Step 1: AI Opportunity Mapping

Where would it be most beneficial to integrate AI into our products?

AI Opportunity Mapping emphasizes the organization and its products.

This could serve as a potential starting point, especially if you're uncertain about which product or service should first adopt AI.

The process typically involves identifying areas where AI technologies could enhance or transform products, addressing problems, and creating value.

During this phase, our main goal is to uncover transformative capabilities from the perspective of specific departments and their products.

It begins with the query: “Which product or service should we apply AI to first to generate value?”

Let’s break this down step by step.

#### Step 1.1: Product and Service Analysis

The journey starts with a thorough analysis using either the organizational diagram or the product mapping method, aimed at identifying, detailing, and prioritizing each product’s pain points.

Following the identification of company challenges, the team pinpoints valuable aspects of the products for AI application, prioritizing these processes based on their impact and benefits.

Throughout this phase, all workshop participants contribute insights into the problem space, allowing us to understand their pain points and expectations, ensuring that the most critical areas are tackled first.

#### Step 1.2: Exploring AI Capabilities with AI Method Cards

Companies often become excited about new technologies, anticipating significant outcomes, only to find that these technologies don’t meet their needs.

A more effective approach is to first determine what you truly require and then choose the specific AI technology that aligns with those needs.

To avoid common pitfalls, we’ll introduce AI Method Cards during the Sprint process to map out AI possibilities and steer participants towards impactful solutions for their business and product requirements.

Think of these cards as a navigational tool that helps organize various AI options systematically. They provide a comprehensive overview, making it accessible for even non-technical users to explore how to apply AI within their business.

How does it work?

To give a snapshot of AI’s current state, we have categorized AI capabilities into 14 parent card categories, arranged by increasing complexity, from simple tasks like “AI finds and organizes information” to more intricate tasks such as “AI controls machines, robots, and vehicles.”

Each category is designed with the user in mind, ensuring clarity and accessibility for all, regardless of expertise. Each primary category features 4–8 specific AI technologies, labeled as child cards.

The back of each card presents three examples of how the particular AI technology can be utilized, demonstrating its wide-ranging applications and confirming its availability.

What AI technologies can you explore?

Let’s examine a few examples to clarify the concept of AI Cards.

• AI finds and organizes information
  • Product need: Streamlining legal research by efficiently locating relevant cases.
  • Technology to use: Utilizing AI and machine learning algorithms for advanced legal document analysis.
  • Example: Everlaw employs AI to quickly analyze extensive legal document repositories, enabling lawyers to swiftly identify pertinent cases and extract crucial insights.
• AI performs simple tasks
  • Product need: Automating mundane tasks to enhance efficiency.
  • Technology to use: Utilizing AI capabilities for automated task completion.
  • Example: Dooer streamlines accounting processes by automating routine tasks like data entry and reconciliation through AI integration.
• AI chats and talks
  • Product need: Improving communication and interaction through conversational AI interfaces.
  • Technology to use: Implementing Natural Language Processing (NLP) and machine learning algorithms for seamless conversation.
  • Example: ASKR.ai enables conversational interactions with data, leveraging advanced NLP technology to allow users to chat with and query their datasets effectively.

These examples illustrate how AI cards can cater to specific product development needs, creating an engaging journey.

Note: In AI Opportunity Mapping, AI Cards serve as fundamental elements for identifying suitable AI solutions, and they are also utilized during Concept Development sessions.

#### Step 1.3: Match AI Cards with Your Products

After gaining a solid grasp of AI capabilities and thoroughly analyzing the challenges faced by different departments and their products, we can identify specific areas in the product portfolio that would benefit from AI solutions.

The team should align each AI Card with the relevant capability in the product diagram.

These cards can be placed wherever deemed most applicable, and it’s essential to explain how these AI capabilities would be implemented in their respective contexts.

All subsequent steps focus on one goal: prioritization.

#### Step 1.4: Evaluate Products and Sketch

In this phase, we aim to identify which products would gain the most from AI, taking into account the quantity and potential impact of the AI Category Cards placed.

We evaluate all departments or respective products that appear most promising.

Additionally, we aim to provide a basic description and illustration of our intended use of AI for these products.

#### Step 1.5: Viability

This phase entails assessing the financial advantages of each AI opportunity, factoring in time savings, return on investment, and quality improvements.

We also evaluate the overall effect of each opportunity on the company's business, including its pain points, core activities, and future plans.

Note: This estimation exercise serves as a rough guide.

#### Step 1.6: Decision-Making

In this final stage, you prioritize the products or services that will focus on creating AI solutions.

You have identified AI opportunities within your business and categorized them based on their value levels.

Additionally, you’ve pinpointed the most valuable AI opportunities for specific products.

Step 2: Concept Development

What will the selected AI product or service look like?

The Concept Development session centers around enhancing or transforming products.

This marks the second potential starting point for your AI product project.

At this juncture, we have either identified which product or service to apply AI to, or you may be progressing from a previous phase (AI Opportunity Mapping).

This stage involves translating high-level objectives and opportunities from earlier stages into a more detailed product concept.

The main goal is to develop one or two AI solution concepts that deliver value to your target area.

Let’s break down the process into specific steps.

#### Step 2.1: Identify and Map Key Steps of the Persona’s Customer Journey

We start by creating a storyboard of the customer journey for a given persona and its value proposition.

A customer journey is a visual narrative that outlines every interaction a customer or persona has with a product.

Developing a customer journey requires a clear understanding of the story being told, which is achieved by breaking down the narrative into approximately 5–10 steps.

At each step, we sketch the interaction, describe the actions taken by the persona, and note any involvement of intelligent software, clarifying the intent behind each step.

The aim is to create a comprehensive customer journey that includes all steps that could be partially or fully automated using AI, thus enhancing the persona’s experience.

This customer journey will serve as the foundation for developing AI solution concepts.

#### Step 2.2: Customer Journey and Product Analysis

In this step, we identify and prioritize areas that could benefit from AI application.

Participants outline and note the pain points in the customer journey and the product itself.

This process highlights the strengths and positive aspects of operations, providing a holistic view of crucial steps and identifying valuable areas to prioritize for product enhancement or transformation.

#### Step 2.3: Exploring AI Capabilities with AI Method Cards

If the AI Method Cards were already introduced during the AI Opportunity Mapping phase, there’s no need to reintroduce them.

If this topic has been covered, we can proceed to the next step; otherwise, please refer back to the AI Opportunity Mapping section for an explanation of the AI Method Cards.

#### Step 2.4: Match AI Cards with the Customer Journey

After thoroughly understanding AI's capabilities and analyzing your challenges, it’s essential to identify where AI solutions are most appropriate.

The objective is to enhance the product using AI.

The team strategically places AI Cards in the process steps based on their respective capabilities, ensuring to explain how each AI technology will improve the product or service.

#### Step 2.5: Prioritize and Decide

In this phase, your team will prioritize which aspects of the product should be enhanced with AI solutions.

This involves identifying the top three AI technologies that hold the most promise for transforming your product and selecting two critical customer journey steps to focus on.

We will also consider both pain points and value creation opportunities to determine which improvements will have the most significant impact.

#### Step 2.6: AI Ethics Check

AI raises ethical challenges in areas like human rights, discrimination, surveillance, transparency, privacy, security, freedom of expression, employment rights, and access to public services.

Thus, understanding the ethical implications of any AI solution we develop is essential.

To facilitate participants in considering these ethical aspects, we utilize AI Ethics Cards. These cards guide discussions surrounding various AI challenges in an organized way. They serve as a deck of cards, each addressing a different ethical topic, providing a broad overview.

How does it work?

AI Ethics Cards are based on research into ethical public discourse and the identification of relevant topics.

To provide an overview of current ethical concerns related to AI, we’ve categorized these topics into 18 categories.

Each card features a main topic and a corresponding description.

These examples illustrate the diverse range of ethical concerns associated with AI, making the topics comprehensible, even for those with limited prior knowledge.

During this workshop, we’ll discuss the impact of ethical considerations on our AI solution.

In your specific concept, we’ll organize the most pertinent cards in their optimal positions and outline the best and worst-case scenarios for this AI Ethics Card.

Our goal is to identify any deficiencies or constraints in your solution to achieve the best possible outcome.

Note: Here, the decision undergoes an ethical review and is made by the stakeholders.

#### Step 2.7: Exploring Data Source Cards for Assessment

In this step, we concentrate on evaluating the redesigned product or service.

We assess the AI solution from three Design Thinking perspectives:

• Technical feasibility
• User value
• Organizational value

We will start with what are referred to as Data Sources Cards.

How does it work?

Data plays a crucial role in AI.

During the workshop, we use Data Sources Cards to help participants explore data sources, even without technical expertise.

These cards showcase various data sources that can serve as a foundation for the team’s AI product project, offering over 60 potential data sources for AI, with the option to include additional sources.

In this process, we meticulously identify the data necessary for our product processes.

In the subsequent step, we will specify and prioritize this data in greater detail.

At this point, we also conduct an initial assessment of the technical feasibility and potential cost savings or value that the solution could deliver to our company.

#### Step 2.8: Exploring Resources and Role Cards for Assessment

To effectively evaluate an AI solution, it’s vital to develop a business case, which includes analyzing potential benefits and cost assessments.

By examining these elements, you’ll gain a clearer understanding of the value of the AI solution and make informed decisions regarding its implementation. Initially, identifying the necessary resources and roles for integrating the AI solution into your product or service is crucial.

How does it work?

During the workshop, we utilize Resources and Roles Cards to help participants identify key resources and roles.

These cards encompass a variety of resources and roles that serve as a starting point for the team to consider for their AI product project.

We provide an overview of potential data sources for AI, offering more than 30 options and the possibility to include additional ones.

Throughout this process, we aim to identify the necessary resources and roles for our product.

Finally, we evaluate the financial aspects, focusing on the areas where your solution delivers benefits compared to its costs.

When assessing the benefits, consider the following:

• How much time is saved?
• What is the return on investment?
• What is the quality increase?

Note: The estimates provided here are approximate.

By the end of the collaborative workshop, you will have developed one or two AI product solution concepts that effectively target your area of interest.

You’ll have identified crucial AI tasks, integrated them into a value-driven workflow, and defined the end product process.

Moreover, you’ll have gained a comprehensive understanding of the data requirements and conducted thorough ethics checks for your use cases.

Curious about what comes next after the workshop?

We recommend structured interviews with subject matter experts (SMEs) to evaluate the feasibility of the ideas. These interviews can be conducted by you or our experts.

Additionally, we can create an implementation roadmap and develop a prototype to validate the concept.

This prototype can be assessed with real users, either as a web application or a similar solution.

Let’s navigate you through this process.

Step 3: Tech Check

How feasible is the concept and implementation of the AI solution?

The technology check primarily focuses on the developed AI solutions.

After concept development, we proceed to a technical check, which emphasizes the practical aspects of bringing the proposed AI solutions to fruition.

Beyond initial stages, the technical check requires the expertise of AI professionals.

What role does an AI expert play?

An AI expert, often a full-stack data science engineer specializing in AI, plays a varied and crucial role, bridging creativity and practicality.

To ensure successful and impactful AI solutions, they perform the following tasks:

• Organize SME interviews: The AI expert schedules and leads discussions with Subject Matter Experts (SMEs) from the team. Through these interviews, they identify the environment, infrastructure, and key data sources, forming a foundation for informed decision-making.
• Technical evaluation: The AI expert evaluates the technical feasibility of the proposed AI solutions, ensuring they align with practical implementation. They review existing AI methods and algorithms to guarantee the successful execution of the AI system.
• Evaluation outcome and roadmap: Following validation of the AI solution, the next step involves forming a clear roadmap to be shared with the team for prototype design.

An AI expert must comprehend both the broader context and the technical intricacies.

To thrive in this field, one requires a diverse skill set, encompassing full-stack data science, technical infrastructure, and the latest AI methods.

This broad expertise guarantees in-depth knowledge across various domains, enabling the expert to manage both conceptualization and technical implementation challenges.

Step 4: Rapid Prototyping

How can we evaluate and refine the AI solution?

Technical prototyping emphasizes hands-on testing and improvement.

We complete this process within seven days.

Phase 1: Design (2 days)

In the initial phase, we create an AI Minimum Viable Product (MVP) concept by analyzing the primary problem in the given use case. This involves dissecting the problem, aligning it with AI technology, and evaluating user and business values. The outcome is a comprehensive map and definition of the problem areas, along with customer requirements. On Day 2, we refine the specifications and create a detailed implementation plan.

Phase 2: Build (5 days)

After the design phase, we construct the prototype through a structured approach that includes exploratory data analysis, feature selection, model validation, and deployment. The results are organized and visualized for clarity and transparency. The end product could be a functional web app featuring dynamic capabilities and understandable metrics.

Phase 3: User Testing (1 day)

If necessary, we conduct user tests to gauge responses to the new solution or process. At the conclusion of the testing phase, we analyze feedback and metrics to identify areas for improvement. We wrap up this productive journey with a tested solution and a satisfied team.

What’s next?

With the insights gained, you can proceed to develop and implement the actual process into your product chain.