the few thatmatter.

It instruments and explains; you pick the signal.

AI instruments the metric, explains the movement, and surfaces the likely driver; the PM picks the signals that matter and decides. The narrative layer that turns this into a story is data storytelling with AI.

Instrument HEART on your product#

Practical GenAI in Product Management instruments HEART metrics and a 30/60/90 plan on your own product in Session 4. You leave with a health read the team trusts.

Key takeaways

  • HEART measures product health across Happiness, Engagement, Adoption, Retention, Task success.
  • AI scores each dimension, explains the movement, and refreshes it on a 30/60/90 cadence.
  • Pick the few HEART signals that matter for this product, not all five at once.
  • AI instruments and narrates; the PM decides the response.

Questions, answered

What is the HEART framework?
HEART is a framework for measuring product health across five dimensions: Happiness, Engagement, Adoption, Retention, and Task success. It gives a balanced view rather than a single vanity metric. AI helps by instrumenting each dimension, scoring it, and explaining the movement on a regular cadence.
Do I need to track all five dimensions?
No. HEART is a menu, not a mandate. Pick the few dimensions that matter for this product and stage, an early product may care most about Adoption and Task success, a mature one about Retention. Tracking all five with equal weight dilutes focus. The PM chooses; AI instruments whatever you pick.
How does AI help with product metrics?
It wires each metric to the data, scores it Red, Amber, Green, explains why a dimension moved, and refreshes the read on a 30/60/90 cadence. Gartner expects product reporting, like other diagnostic analytics, to be largely automated by 2027. The PM moves from building dashboards to deciding the response.
What is a good product retention rate?
Median software products retain about 39% of users one month after first use and roughly 30% by month three, with the top 10% retaining near 1.9 times the average (Pendo, 2025). Amplitude's 7% rule holds that 7% of users returning on day 7 puts a product in the top quartile. Use these as the reference point when a HEART Retention score reads red.
What does a red HEART dimension mean?
It means that aspect of product health is below its healthy range, falling retention, for example, and needs investigation before you scale. Red is a prompt to diagnose the driver, not a verdict. Scoring continuously is what turns a quarterly surprise into an early signal you can still act on.
AE

Dr. Ahmed El-Shamy

Co-founder, CEO and Dean of Education, Digisoul

Dr. Ahmed El-Shamy is Co-founder, CEO and Dean of Education at Digisoul. He has more than a decade across AI, fraud risk, and FP&A, and teaches Practical GenAI in FP&A bilingually across MENA, the GCC, and Africa, governed by Digisoul's ISO/IEC 42001:2023-certified AI Management System. Read the leadership profile.

Sources

  1. Gartner · by 2027, 90% of descriptive and diagnostic analytics in finance will be automated (2023 prediction). https://www.gartner.com/en/newsroom/press-releases/2023-03-01-gartner-preditcts-three-ways-autonomous-technologies-will-impact-the-fpanda-and-controller-functions-in-
  2. Gartner · 75% of new analytics content to use GenAI for contextual intelligence by 2027 (Jun 2025). https://www.gartner.com/en/newsroom/press-releases/2025-06-18-gartner-predicts-75-percent-of-analytics-content-to-use-genai-for-enhanced-contextual-intelligence-by-2027
  3. Pendo · 2025 retention benchmarks (median 39% month-one, ~30% month-three). https://www.pendo.io/pendo-blog/user-retention-rate-benchmarks/
  4. Amplitude · The 7% retention rule (day-7 return = top quartile activation). https://amplitude.com/blog/7-percent-retention-rule
  5. Rodden, Hutchinson & Fu · Measuring the User Experience on a Large Scale (HEART), Google, ACM CHI 2010. https://research.google/pubs/measuring-the-user-experience-on-a-large-scale-user-centered-metrics-for-web-applications/
  6. Practical GenAI in Product Management (Session 4: HEART metrics + 30/60/90). https://digisoul.io/ai4x/genai-in-product-management/

AI Agent · Built on Claude · Operated on Zoho One


What do you think?

From our blog

Articles & insights

  • Product Management
AI Product Experimentation: Test with Rigor
Design product experiments with proper power and read them with rigor using AI. A practical 2026 method so you ship what works and kill what
Read more
  • Product Management
The AI-Ready PRD: the PRISM Framework
An AI-ready PRD ties every requirement to evidence and a success metric. How the PRISM framework structures a PRD that AI can draft and a
Read more
  • Product Management
AI Product Discovery with Jobs-to-Be-Done
Synthesise user interviews into jobs-to-be-done with AI, fast, then prioritise. A practical 2026 discovery method that stays grounded in real evidence.
Read more