top of page
  • Writer's pictureAiAnalyst

Only 5% of AI Products Will Succeed

Updated: Sep 25, 2023

Artificial Intelligence (AI) has become a buzzword in the business world, with companies eager to tap into its potential for transforming their operations and enhancing customer experiences. However, despite the hype, the reality is that the majority of AI products fail. According to a Harvard Business School professor, a staggering 95% of all products miss the mark. But this doesn't mean that AI is destined for the same fate. By adopting a customer-centric approach and following some key principles, companies can increase their chances of being part of the successful 5%.

Putting the Customer Problem at the Center

One of the common pitfalls in product building is the tendency to prioritize technology over the customer's problem. Many people believe that a great idea or a technological breakthrough is the key to success. However, this technology-first thinking can lead to failure. Instead, companies should always start by understanding and addressing the customer's pain points. Before integrating AI into a product, it is crucial to identify the specific customer challenges or questions that can be solved using AI capabilities.

For example, a team we worked with developed an AI assistance chatbot that answers users' questions within a customer's app based on their knowledge base and web content. By leveraging AI, the team was able to provide instant answers, eliminating the pain of prolonged content searches. This customer-centric approach not only solved a common problem but also demonstrated the value of AI in enhancing the overall user experience.

Driving Differentiated Value Through Iteration

Once the customer problem is identified, it is essential to determine what makes the product indispensable. Data plays a critical role in this process. Companies need to continuously collect and analyze user behavior data to fine-tune their AI models and create a more personalized customer experience. This iterative approach allows for constant improvement and ensures that the AI models remain accurate and effective.

By leveraging behavioral data insights, companies can inform their AI models, leading to more accurate predictions and improved personalization. This data-driven feedback loop creates a competitive advantage, as companies can continuously refine their models based on real-time user interactions. The more data is collected and analyzed, the better the models become, resulting in increased product usage and customer satisfaction.

Balancing Time-to-Market with Customer Trust

In today's fast-paced business landscape, time-to-market is crucial. Companies that take too long to launch their products risk losing out to competitors. However, it is equally important to balance speed with customer trust. Deploying AI solutions that are not yet fully trusted or reliable can have serious consequences and undermine the success of the product.

Transparency is key to building customer trust. By providing visibility into how AI models generate outputs, companies can address concerns and alleviate doubts. For example, our customers use Amplitude to personalize messaging with a feature called predictive cohorts. Initially, some customers were hesitant to take recommended actions because they didn't understand why certain users were identified by the AI model. To build trust, we showed customers the underlying data that led to the model's conclusions, increasing their confidence in the predictive cohorts feature.

By finding the right balance between time-to-market and customer trust, companies can ensure that their AI products deliver value while maintaining customer confidence.

The Data Challenge

Implementing AI successfully requires good quality and relevant data. Companies need access to properly labeled data sets without biases that could skew the results. However, collecting and preparing such data can be time-consuming and expensive.

For example, if a company wants to use AI to detect cats in a hen house, they would need a large dataset of labeled images with cats and without cats. However, if the company later realizes they need to know the exact number of cats in each image, the dataset would need to be re-labeled, adding complexity and cost to the process.

The challenge of data collection and labeling becomes even more pronounced when organizations try to achieve highly specific segmentation. If the data collected does not align with the desired segmentation, the insights derived from the data will be irrelevant and ineffective.

To overcome these challenges, companies need to be clear about the decisions they want to make and ensure that the data collected is representative of those decisions. Additionally, when predicting rare events or automating business processes, it may be necessary to build systems that capture the necessary data before developing AI models.

Learning from AI Failures

AI is not immune to failures, and companies must learn from these experiences to improve their future AI initiatives. One notable example is the case of Zillow, a real estate company that wrote down $304 million worth of homes purchased based on the recommendations of its AI-powered Zillow Offers service. The company faced unforeseen events such as the global pandemic and supply-demand imbalances, which were not accounted for in its AI models. This failure highlights the importance of understanding the limitations of AI and not relying solely on historical data.

AI algorithms learn from the past, so if something unprecedented occurs, they may not be able to predict or handle it effectively. Additionally, AI lacks common sense, and certain factors or variables that were not considered in the training data may significantly impact the accuracy of AI models.

To mitigate the risks associated with AI failures, companies should prioritize factors such as cybersecurity, compliance, explainability, and personal privacy. By addressing these concerns and being aware of the limitations of AI, businesses can make more informed decisions and minimize the potential negative impacts.

The Future of AI and Its Impact on Jobs

There is a common fear that AI will replace human jobs and lead to widespread unemployment. While AI will certainly automate certain tasks, it is expected to create more jobs than it eliminates. According to Semrush, AI technology is projected to create 12 million more jobs than it replaces.

However, it is crucial to prepare for the changing job landscape. Companies should invest in upskilling and reskilling their workforce to adapt to the evolving AI-driven environment. Additionally, certain industries, such as technology, financial services, healthcare, education, and government, are expected to experience higher demand for AI specialists.

While AI may automate certain job functions, there will always be a need for human expertise in areas that require creativity, critical thinking, and emotional intelligence. By embracing AI and leveraging its capabilities, companies can enhance productivity, reduce errors, and create new opportunities for their employees.

AI in Retail: Personalization and Hyperpersonalization

The retail industry has embraced AI to provide personalized experiences and make data-driven decisions. According to Juniper Research, machine learning spending in the retail industry is expected to grow by 230% between 2019 and 2023. Retailers are leveraging AI to hyperpersonalize their offerings and enhance customer engagement.

By analyzing customer data, retailers can understand individual preferences and tailor product recommendations accordingly. This level of personalization improves customer satisfaction and increases the likelihood of conversions. Additionally, AI-powered chatbots are being used in customer service to provide quick and accurate responses, reducing costs and enhancing the overall shopping experience.

The adoption of AI in retail is driven by the desire to stay competitive and meet customer expectations. Retailers that leverage AI effectively can gain a significant advantage in a highly competitive market.

Chatbots: Enhancing Customer Service with AI

Chatbots have become increasingly popular in customer service, offering quick and efficient support to customers. According to Salesforce, 23% of customer service companies are currently using AI, and IBM reports that chatbots can reduce customer service costs by up to 30%.

Chatbots provide immediate responses to customer inquiries, ensuring round-the-clock availability and improving response times. They can handle routine queries, freeing up customer service representatives to focus on more complex issues. However, it is important to strike the right balance between chatbot automation and human interaction, as some customers still prefer speaking to a live representative.

Despite the benefits of chatbots, there are challenges to overcome. Customers may have concerns about the accuracy and understanding of chatbot responses. To address these concerns, companies should invest in training chatbots to better understand and respond to customer queries. By continuously improving chatbot capabilities, businesses can enhance customer service and streamline their operations.

The Growing AI Market and Its Potential

The AI market is experiencing rapid growth, with a projected market value of $60 billion by 2025. Companies across industries are investing in AI to drive innovation and gain a competitive edge.

The United States leads in AI adoption, followed by countries like India, Canada, Germany, and the United Kingdom. Private sector companies are planning significant investments in AI systems, with 44% of them planning to invest in AI in 2023, according to Info-Tech Research Group.

The retail industry is expected to invest heavily in AI, with a projected market size of $20.05 billion by 2026. Machine learning spending in various sectors, including technology, financial services, healthcare, education, and government, is also on the rise.

The future of AI looks promising, with advancements in natural language processing, machine learning, and deep learning. Companies like IBM and Google are at the forefront of AI development, pushing the boundaries of what is possible. As AI continues to evolve, it will play an increasingly vital role in shaping the business landscape.

Overcoming Obstacles and Reaping the Benefits of AI Adoption

While AI adoption presents challenges, such as the lack of skilled individuals, unclear ROI metrics, and fear of job replacement, the benefits outweigh the obstacles. AI can reduce operational time, provide valuable business insights, automate repetitive tasks, enhance productivity, and deliver better customer service.

To maximize the benefits of AI adoption, companies should invest in training their employees, establish clear metrics to measure ROI, address governance concerns, and ensure transparency in AI systems. By taking a strategic and thoughtful approach to AI implementation, businesses can harness its full potential and drive growth.


AI holds tremendous potential for businesses, but success lies in building customer-centric products, iterating based on data-driven insights, balancing time-to-market with customer trust, and addressing the challenges associated with data quality and AI limitations. With careful planning and execution, companies can navigate the path to AI success and become part of the 5% that achieves remarkable outcomes. The future of AI is promising, and those who embrace it will be at the forefront of innovation and transformation in their industries.


Obtuvo 0 de 5 estrellas.
Aún no hay calificaciones

Agrega una calificación


bottom of page