SAB Innovation Champions
DAY 4
17 June  |  Opportunity Deepening

Rapid Experiments

Today is about turning promising ideas into evidence. You will take your strongest SAB innovation opportunity and design a simple experiment to test whether it is worth developing further. The goal is not to build the full solution. The goal is to learn quickly, safely and cheaply before investing more time, money or effort.

Start Day 4 Experiment library
Today's outcomes

By the end of today, you will have

A clearer opportunity statement.
A sharper understanding of the riskiest assumption behind your idea.
A selected rapid experiment method.
A simple experiment plan.
A clear success measure.
A practical next step for testing the idea.
Stronger evidence to support your pitch and submission.
Mindset

The mindset for rapid experiments

A rapid experiment is not a mini version of the final solution. It is a focused test designed to answer one important question.

The best experiments are small, safe and specific. They help you learn whether customers, colleagues, partners or stakeholders care enough to take action.

Test the riskiest assumption first.
Keep the experiment small.
Look for behaviour, not opinions.
Decide success before you start.
Learn quickly and adjust.
Protect trust, data and the SAB brand.
Workshop flow

Day 4 flow

1

Reframe the opportunity

Clarify the customer, problem and value behind your idea.

2

Find the riskiest assumption

Identify what must be true for the idea to work.

3

Choose the right rapid experiment

Select the simplest test that gives useful evidence.

4

Design the experiment

Define the method, participants, evidence, success measure and timebox.

5

Prepare the test

Create the script, prototype, landing page, survey, demo or material needed.

6

Plan what happens next

Decide what you will do if the experiment passes, fails or gives mixed evidence.

The basics

What is a rapid experiment?

A rapid experiment is a quick, low-cost and low-risk test that helps you learn whether an idea deserves more investment. It is useful when the team has a promising idea but still needs evidence.

A good rapid experiment answers one question: what do we need to learn before we build more?
Idea  +  Riskiest assumption  +  Small test  +  Evidence  =  Better decision
Step 2

Find the riskiest assumption

Every idea is built on assumptions. Some are minor. Others can break the idea. The riskiest assumption is the thing that must be true for your idea to succeed, but is still uncertain.

Examples

  • Customers will trust AI to support this decision.
  • Branch staff will use the tool during a real customer conversation.
  • Customers will share the information needed.
  • The process can be simplified without adding risk.
  • The value is important enough for people to change behaviour.
  • SAB can deliver the service safely and responsibly.
SAB Bank Challenges

Reconnect to the real challenge

"How might AI enable SAB to reinvent banking through disrupting new products, services and business models?"

Before choosing a rapid experiment, return to the core challenge. These seven SAB Bank Challenges are the opportunity spaces for innovation. Your task is to choose one challenge, identify what must be true for a strong idea to work, and then design a small experiment to collect evidence.

The goal is not to prove the idea is perfect. The goal is to learn quickly, safely and practically.

Seven opportunity spaces

Choose one challenge

1 Challenge 1

Private Banking and High Affluent

SAB Bank's Private Banking and High Affluent customers expect highly personalised, proactive and trusted financial experiences, but traditional banking models can make it difficult to deliver this at scale.

Who it helps: Private Banking and High Affluent customers.
AI opportunity: use AI to support personalised insights, next-best actions, relationship manager preparation, proactive wealth guidance and trusted life-stage financial support.
Rapid experiment angle: test whether customers or relationship managers value a more proactive, AI-supported wealth experience.
Suggested experiments
  • Customer discovery interviews
  • Comprehension test
  • Concierge test
  • Demo pitch
  • Landing page test
2 Challenge 2

Upper Mass and Youth

SAB Bank's Upper Mass and Youth customers are increasingly digital-first, but the bank must find better ways to attract, engage and retain them with simple, relevant and confidence-building experiences.

Who it helps: Upper Mass, youth and emerging affluent customers.
AI opportunity: use AI to create simpler digital guidance, personalised money confidence tools, savings nudges, product matching and more engaging mobile-first experiences.
Rapid experiment angle: test whether younger and digital-first customers understand, trust and engage with a new AI-supported banking experience.
Suggested experiments
  • Fake door test
  • Landing page test
  • Paper prototyping
  • Usability testing
  • A/B testing
3 Challenge 3

SMEs

SAB Bank's SME customers need faster access to finance, clearer cashflow insight and practical business support, but banking services can still feel complex, slow and reactive. There is an opportunity to help SMEs access shared and fractional services such as finance, HR, legal, marketing and operations support, alongside banking products, so they can grow more confidently.

Who it helps: SME owners, founders, finance managers and growing businesses.
AI opportunity: use AI to support cashflow insight, funding readiness, working capital guidance, business health scoring and access to trusted shared or fractional services.
Rapid experiment angle: test whether SMEs would value SAB as both a banking partner and a growth support platform.
Suggested experiments
  • Customer discovery interviews
  • Value proposition flyer
  • Concierge test
  • Landing page test
  • Sales pitch
  • Scorecards
4 Challenge 4

Mid-Market and Large Corporates

SAB Bank's mid-market and large corporate clients need faster, smarter and more valuable banking support across treasury, trade, lending and risk, but current services can be fragmented and manually intensive. There is also an opportunity to help mid-market firms access specialist shared and fractional services that improve finance operations, working capital, compliance, procurement and growth capability.

Who it helps: mid-market companies, large corporates, CFOs, treasury teams, procurement teams and relationship managers.
AI opportunity: use AI to support treasury insight, trade finance workflows, risk monitoring, lending preparation, working capital optimisation and access to specialist shared services.
Rapid experiment angle: test whether corporate clients value an AI-supported advisory, treasury or operational support layer beyond traditional banking products.
Suggested experiments
  • Ask an expert
  • Customer discovery interviews
  • Demo pitch
  • Concierge test
  • Solution interview
  • Dashboard prototype
5 Challenge 5

Investment, Balance Sheet and Capital Optimisation

SAB Bank's investment, balance sheet and capital decisions depend on complex data, but the bank needs better AI-enabled insight to improve forecasting, optimise performance and unlock new financial value.

Who it helps: treasury, finance, investment, risk, strategy and senior decision-makers.
AI opportunity: use AI to improve forecasting, scenario planning, capital allocation, portfolio insight, risk signals and performance optimisation.
Rapid experiment angle: test whether AI-generated insight improves decision speed, confidence or quality in complex financial decisions.
Suggested experiments
  • Data mining
  • Dashboard
  • Ask an expert
  • Scorecards
  • Wizard of Oz
  • A/B testing
6 Challenge 6

Culture, Productivity and Ways of Working

SAB Bank needs to uplift culture, productivity and ways of working so employees can make better decisions, retain talent, improve work-life balance and deliver innovation faster.

Who it helps: SAB employees, managers, leaders, HR, operations and innovation teams.
AI opportunity: use AI to reduce repetitive work, improve decision support, capture knowledge, support wellbeing, improve collaboration and speed up innovation delivery.
Rapid experiment angle: test whether employees adopt and value an AI-enabled tool, workflow or support service that improves how work gets done.
Suggested experiments
  • Open-ended survey
  • Experience sampling
  • Dogfooding
  • Fake door test
  • Usability testing
  • NPS survey
7 Challenge 7

Funding, Marketplaces and Growth Platforms

Access to new opportunities is challenging for customers, SMEs and entrepreneurs, while funders often struggle to identify, assess and match with the right opportunities. SAB Bank has an opportunity to create new funding, marketplace and growth-platform models that connect capital with credible opportunities in a more trusted, efficient and scalable way.

Who it helps: SMEs, entrepreneurs, funders, investors, partners, ecosystem players and growth companies.
AI opportunity: use AI to match capital with opportunities, assess credibility, support due diligence, create trusted marketplaces, connect partners and help customers access funding or growth support.
Rapid experiment angle: test whether customers, funders or ecosystem partners would engage with a trusted SAB-backed platform or marketplace model.
Suggested experiments
  • Landing page test
  • Fake door test
  • Event test
  • Demo pitch
  • Concierge test
  • Pre-sales style commitment test

Activity: Choose your SAB challenge

Choose one SAB Bank Challenge and complete the canvas below before selecting your rapid experiment.

AI Turn my SAB challenge into a rapid experiment
You are an innovation coach helping me turn an SAB banking challenge into a practical rapid experiment.

Overall programme challenge: "How might AI enable SAB to reinvent banking through disrupting new products, services and business models?"

Selected SAB Bank Challenge: [INSERT SELECTED CHALLENGE]
My early idea is: [INSERT IDEA]
The target customer, colleague or stakeholder is: [INSERT TARGET USER]

Help me:
1. Clarify the real problem behind this challenge
2. Identify the strongest AI-enabled opportunity
3. Define the new product, service or business model possibility
4. Identify the riskiest assumption
5. Recommend the best rapid experiment method
6. Explain why this experiment is the right choice
7. Design the experiment in simple steps
8. Define what evidence we need
9. Create a clear success measure
10. Highlight trust, data, compliance or customer experience risks
11. Suggest what to do if the experiment passes
12. Suggest what to do if the experiment fails

Keep the recommendation practical, low-risk and suitable for SAB and the Saudi banking environment.
AI Compare challenge opportunities
You are helping me compare SAB banking innovation opportunities.

Overall challenge: "How might AI enable SAB to reinvent banking through disrupting new products, services and business models?"

Compare these seven challenge areas:
1. Private Banking and High Affluent
2. Upper Mass and Youth
3. SMEs
4. Mid-Market and Large Corporates
5. Investment, Balance Sheet and Capital Optimisation
6. Culture, Productivity and Ways of Working
7. Funding, Marketplaces and Growth Platforms

Score each challenge from 1 to 5 against:
- Customer or stakeholder value
- Strategic fit for SAB
- AI advantage
- Feasibility
- Differentiation
- Revenue or efficiency potential
- Ease of testing
- Trust and compliance readiness

Then recommend the top 3 opportunity areas to explore first.
For each top opportunity, suggest:
- One idea direction
- One riskiest assumption
- One rapid experiment
- One success measure

Keep the output practical, banking-relevant and suitable for a live innovation workshop.
Choose the challenge carefully. The better the challenge, the sharper the experiment. The sharper the experiment, the faster SAB can move from idea to evidence.
Step 3

Rapid Experiment Library

Choose the right experiment for your SAB idea. Filter by what you need to learn, then pick the simplest test that gives useful evidence.

No experiments match this filter.

A. Generative market research

Use these when you need to understand customers, behaviours, problems, needs or market context.

Contextual inquiry

Medium

Best when: you need to observe how the problem happens in the real environment.

Tests: real behaviour, workflow, workarounds and hidden needs.
Run: observe users in context, ask clarifying questions, capture workflow and pain points.
Evidence: observations, quotes, repeated behaviours and real friction points.
SAB example: observe how relationship managers prepare for customer meetings and where manual effort slows them down.
AI prompt

Customer discovery interviews

Low–Med

Best when: you need to understand whether the problem is real and important.

Tests: pain points, current behaviour, alternatives and urgency.
Run: interview relevant users, ask about the last time the problem happened, listen for patterns.
Evidence: customer stories, repeated needs, quotes and problem themes.
SAB example: interview SME customers about the hardest part of accessing financial guidance.
AI prompt

Data mining

Med–High

Best when: you have access to existing data and need to find patterns.

Tests: behavioural patterns, trends, segments and opportunity signals.
Run: identify useful data, extract key fields, look for patterns and draw conclusions.
Evidence: data patterns, usage trends, customer segments and correlations.
SAB example: analyse service request data to identify repeated customer pain points.
AI prompt

Experience sampling

Medium

Best when: you need to capture repeated behaviours or feelings close to the moment they happen.

Tests: real-time behaviour, frequency, triggers and context.
Run: ask selected participants to record short responses when the situation occurs.
Evidence: time-based behaviour logs, patterns and repeated pain points.
SAB example: ask staff to capture moments when customers need help understanding a digital journey.
AI prompt

Focus groups

Medium

Best when: you want to explore perceptions and generate ideas with a small group.

Tests: group perceptions, language, reactions and shared needs.
Run: facilitate a structured group discussion with a clear purpose and neutral questions.
Evidence: group themes, reactions, language and idea triggers (useful for exploration, not proof).
SAB example: run a focus group with young professionals to explore expectations for AI banking support.
AI prompt

Open-ended survey

Low–Med

Best when: you need written feedback from a wider group.

Tests: needs, opinions, language and patterns in free-text responses.
Run: ask a small number of non-leading open questions and analyse recurring themes.
Evidence: written comments, themes and customer language.
SAB example: survey employees on where AI could reduce repetitive work.
AI prompt
B. Evaluative market experiments

Use these when you need to test whether a value proposition, message, offer or market signal creates interest.

Closed-ended survey

Low–Med

Best when: you need structured quantitative feedback.

Tests: preference, priority, satisfaction, segmentation or demand signals.
Run: define the learning goal, write structured questions, collect and analyse responses.
Evidence: scores, rankings, percentages and patterns.
SAB example: ask customers to rank which AI banking support features would be most useful.
AI prompt

Comprehension test

Low

Best when: you need to check whether people understand the idea.

Tests: clarity of the value proposition and message.
Run: show the idea briefly, remove it, ask participants to explain it in their own words.
Evidence: misunderstandings, clearer language and message gaps.
SAB example: test whether customers understand a new AI financial guidance service.
AI prompt

Secondary market research

Medium

Best when: you need existing evidence about the market, trends or competitors.

Tests: market context, trends, size, benchmarks and existing approaches.
Run: search credible sources, summarise findings, identify gaps and implications.
Evidence: market facts, examples, competitor signals and trend insights.
SAB example: research how leading banks are using AI for customer onboarding.
AI prompt

Value proposition: broken promise

Low–Med

Best when: you want to test whether an idea is interesting enough to be shared.

Tests: referral interest and buzz.
Run: share a concept with a small group, ask them not to share, then track whether new people still express interest.
Evidence: referral rate and organic interest.
SAB example: test whether an internal AI productivity idea spreads beyond the first pilot group.
AI prompt

Value proposition: event

Medium

Best when: you want to test interest through attendance or participation.

Tests: demand, relevance and willingness to spend time.
Run: create a small event, invite the target audience, measure sign-ups, attendance and feedback.
Evidence: attendance, engagement, questions and follow-up interest.
SAB example: host a short internal session on AI support for frontline banking teams and measure sign-ups.
AI prompt

Value proposition: fake door

Low

Best when: you want to test real interest before building the solution.

Tests: clicks, sign-ups or intent to learn more.
Run: create a button, card, message or page for the concept, then measure who clicks or registers interest.
Evidence: click-through rate, sign-ups and interest data.
SAB example: add a "Join AI savings coach pilot" button to an internal prototype and track interest.
AI prompt

Value proposition flyer

Low

Best when: you need a simple conversation starter for a concept.

Tests: clarity, interest and willingness to take a next step.
Run: create a one-page flyer with the promise, benefits and call to action, then show it to target users.
Evidence: reactions, questions, objections and sign-ups.
SAB example: create a flyer for a new SME advisory concept and test reactions with relationship managers.
AI prompt

Value proposition: high bar

Low–Med

Best when: you want to test serious interest without charging money.

Tests: commitment, motivation and quality of demand.
Run: add meaningful effort, such as an application form, meeting request or detailed sign-up. Measure completion.
Evidence: high-intent leads and commitment signals.
SAB example: ask teams to complete a detailed application for an AI service pilot and measure serious demand.
AI prompt

Value proposition: landing page

Low–Med

Best when: you need to test demand for a product, service or feature.

Tests: conversion, interest and value proposition strength.
Run: create a simple page explaining the idea and include a clear call to action.
Evidence: visits, clicks, sign-ups and conversion rate.
SAB example: create a landing page for a digital wealth coaching concept and measure sign-ups.
AI prompt

Value proposition: pocket

Low

Best when: you have a simple prototype or mock-up you can show during a conversation.

Tests: initial reaction, understanding and perceived usefulness.
Run: show a simple prototype, sketch or mobile mock-up and discuss the reaction.
Evidence: user reactions, questions, objections and feature expectations.
SAB example: show a mobile mock-up of an AI banking assistant during short customer interviews.
AI prompt

Value proposition: pre-sales

Med–High

Best when: you need strong evidence of willingness to pay or commit.

Tests: real commercial commitment.
Run: offer the future solution for purchase, pre-order, commitment or formal pilot interest.
Evidence: orders, deposits, letters of intent or pilot commitments.
SAB example: for internal use, replace payment with a formal business sponsor commitment or pilot agreement.
AI prompt

Value proposition: sales pitch

Medium

Best when: you need to test whether stakeholders or customers will take a next step.

Tests: pain, value, objections and commitment.
Run: pitch the idea, listen to objections, ask for a clear next step.
Evidence: objections, commitments, meeting requests and decision signals.
SAB example: pitch a new customer onboarding improvement to a business owner and ask for pilot sponsorship.
AI prompt

Value proposition: video

Medium

Best when: you need to explain a concept quickly and test excitement.

Tests: understanding, interest, sharing and emotional response.
Run: create a short concept video, share it with target users and track response.
Evidence: views, shares, comments, sign-ups and qualitative feedback.
SAB example: create a short video explaining an AI-powered banking journey and test internal reactions.
AI prompt
C. Generative product research

Use these when you need to explore possible solutions, competitors, user behaviour or product design options.

Analog / digital test

Low–Med

Best when: you want to test a digital idea manually or physically first.

Tests: workflow, value and feasibility before building technology.
Run: create a manual or paper-based version of the service and test it with users.
Evidence: workflow insights, user reactions and design requirements.
SAB example: test an AI recommendation journey using a human-prepared recommendation sheet.
AI prompt

Ask an expert

Low

Best when: you need specialist insight before testing.

Tests: expert assumptions, risks, constraints and options.
Run: identify relevant experts, prepare questions, capture advice and convert it into testable assumptions.
Evidence: expert insight and better hypotheses.
SAB example: speak to compliance, data, risk or frontline experts before testing an AI feature.
AI prompt

Competitor analysis

Medium

Best when: you need to understand existing solutions and differentiation.

Tests: market landscape, alternatives and positioning.
Run: research competitors, substitutes and adjacent examples. Compare features, value and gaps.
Evidence: competitor map, differentiation and opportunity gaps.
SAB example: compare AI personal finance features from leading regional and global banks.
AI prompt

Competitor usability testing

Medium

Best when: you want to learn from how users interact with existing alternatives.

Tests: user expectations, frustrations and improvement opportunities.
Run: ask users to complete tasks using competitor or substitute solutions and observe their behaviour.
Evidence: usability issues, expectations and feature ideas.
SAB example: observe users completing onboarding in another banking app and note friction points.
AI prompt

Concierge test

Medium

Best when: you want to test a service manually before automating it.

Tests: whether the service creates value when delivered by humans.
Run: deliver the service manually to a small group and capture feedback.
Evidence: customer value, workflow needs and minimum feature requirements.
SAB example: manually provide AI-style financial guidance to a small internal test group before building automation.
AI prompt

Demo pitch

Low–Med

Best when: you need to test whether a demonstrated solution creates commitment.

Tests: interest, objections and willingness to take the next step.
Run: prepare a simple demo, pitch it to target users or stakeholders, and ask for a next step.
Evidence: reactions, objections and commitments.
SAB example: demo a prototype of a staff AI assistant to branch managers and ask for pilot interest.
AI prompt

Dogfooding

Low

Best when: your team can use the idea themselves.

Tests: internal usability, practical value and workflow issues.
Run: use the prototype or process in your own work and capture what works and what fails.
Evidence: first-hand observations and improvement opportunities.
SAB example: the team uses its own AI workflow tool for one day and logs pain points.
AI prompt

Picnic in the graveyard

Medium

Best when: similar ideas have failed before and you need to learn why.

Tests: past failure patterns, risks and avoidable mistakes.
Run: research failed or discontinued solutions, interview users if possible, and identify what to avoid.
Evidence: failure insights, risk factors and differentiation ideas.
SAB example: study failed digital wallet or AI banking initiatives to identify trust, adoption or usability issues.
AI prompt

Solution interview

Low–Med

Best when: you have a clear problem and want feedback on a proposed solution.

Tests: whether the proposed solution addresses the problem.
Run: confirm the user has the problem, show the solution or prototype, then gather reactions and requirements.
Evidence: feature needs, objections and value signals.
SAB example: show a concept for AI-supported SME cash flow advice and ask users what would make it useful.
AI prompt
D. Evaluative product experiments

Use these when you need to test usability, traction, product experience, feature performance or evidence of fit.

Dashboards

Medium

Best when: you need to track progress or performance over time.

Tests: whether key metrics show improvement or traction.
Run: define metrics, create a simple dashboard and update it regularly.
Evidence: trend data, progress and performance signals.
SAB example: track pilot usage, satisfaction and time saved for an AI staff assistant.
AI prompt

Net promoter score survey

Low

Best when: you need a simple loyalty or recommendation signal.

Tests: satisfaction, recommendation likelihood and reasons behind the score.
Run: ask "how likely are you to recommend this?" and follow up with why.
Evidence: NPS score, comments and improvement themes.
SAB example: ask pilot users whether they would recommend the new service to colleagues.
AI prompt

Paper prototyping

Low

Best when: you need to test screens, flows or service journeys quickly.

Tests: usability, comprehension and journey logic.
Run: sketch the journey or screens on paper, ask users to complete tasks and observe confusion.
Evidence: usability issues, missing steps and design improvements.
SAB example: test a paper version of an AI loan guidance flow before building it digitally.
AI prompt

Product / market fit survey

Low–Med

Best when: users have already experienced the core idea and you need to measure value.

Tests: how disappointed users would be if the solution disappeared.
Run: ask users how they would feel if they could no longer use the product or service.
Evidence: product-market fit signal and user value feedback.
SAB example: ask pilot users how they would feel if the AI support tool was removed.
AI prompt

Usability testing

Low–Med

Best when: you need to know whether people can use the solution.

Tests: task completion, confusion, frustration and usability problems.
Run: give users tasks, ask them to think aloud, observe and record issues. Start with five users.
Evidence: task success, usability issues and improvement actions.
SAB example: test whether customers can complete a new digital banking journey without support.
AI prompt

Wizard of Oz

Med–High

Best when: you want to test an AI-enabled or automated experience before building the technology.

Tests: whether users value the experience when it appears automated.
Run: create the front-end experience and manually perform the back-end work behind the scenes.
Evidence: usage behaviour, value signals and workflow requirements.
SAB example: simulate an AI banking assistant manually before building the actual AI system.
AI prompt

A/B testing

Medium

Best when: you need to compare two versions of one element.

Tests: which message, design or option performs better.
Run: create version A and version B, show both to comparable users, and measure one clear metric.
Evidence: comparative performance data.
SAB example: test two messages for a new savings feature and compare click-through rates.
AI prompt

Off-brand testing

Low–Med

Best when: you need to test the idea without the influence of the SAB brand.

Tests: true interest without brand trust or bias.
Run: present the idea without brand signals and measure response.
Evidence: unbiased demand or interest signals.
SAB example: test whether customers respond to a new financial coaching concept without using SAB branding.
AI prompt

Scorecards

Low

Best when: you need to prioritise between ideas, assumptions or experiments.

Tests: relative strength based on agreed criteria.
Run: choose criteria, score each option, weight the results and rank priorities.
Evidence: clear decision ranking and rationale.
SAB example: score ideas by customer value, feasibility, risk, strategic fit and evidence strength.
AI prompt
Start here

Recommended experiments for SAB Day 4

If you are unsure where to begin, these ten work well for banking ideas and fit inside the day.

Customer discovery interviews

Best for testing whether the problem is real and important.

Comprehension test

Best for checking whether people understand the idea and value proposition.

Fake door test

Best for testing whether customers or colleagues show real interest before anything is built.

Landing page test

Best for testing interest in a new digital banking product, service or feature.

Paper prototyping

Best for quickly testing screens, journeys, flows or service concepts.

Concierge test

Best for testing a service manually before building automation or AI.

Wizard of Oz test

Best for testing an AI-enabled experience before the technology is fully built.

Demo pitch

Best for testing whether stakeholders, customers or partners will take a next step.

Usability testing

Best for testing whether people can actually use the proposed solution.

Scorecards

Best for helping teams choose the strongest idea, assumption or experiment.

Decision helper

How to choose the right experiment

Start from what you most need to learn, then pick one method from that group.

You do not understand the customer yet
Contextual inquiryCustomer discovery interviewsOpen-ended survey
You need to test whether people care about the idea
Fake doorLanding pageEventFlyer
You need to test whether people understand the idea
Comprehension testDemo pitchVideo
You need to test whether the solution could work
Paper prototypingConciergeWizard of OzSolution interview
You need to test usability
Usability testingCompetitor usability testingA/B testing
You need to choose what to focus on
ScorecardsDashboardsData mining
Step 4

The Rapid Experiment Canvas

Complete the canvas for your chosen experiment, then copy it into your notes or an AI tool.

AI prompts for Day 4

Use AI to design and run your experiment

Copy a prompt, paste it into your AI tool, and fill in the brackets. Keep your inputs public or non-sensitive.

1 Choose my rapid experiment
You are an innovation coach helping me design a rapid experiment for a banking innovation idea.

My idea is: [INSERT IDEA]
My target user or customer is: [INSERT USER]
The problem I think we are solving is: [INSERT PROBLEM]
The value proposition is: [INSERT VALUE PROPOSITION]

Help me:
1. Identify the riskiest assumption
2. Choose the best rapid experiment method from the SAB Day 4 Rapid Experiment Library
3. Explain why this experiment is the right choice
4. Design the experiment in simple steps
5. Define the evidence we need
6. Define a clear success measure
7. Suggest how we can run it in 5 working days or less
8. Suggest how AI could help us prepare, run or analyse the experiment

Keep the recommendation practical, low-risk and suitable for a bank.
2 Build my experiment plan
You are helping me turn my SAB Innovation Champions idea into a practical rapid experiment.

Use this structure:
Idea: [INSERT IDEA]
Riskiest assumption: [INSERT ASSUMPTION]
Chosen experiment: [INSERT EXPERIMENT METHOD]

Create a one-page experiment plan with:
1. Experiment title
2. Learning goal
3. Hypothesis
4. Target participant or user
5. Method
6. Step-by-step plan
7. Questions or tasks
8. Evidence to collect
9. Success metric
10. Timebox
11. Risks and ethical considerations
12. What we will do next if the evidence is positive
13. What we will do next if the evidence is negative

Make it simple, realistic and suitable for SAB.
3 Write my test script
You are helping me prepare a rapid experiment script for a SAB banking innovation idea.

My idea is: [INSERT IDEA]
My chosen experiment method is: [INSERT METHOD]
My target participant is: [INSERT PARTICIPANT]
My learning goal is: [INSERT LEARNING GOAL]

Create:
1. A short introduction script
2. Five non-leading questions or tasks
3. Follow-up questions to understand behaviour
4. A simple note-taking template
5. A closing script
6. Guidance on what evidence to capture

Keep it professional, clear and suitable for a banking environment.
4 Analyse my experiment results
You are helping me analyse the results of a rapid experiment.

Experiment method: [INSERT METHOD]
Hypothesis: [INSERT HYPOTHESIS]
Success measure: [INSERT SUCCESS MEASURE]
Evidence collected: [PASTE NOTES, DATA OR FEEDBACK]

Help me:
1. Summarise the key evidence
2. Identify patterns
3. Separate facts from opinions
4. Decide whether the experiment passed, failed or needs more testing
5. Explain what we learned
6. Recommend the next step
7. Suggest how to present this evidence in a pitch

Be honest. Do not overclaim. Highlight uncertainty.
Stay safe

Experiment ethics for banking

Rapid experiments are powerful, but in a bank they must protect customers, data and trust. Keep these in mind for every test.

  • Do not mislead customers in a way that damages trust.
  • Do not collect unnecessary personal data.
  • Do not use confidential, sensitive or regulated information.
  • Make internal tests safe and transparent.
  • Keep experiments small before scaling.
  • Use evidence to learn, not to prove you were right.
  • Ask for guidance when compliance, legal, data or customer risk is involved.
  • Be careful with fake door, off-brand and Wizard of Oz tests.
  • Make sure all tests protect customer trust and SAB's reputation.
Before you leave today

Day 4 action checklist

Tick each item as you complete it. Progress is saved on this device for the session.

0% complete
  • We have selected one idea to test.
  • We have written the riskiest assumption.
  • We have chosen one rapid experiment method.
  • We have defined the target user or stakeholder.
  • We have written the success measure.
  • We know what evidence we need.
  • We know what we will build or prepare.
  • We have a 5-day testing plan.
  • We know what we will do if the test passes.
  • We know what we will do if the test fails.
Quick access

Move across the programme

Back to Day 2 Guide Back to Day 3 Guide Day 5 Guide
✓ Copied