Product discovery and customer validation
Welcome to the Product discovery and customer validation topic.
Here you train and certify the ability to reduce product uncertainty with useful evidence before investing in building: separating the problem from the solution, talking to the right people to learn rather than to confirm, turning conversations into actionable insights, designing the smallest test that challenges the riskiest assumptions, and deciding with judgment whether to persevere, iterate, pivot, or stop. You will see how to sustain this work as a weekly team habit and how to use artificial intelligence to speed it up without delegating judgment or mistaking synthetic responses for evidence. It is a skill that complements Product Owner training, and it is also useful for product leads, designers, developers, Scrum Masters, and analysts who take part in discovery.
The **"Product discovery and customer validation"** academic diplomas certify that their holder has successfully completed specialized training in product discovery: reducing uncertainty about what deserves to be built with useful evidence from customers and the market, from the conversation with the customer to the decision, as a continuous team practice and also with artificial intelligence.
This diploma certifies that the student has acquired the knowledge and judgment needed to discover and validate with customers **with rigor, honesty about the evidence, and independent judgment** in the usual tasks of product discovery: interviewing and recruiting, synthesizing and framing the problem, mapping assumptions and designing experiments, setting success criteria and deciding, sustaining the team's cadence, and using artificial intelligence without delegating judgment.
## Competencias acreditadas
Holding these up-to-date diplomas certifies that the holder is able to:
### 1) Separate problem space from solution space
Tell which statements belong to each space, reframe a request that carries the solution inside it in terms of who, in which situation, what they are trying to achieve, and what gets in their way, and return to the problem framing when a solution fails.
### 2) Interview to learn, not to confirm
Set the learning goal before the script, ask for and follow concrete stories from the past, turn opinion or future-intention questions into behavior questions, detect confirmation bias, and keep discovery separate from pitching a solution.
### 3) Decide who to listen to
Translate a segment into inclusion, exclusion, and variety criteria, write a screener that does not give away the objective, correct own-network and survivorship bias, and size the sample according to whether the goal is to discover patterns or to measure their frequency.
### 4) Capture quality signal
Separate in the notes the quote, the fact with its context, the marked interpretation, and the open question, turn adjective-based notes into concrete facts, close each interview with its snapshot, and treat a tool's summary as an interpretation to be checked against the transcript.
### 5) Handle customer data with care
Identify the personal data processed when recording, transcribing, and analyzing an interview, apply prior notice, a visible indicator, and consent for each recording, check what a tool does with the data, anonymize before using an AI assistant, and recognize when AI as an interviewer or emotion detection triggers transparency obligations or risks of prohibition.
### 6) Synthesize up to the actionable insight
Classify a statement as observation, pattern, insight, or implication, group and count recurrence across different sources, treat exceptions as a possible signal of another segment, and formulate insights with a quantified pattern, a cause marked as inference, and a concrete implication.
### 7) Frame the problem as the job the customer wants to get done
Rewrite a problem stated as a missing feature without the solution inside it, distinguish the school of the job as progress from the school of the job as a measurable outcome, write a job story completed with the current alternative, and check that the framing leaves several solutions open.
### 8) Map and prioritize assumptions
Classify assumptions by desirability, viability, feasibility, usability, ethics, or adaptability, place them on the importance-and-evidence matrix, recognize when the critical assumption is not the usual one, and reword vague assumptions into ones that can be tested.
### 9) Design minimal experiments to learn
Set up prototype tests with tasks as goals, neutral moderation, and observable criteria, choose the minimum fidelity that answers the question, and choose between concierge, Wizard of Oz, landing page, single-feature MVP, or functional prototype according to the hypothesis, with the ethical and legal cautions of what is hidden or generated with AI.
### 10) Set the success criterion before launch
Define the primary metric, threshold, sample, and date before the experiment, distinguish primary, secondary, guardrail, and diagnostic metrics, estimate whether the sample can detect the effect sought, recognize p-hacking and HARKing, and state as directional the evidence that is not causal.
### 11) Weigh the evidence and decide with it
Place a data point on the strength-of-evidence gradient, tell low-cost signals from those that cost something to whoever produces them, mitigate the four common biases, and decide whether to persevere, iterate, pivot, or stop without giving in to the sunk cost fallacy, documenting what was learned.
### 12) Sustain discovery as a team habit
Organize the product trio's weekly cadence with continuous recruiting, bring what is learned into the work of the sprint without creating a handoff, keep a traceable research repository, and build the opportunity solution tree from real interviews, choosing an opportunity before ideating and comparing several solutions.
### 13) Discover with artificial intelligence without delegating judgment
Tell the uses of AI that the evidence supports from those that remain human, apply compared synthesis, decide when an AI can moderate an interview, place synthetic users on the lowest rung of evidence, and, in a product whose core is a model, validate value with acceptable reliability through error analysis.
This diploma certifies that the student has acquired the knowledge and judgment needed to discover and validate with customers **with rigor, honesty about the evidence, and independent judgment** in the usual tasks of product discovery: interviewing and recruiting, synthesizing and framing the problem, mapping assumptions and designing experiments, setting success criteria and deciding, sustaining the team's cadence, and using artificial intelligence without delegating judgment.
## Competencias acreditadas
Holding these up-to-date diplomas certifies that the holder is able to:
### 1) Separate problem space from solution space
Tell which statements belong to each space, reframe a request that carries the solution inside it in terms of who, in which situation, what they are trying to achieve, and what gets in their way, and return to the problem framing when a solution fails.
### 2) Interview to learn, not to confirm
Set the learning goal before the script, ask for and follow concrete stories from the past, turn opinion or future-intention questions into behavior questions, detect confirmation bias, and keep discovery separate from pitching a solution.
### 3) Decide who to listen to
Translate a segment into inclusion, exclusion, and variety criteria, write a screener that does not give away the objective, correct own-network and survivorship bias, and size the sample according to whether the goal is to discover patterns or to measure their frequency.
### 4) Capture quality signal
Separate in the notes the quote, the fact with its context, the marked interpretation, and the open question, turn adjective-based notes into concrete facts, close each interview with its snapshot, and treat a tool's summary as an interpretation to be checked against the transcript.
### 5) Handle customer data with care
Identify the personal data processed when recording, transcribing, and analyzing an interview, apply prior notice, a visible indicator, and consent for each recording, check what a tool does with the data, anonymize before using an AI assistant, and recognize when AI as an interviewer or emotion detection triggers transparency obligations or risks of prohibition.
### 6) Synthesize up to the actionable insight
Classify a statement as observation, pattern, insight, or implication, group and count recurrence across different sources, treat exceptions as a possible signal of another segment, and formulate insights with a quantified pattern, a cause marked as inference, and a concrete implication.
### 7) Frame the problem as the job the customer wants to get done
Rewrite a problem stated as a missing feature without the solution inside it, distinguish the school of the job as progress from the school of the job as a measurable outcome, write a job story completed with the current alternative, and check that the framing leaves several solutions open.
### 8) Map and prioritize assumptions
Classify assumptions by desirability, viability, feasibility, usability, ethics, or adaptability, place them on the importance-and-evidence matrix, recognize when the critical assumption is not the usual one, and reword vague assumptions into ones that can be tested.
### 9) Design minimal experiments to learn
Set up prototype tests with tasks as goals, neutral moderation, and observable criteria, choose the minimum fidelity that answers the question, and choose between concierge, Wizard of Oz, landing page, single-feature MVP, or functional prototype according to the hypothesis, with the ethical and legal cautions of what is hidden or generated with AI.
### 10) Set the success criterion before launch
Define the primary metric, threshold, sample, and date before the experiment, distinguish primary, secondary, guardrail, and diagnostic metrics, estimate whether the sample can detect the effect sought, recognize p-hacking and HARKing, and state as directional the evidence that is not causal.
### 11) Weigh the evidence and decide with it
Place a data point on the strength-of-evidence gradient, tell low-cost signals from those that cost something to whoever produces them, mitigate the four common biases, and decide whether to persevere, iterate, pivot, or stop without giving in to the sunk cost fallacy, documenting what was learned.
### 12) Sustain discovery as a team habit
Organize the product trio's weekly cadence with continuous recruiting, bring what is learned into the work of the sprint without creating a handoff, keep a traceable research repository, and build the opportunity solution tree from real interviews, choosing an opportunity before ideating and comparing several solutions.
### 13) Discover with artificial intelligence without delegating judgment
Tell the uses of AI that the evidence supports from those that remain human, apply compared synthesis, decide when an AI can moderate an interview, place synthetic users on the lowest rung of evidence, and, in a product whose core is a model, validate value with acceptable reliability through error analysis.
Minimum evaluation score to earn or renew the skill: 50 (Qualified 50-64 | Professional 65-79 | Advanced 80-89 | Authority 90-100)
Evaluation for Agile Club members. Sign in or join the club if you are not a member yet.
In your personal area you can manage your assessments and earned diplomas.