The AI Product Charter

How to validate AI-powered Products

TLDR

Before diving headfirst into developing a new product or concept, it's crucial to first assess if it's a venture worth embarking on. The initial step involves researching whether your idea is already out in the market. Discovering a similar existing product isn't necessarily a deterrent, but it does call for a careful analysis of the market you're about to enter.

In this post, we will explore:

1. How do you validate an AI-powered product idea?

2. AI Product Charter: What is it? Why is it needed?

3. The AI Product Charter framework

---

The statistics confirm that:

• 9 out of 10 startups fail (Startup Genome, 2023)

• Data from Pitchbook shows that 3,200 venture-backed firms in the U.S. have gone out of business in 2023. Those startups had raised $27.2 billion collectively.

• 42% of startups fail due to lack of a market need (lack of product market fit) (Forbes, 2023) 

• 2 out of 10 new businesses fail in the first year of operations (Bureau of Labor, 2023)

• 38% of Venture Capital firms did not actively invest in 2023 (Source: Pitchbook)

And you may wonder:

• Why do 90% of startups fail? Why have 50% of companies in the Fortune 500 vanished since 2000?

• How can I create a defensible business moat while embracing new AI technology?

• What should I do to not stay behind?

How do you validate an AI-powered product idea?

First of all, let’s define AI-powered products.

AI-powered products are software or hardware products that rely on foundational AI technologies such as machine learning, deep learning or generative AI to drive their core functionality, features and innovation.

Examples of AI-powered products:

1. Smart Assistants: Devices like Amazon Alexa, Google Home, and Apple's Siri use AI for voice recognition and Natural Language Processing (NLP) to perform tasks based on user commands.

2. Autonomous Vehicles: Tesla cars, as an example, are equipped with sensors, cameras, and advanced navigation algorithms to improve the driver experience.

3. Personalized Recommendations: E-commerce and streaming services like Amazon, Netflix, YouTube, Facebook and Spotify use recommendation algorithms to analyze user preferences and behavior to recommend products or content.

4. Image Generation: AI software such as Adobe Firefly use generative AI algorithms to transform creative concepts into vivid, high-quality visuals, enhancing the creative process and output.

The initial step before starting to develop a product involves researching whether your idea is already a commercial product in the market.

In this context, several critical questions come into play:

1. Can you improve significantly? - Determine if you can offer a markedly better product. For instance, consider the iPod, which outpaced other mp3 players largely due to its smooth integration with iTunes. Or Google Maps, which gained an edge over MapQuest with its superior accuracy, user-friendliness, and feature set. Similarly, Figma's web-first approach gave it a competitive advantage over Adobe in design collaboration. If your idea does not promise a significant and quantifiable improvement over existing products, it might be wise to rethink your strategy.

2. Is the idea fundamentally sound? - Evaluate the core value of your idea. Is it intrinsically a scalable, technically viable and sound concept?

3. Do you have the right team to execute it? - Consider your team’s unique qualifications and whether they align with your idea’s strengths.

If the market already has a similar product and you can not significantly surpass its capabilities, it is probably more strategic to explore other ideas. Marginal improvements on existing products often lead to challenging market battles, and your efforts might be better utilized elsewhere.

On the other hand, just because your idea already exists in some form does not automatically invalidate its potential for success. It is essential not to jump to conclusions about its necessity or success probability. Instead, you need to apply a product validation framework to critically assess the viability and potential of your concept. This is where the AI Product Charter comes into play.

AI Product Charter: What is it? Why is it needed?

Until now, the success and viability of a new concept or product idea was validated with a standard product validation framework such as Jobs To Be Done. This method focuses on understanding customer needs, the context in which a product is used, and the distinct function that customers expect the product to fulfill in their daily lives. It is a customer-centric approach that emphasizes a product’s functionality and user satisfaction.

However, with the advent of advanced technologies like Generative AI, Artificial General Intelligence (AGI), and Artificial Super Intelligence (ASI), the landscape has changed significantly. These technologies are not just another incremental step in the evolution of AI, but represent a paradigm shift in capability, complexity, and potential impact. Thus, a more comprehensive and robust validation framework is needed to address these changes:

1. Understanding Technical Complexities: These technologies are inherently complex and evolving rapidly. A validation framework needs to comprehend the technical nuances, the potential for continuous learning and evolution of these systems, and their interaction with diverse and dynamic environments.

2. Ethical and Societal Implications: AGI and ASI, especially, raise significant ethical and societal questions. A framework should consider the broader implications of deploying such technologies, including privacy concerns, ethical use, potential biases, and the impact on employment and societal implications.

3. Risk Assessment: The risks associated with Generative AI, AGI, and ASI are not only technical but also existential. The framework should include a thorough risk assessment strategy, evaluating both short-term and long-term risks, and plan for unforeseen consequences.

4. Regulatory Compliance and Standards: As these technologies are relatively new, regulatory frameworks are still catching up. The validation process should anticipate and adapt to potential regulatory changes and ensure compliance with existing and future standards.

5. Scalability and Integration: Given the expansive nature of these AI technologies, the framework should assess scalability and the ability to integrate with existing systems and infrastructures. It should consider the technological ecosystem in which the product will operate.

6. Market and User Adoption: Understanding the market readiness and user acceptance for products based on these advanced AI technologies is crucial. The framework should evaluate public perception, potential market size, and adoption challenges.

7. Sustainability and Longevity: With the rapid pace of technological advancement, it's important to consider the sustainability and future relevance of the product or concept. The framework should evaluate the potential for future development, adaptation, and long-term viability.

The AI Product Charter framework

The AI Product Charter framework, structured like the layers of an onion, offers a comprehensive approach for validating AI products. Each layer builds upon the previous, ensuring a holistic and strategic product validation process. Below we expand on each layer with specific examples:

1. AI Product Leadership: This layer sets the overarching strategy for the AI product. It should define the governing principles on how the product will penetrate the market, how it will adopt users and how it will grow long-term. For example, a company developing an AI-driven health monitoring app should determine its market entry strategy, user adoption plans, and long-term growth. Leadership should decide whether to target a niche market initially, like athletes or seniors, and how to expand over time.

   

   Butterfly scanlab (Source)

   AI-powered product example: Butterfly Network entered the medical imaging market with its AI-powered handheld ultrasound, initially targeting underserved healthcare professionals with an accessible and cost-effective solution. They facilitated user adoption through intuitive design and comprehensive training, leveraging AI for enhanced diagnostics. Over time, the company expanded its offerings and market reach, innovating beyond initial applications to include specialties and veterinary medicine, exemplifying strategic growth in AI product leadership.

2. Challenge: The AI product should resolve a genuine need or problem at its core. A critical question to ask is: if this AI-powered product vanished, will the customers be impacted in any way? If so, how much? How are the customers solving this challenge now? For instance, if an AI-powered language translation tool ceased to exist, would businesses and travelers who rely on it for real-time, accurate translations be significantly impacted? Currently, these users might be using less sophisticated translation apps or human translators, which might not offer the same speed and convenience.

AI-powered product example: Grammarly, an AI-powered writing assistant, significantly enhances writing quality by offering real-time, context-specific corrections beyond basic spellcheck capabilities. Its absence would force users to revert to less sophisticated tools or manual editing, impacting efficiency and accuracy. This dependency underscores the AI tool's value in addressing the genuine need for advanced, convenient writing assistance.

3. AI Golden Metrics: This is a fundamental layer, where product leaders need to define what success looks like for this product and how to measure it. Examples of evaluation metrics (e.g., quantitative, health, user engagement) can be found in a previous post here. These metrics need to align with the primary objectives set in #4. There is no need to add complexity at this stage by defining OKRs, guardrail and north star metrics. These will follow after the AI product validation phase has been accomplished. For example, for an AI-driven content recommendation system, success metrics might include user engagement rate, the accuracy of recommendations, and user retention rate.

4. Core Objectives: After setting the AI Golden Metrics, it is crucial to define the main product objectives, which need to be aligned with all the previous layers #1 - #3. For example, if the objective of an AI-driven financial advisory tool is to democratize investment advice, the objectives might include reaching a broad user base, ensuring high accuracy in investment suggestions, and maintaining user trust and satisfaction.

5. AI Hypothesis: A testable hypothesis needs to be formulated. For example, for an AI-based predictive maintenance tool for manufacturing equipment, the hypothesis might be: "Implementing AI predictive analysis will reduce equipment downtime by 30% and save costs by 20%."

6. AI Technical Matrix: Assessing the technical feasibility of an AI concept or product involves evaluating the AI model's capabilities, data requirements, and integration with existing systems. For an AI-powered chatbot for customer service, this might involve analyzing natural language processing capabilities, data privacy concerns, and integration with the company’s existing CRM software at a high level.

7. AI risks: A thorough assessment of risks is crucial, especially during market adoption and scalability phases. For an autonomous driving AI system, risks could include technical reliability, regulatory compliance, ethical considerations around decision-making in unforeseen circumstances, and public perception and acceptance.

AI Product Charter (Source: Hatch Labs)

Each of these layers is critical for the successful development, launch, and scaling of AI products, ensuring they are not only technologically advanced but also ethically responsible, market-ready, and user-centric.

---

Hatch Labs is deeply committed to launching AI-powered products and services which are aligned with the above principles. Feel free to directly contact Eva Agapaki for more information or inquiries on the above!