🔨 Train AI to Identify the Most Valuable Product Research Avenues Next Quarter (Thanks to Reforge)

Cut decision-making time by 40%. Eliminate ineffective research methods.

The single most powerful habit for making better product decisions:

Opening up my notes from the User Insights for Product Decisions course by Reforge.

(Highly recommend you take the whole course. This is just the tip of the iceberg of what’s inside that course.)

Over the last 6 months, I've referred to these notes every single week—and along the way, I've:

• Reduced the time it took to decide how to frame research;

• Done more impactful research, without wasting time on methods that wouldn’t have answered the questions I had.

But everything can be made a bit better.

What I realised was: I still went through a few mental steps every time to decide.

So instead, I turned it into an AI prompt anyone can use:

I am going to train you on the "Reforge User Insights Decision" framework.

First, I will give you a

• Description of the project

• Development Stage

• Evidence Needs

• Splash zone of the decision

Development stage can be either Discover, Design, Develop, Deploy

Evidence Needs map to the stages.

For Discover, there are 3 possible Evidence Needs

• Problem Identification, asking "What problems do users have? Which users have problems?"

• Problem Understanding, asking "Why do users have this problem? What is the scope? How frequent/intense? How do users solve now? Are solutions working? How much do users value?"

• Problem Prioritization, asking "How big is the market for this problem? How does this problem rank compared to others? How does this problem leverage our capabilities?"

For Design, there are 3 possible Evidence Needs

• Solution Ideation, asking "What are possible approaches to solving?"

• Solution Design, asking "What solution features or elements do users value?"

• Solution Validation, asking, "Does my solution design work? What elements work or don't work? Why?"

For Develop, there are 3 possible Evidence Needs

• Solution Usability, asking "Can users do what you want them to do? Do users do what you expect them to do?"

• Solution Discoverability, asking "Can users find the solution in the context of the broader product experience?"

• Solution Value, asking "Does the solution create value for users in the broader product?"

For Deploy, there are 3 possible Evidence Needs

• Solution Performance, asking "How does solution impact key product metrics?"

• Solution Disruption, asking "Are there any unintended consequences to product metrics? Or to the overall experience?"

• Solution Satisfaction, asking "Are users satisfied with the solution? Why or why not?"

Finally, Splash Zone refers to the impact of the decision.

The bigger the splash the change makes, the more evidence needs we have.

• If small, we could do with a small sample size and a simple method.

• If large, we might need several rounds of research, from several angles, with more people.

3 inputs here:

• Work required (Small to Large)

• Internal stakeholders involved (Few to Many)

• User impact (Small to Large)

After I give you these things, you will need to come up with a Decision Description.

Decision Description needs to be framed as "I have a decision to make, and I will go through this piece of research to answer the Evidence Needs questions, and then I will have the answer to the question I have". So it needs to be framed as a question, that when answered with research, can be clearly stated as answered.

BAD EXAMPLE: "Decide if this product is good"

GOOD EXAMPLE: "Decide what things can we do to improve product growth in the India market"

A few rules for the Decision Description:

- Rule #1: Use visceral TANGIBLE language

- Rule #2: Be super specific (specificity clarifies what kind, what happens when/if, what does it mean, so that you can what)

VAGUE: Gain financial freedom

SPECIFIC: Overcome bad spending habits and build positive ones so you can start investing in stocks, real estate, and digital property this month.

Do you understand?

---

Remember: AI is still new, so if you don’t like what you see, run the prompt again or fill it in with your own ideas.

Give it a try and let me know how it goes!

Chat soon,

George

---

PS… If you enjoyed this, will you take 6 seconds and forward this edition to a friend? It goes a long way in helping me grow the newsletter (and help more Product Managers).

And finally, I’d love your input on how I could make these posts even more useful for you! So please leave a comment with:

• Ideas you’d like covered in future editions

• Your biggest takeaway from this edition

I read & reply to every single one of them!