You are defining your AI product opportunity wrong. Here's how:

AI-enabled services and products have redefined new user possibilities.

New innovations. New user possibilities.

But also, the rise of new user problems.

After all, AI is just a tool.

It's a whole new design medium and canvas for us to innovate.

Build new products. Solve user points through AI capabilities (that were previously not possible). Leverage AI to benefit society and users for good

But…

What exactly can AI do?

How do we design solutions for something we don't know?

AI is complex, and a whole new design world. It can be intimidating to work with AI as a Designer, due to the complexity of AI technology.

1. AI capabilities are still uncertain — we don't know what AI can do, and it can be a blue sky field of possibilities and limitations
2. AI output is complex and can be infinite based on contextual scenarios (etc Tesla, Siri)

In Pt 1 of the Human-Centered AI series, I'll break down how we can approach design with AI through a 5-step framework.

First off, what is the Human-Centered AI methodology?

It's a unique Human-Centered AI methodology I have spent 2+ years of research on, developed in a prior research collaboration with AI Singapore.

Essentially the new Design Thinking for AI. Read why we need a new product design approach for AI here.

It fuses business, technology, and design for ethical AI Product Innovation through 5 steps.

1. Define

Define and identify business opportunities, user painpoints, and areas where AI can add value

2. Align

Align business and user needs to achievable data needs and AI inputs

3. Ideate

Brainstorm and generate ideas for new design possibilities enabled by AI capabilities

4. Explain

Explain inner workings of the AI model, and communicate what AI does for users to promote user understanding and trust

5. Impact

Consider the impact of AI solutions across different aspects of society in unintended consequences

🔈R I#39m running a series of Human-Centered AI Design Sprint workshops at the upcoming UX STRAT APAC Conference in Bangkok on 19–20 February.

👉🏻 Register here at: https://strat.events/asia/program

In Part 1 of today's article, we will go delve in depth into Step 1: Defining the AI opportunity.

Before we dive in, let's first understand how AI works.

To identify AI-enabled opportunities, we first need to understand how an AI model and life cycle works. On a high level.

A quick dummy's guide to how AI models work

AI model stages can be generally classified into 4 stages:

1. Data Preparation: Gathering and organizing relevant data.
2. AI Modeling: Building the model using algorithms and data.
3. Simulation & Test: Evaluating the model's performance and accuracy.
4. Deployment: Implementing the model in real-world applications.

This is the big picture. Think of how to get AI outputs, you need to input the AI model with data it can learn and train from. And what are tangible AI outputs that can be achieved from the model.

Amid the training process, it's going to make errors and learn from feedback. You need to design for error scenarios, how to help users know what feedback to implement

The big picture of AI model design

We can draw parallels between AI model development and user design thinking:

1. Identify user needs → Determine how AI can help
2. Collect Data → Train Models (input the AI model with user data)
3. Design to meet user's mental model → Test prototypes
4. Explain AI → Build trustworthy and explainable AI UX for users
5. Incorporate feedback → Handle failures gracefully

And this can be categorised into 3 core facets:

1. Identify — Identify if AI adds unique value (User needs + Defining AI Success)
2. Translate — Translate user needs to attainable data inputs (Data collection + Evaluation)
3. Explain — Explain how AI WORKS + manage users expectations (Explainability + Trust)

Now let's link this back to our original Human-Centered AI methodology.

Step 1: Identify AI-enabled Opportunities

This is where you identify current points and potential areas where AI can provide value.

Instead of asking "Can we use AI to solve XX" or "How might we solve XX", ask — "Can AI solve this problem in a unique way?

A common fallacy I see is businesses jumping off the AI hype train to build an AI-enabled app or product feature.

How? Through AI, they say.

However upon further probing, the user problem can be solved through conventional technology.

Or adding AI doesn't actually solve the user painpoint.

When tapping into AI innovation, the key question to ask is really — what AI capabilities can we leverage for disruptive innovation?

How does it provide a new user opportunity?

Some questions to ask to identify if AI adds unique value:

• Pattern: What is the user need and associated pattern of behaviour?
• Success: What is the definition of success to optimise for?
• Value: What is the additional value add of AI compared to a non-AI solution?
• Type: Who are the stakeholders involved?

Here's an example to illustrate.

First, you should already have your existing user personas, or user journey roadmaps to begin with.

• How might we: How might we improve flight booking conversion for users of our app?
• User journey: Flight booking experience for user on app
• User personas: Primary user persona (Parents looking for family vacations, Corporates travelling for business, individual travellers etc)

Then, extract insights from your user research. Craft and define the opportunity direction — include both user and business perspectives.

Here's an example.

Opportunity Direction: Improve flight booking experience (user value) + conversion of flight bookings (business value) of flight app

• User need: When to book for best price
• User pattern: Check multiple sites and find the best price to book
• Success: Buying at the right time and price
• Optimize: Best time to buy, avoid false price predictions

From there, identify touchpoints in your service journey that can potentially present themselves as AI opportunities.

And then define the AI-opportunity direction as follows:

By now, you will have an AI-enabled opportunity direction defined.

We have mapped user insights into a potential AI opportunity direction. But this is just the start.

The next step is assessing how feasible it is to turn user needs into achievable data inputs.

That's where step 2 of the HCAI framework comes in:

Step 2: Aligning user needs to data inputs

I'll cover how to achieve this step in Part 2 of my HCAI article series. To stay updated, hit a follow and turn on notifications if you haven't!

— -

In the meantime, check out my series on how to build a Human-Centered AI product:

👇🏻 Learn how to build a Human-Centered AI product

• Step 1: DEFINE your AI-enabled product opportunity
• Step 2: ALIGN your user needs to data
• Step 3: IDEATE design possibilities with AI capabilities

— -

If you made it this far, thank you for the read!

I hope you have found this article insightful, and always love connecting with fellow AI enthusiasts and AI design thinkers.

If you found this post insightful, hit the like button and follow me on Medium for more 🔔 — I appreciate it!

I'm building Human-Centered AI initiatives for 2024, and looking to connect with fellow AI enthusiasts and AI design thinkers for collaborations.

To learn more, drop me a follow and message on Linkedin!

Here are some additional HCAI resources that I have been inspired by in my research. Do check them out: - Google People + AI Research - IBM Human-Centered AI Research - AIxDesign Co