From Data to Design: How AI Transforms Business Strategy
Enhancing Products and Identifying Market Gaps with AI Data Analysis Tools
How would you take a million dollar business idea from inception to reality? What steps do you take to ensure your product meets — and exceeds — your users’ needs?
Would you try to find an untapped niche, or try to outperform the competition by always staying one step ahead?
These are questions that every founder has contended with at some point, and the path forward is rarely clear. To reduce this uncertainly, businesses of all sizes often rely on data-driven market analysis and user research. In fact, data-driven companies are 23 times more likely to acquire customers and 19 times more likely to be profitable compared to companies that don’t make use of customer analytics.
Traditional methods of large-scale data analysis have struggled for years to convert unstructured text data into meaningful market insights. For problems that require you to consider complex relationships and strategic competitor moves that aren’t easily converted to numerical data, we have always had to rely on human experience and intuition to stay competitive. While a small startup may choose to collect and analyze this data themselves, hiring someone with the skills and business savvy to extract meaningful insights from data and use it to formulate a corporate strategy often requires some pretty deep pockets. Modern AI techniques, particularly Large Language Models (LLMs) and Natural Language Processing (NLP), are now making it possible to analyze complex unstructured text data effectively with minimal resources.
With over a decade of experience in product design, most recently serving as the Head of AI Architecture at NAX Group, Elena Vid has spent her career working to solve such problems. As more and more companies integrate Generative AI into their workflows and increasingly choose to automate aspects of product design and development, it’s worth taking a step back to see where you can — and where you shouldn’t — fit AI into your business strategy.
You Don’t Have to Ship an AI Product
When it comes to UX design, there’s a lot that already works.
Rather than overhauling existing systems or introducing complex new interfaces, research consistently shows that simplicity and familiarity in design leads to higher user adoption and satisfaction. As Elena notes, “Embedding AI in everyday actions creates significant adoption, which is very important for AI tools.”
In this way, AI can act as an “invisible aid” in the product design process, benefiting the user without the friction that often accompanies technological change.
Elena leverages AI to analyze vast amounts of data to inform her product strategy and design decisions. Take competitive gap analysis, for example. If you want to know where you stand amongst your competitors, there are a lot of relevant factors to keep track of. You can employ social listening to see what people are saying about your company or product on social media, look at market reports to identify trends and areas for growth, and watch your competitors for unusual movements such as hiring waves and mergers/acquisitions. This data can be difficult to acquire and even harder to make sense of, but if done correctly, it is less prone to bias than traditional methods of user research such as focus groups. For a fraction of the cost of a human data analyst, LLMs can make sense of these disparate sources of data and identify emerging trends and market opportunities to inform your Go-To-Market (GTM) strategy.
AI Consulting Tools for Long-Term Business Growth
While developing a business consulting tool for startups and ventures forming a long-term growth strategy, Elena has utilized similar techniques in her own work. Through social listening, market reports, and competitor data, she is testing the efficacy and logic flow for creating “digital twins” of markets, enabling companies to simulate different market scenarios and manage risks more effectively. By automating data collection and analysis, this tool would allow for real-time adaptation and rapid iteration based on the most current data, ensuring that strategies are always aligned with the latest market dynamics.
Since trends change so quickly, this evaluation of your position in the market is best done quickly and often. AI allows you to do these analyses much quicker when integrated seamlessly into the design process. AI does not necessarily have to be included in the product — it is just as helpful, if not more so, to integrate AI into your tech stack to enhance the product.
For businesses trying to scale AI solutions, focus on creating and shipping a minimum viable product (MVP) before going back to implement some of the more complex functions and best practices. Don’t get too invested in the perfect toolchain; instead make sure your product is meeting the needs of your target customers and then branch out. Building complex data infrastructure, like vector stores or advanced data labeling systems, can be costly and slow down the initial go-to-market process, and are more valuable once you have made some revenue to reinvest into your product.
As Giacomo Marzi, researcher at The University of Trieste explains, “Developing new products and services beyond what is required by the needs of users, market demand, and the resources of companies ranks among the top 10 risks leading to new product development (NPD) failures.”
Over-engineering a product often leads to increased complexity without providing additional value to users. By focusing on shipping an MVP first, businesses can avoid the pitfalls of unnecessary complexity and better align their offerings with customer needs.
What is Your Use Case?
Whenever a new technology becomes available that promises to revolutionize the way we approach a problem, it’s easy to jump on the hype train. Cloud computing promised to reform infrastructure costs and scalability, while cryptocurrency promised to transform financial services and usher in the era of Web3.
Both have brought with them unforeseen costs and drawbacks. Unpredictable pricing and vendor lock-in have caused many companies to reduce their reliance on cloud providers and instead shift to multi-cloud environments, on-prem storage, and edge computing, whereas market volatility and regulatory uncertainty have resulted in hesitancy to shift the financial system towards crypto.
Even if these technologies have not followed through with their grandest visions of the future, they have still radically changed the IT and software development landscape. Taking a look at historical examples makes it clear that we should approach AI with the same appreciation and caution that we give its predecessors.
For an example of a relevant use case, Elena explains in detail a product she worked on for art collectors to navigate the complex world of acquiring new works for their collection.
AI Tools for Collectors in the Fine Arts Market
In the world of fine art, making informed decisions about new acquisitions is a delicate balance of knowledge, taste, and strategy. Unlike auction houses, where sales data is often publicly available, private galleries operate in a more opaque manner. Prices are rarely fixed or openly displayed, and instead of offering a straightforward sale, galleries often choose buyers based on relationships that will benefit the artist or gallery in the long term. This intricate dynamic leaves many collectors navigating a complex and nuanced art landscape, one where traditional market analysis tools fall short.
For prestigious collectors, there are a staggering amount of factors to consider when deciding if a piece is worth adding to their collection. A piece’s provenance — its ownership history and exhibition record — plays a major role in its value. If a work has previously been sold at auction or in a gallery, its historical price sets a benchmark for its monetary worth, influencing a piece’s prestige and rarity.
Once you factor in the condition of the piece, authenticity certificates, critical acclaim, the artist’s career trajectory, and demand from other distinguished collectors, suddenly there are many more subjective factors to consider when speculating about trends in the secondhand art market. A great deal of this information is gatekept by major players in the art world as well, making it even more difficult to judge a piece’s relevance to your collection.
When you have this much data, so much of which is qualitative and not numerical in nature, it’s difficult to even know where to begin analyzing this. Elena approached this by creating a knowledge graph, essentially an interconnected network of data points, and using LLMs and NLP models to extract insights and provide collectors with an extensive overview for how a specific artwork can benefit their and the gallery’s reputation.
By using AI to gain insights from this data, Elena explains, “It helps collectors understand how they can curate meaningful collections that reflect their personal values, while also gaining a deeper understanding of the evolving art landscape.”
The complexity and variety of relevant unstructured text data makes this tool a prime candidate for integrating AI into a product, as traditional statistical methods like linear or multiple regression analysis would have been insufficient to accurately analyze the problem space, and likely too costly and time-consuming to develop relevant metrics that could be used in such a calculation.
Once you know what problem your product is trying to solve, ask yourself: “Is AI the best solution to this problem? Is there a simpler way to approach this?” Oftentimes the simpler solution will be more consistent and dependable, resulting in a better product for your end users.
However, if your problem involves large-scale data analysis with unstructured text, such as text classification, summarization, or semantic understanding, LLMs or smaller NLP models tend to be among the most effective solutions, particularly for tasks that require understanding complex context or relationships in language. While LLMs offer powerful capabilities, simpler NLP models or traditional methods may be more efficient and cost-effective for smaller-scale or less complex tasks.
Overcoming Challenges in AI Adoption
One of the major challenges people often face when it comes to integrating AI into product design is getting other people on board with adopting unfamiliar technology. Some organizations are still hesitant to incorporate AI for a number of reasons. There are valid concerns about data privacy and security risks since large amounts of sensitive data is often required for analysis, and a lack of in-house expertise makes it difficult for organizations to understand how to implement AI effectively, leading to fears of mismanagement or poor execution. Employees and clients may also have reservations regarding job displacement or ROI of unfamiliar AI tools. Many people are also process-oriented when it comes to their workflow rather than outcome-oriented, and disrupting that process can be quite uncomfortable.
When automation is the right business decision, it is important to highlight where the benefits outweigh the costs. When pitching AI integration to a product lead or to a team of executives, make your presentation interesting, and clearly outline the use cases. If you can’t get someone excited about this technology that you believe in, you’ll have an uphill battle trying to convince them that it’s worth considering.
“I wouldn’t pursue them, or try to convince them,” Elena explains. “I would make it engaging for them by showing them how interesting and eye-opening AI technology can be.”
You should always listen to and address their concerns about using AI as a part of their toolchain. Successful leaders actively acknowledge their own biases and seek out diverse perspectives both internal and external to their organization to help mitigate them. If you’re facing cultural pushback against the use of AI tools, don’t try to force their adoption. It’s better to maintain a positive and trusting relationship with your clients and coworkers than it is to impose your preferred solution regardless of its reception. Team members should always feel comfortable providing honest, constructive feedback.
Future Trends and Predictions in AI Product Design
Elena imagines a future where AI becomes a pivotal tool in creating highly personalized and customized experiences for users, moving beyond the traditional one-size-fits-all approach. She believes that the key to successful AI product design lies in its ability to listen to consumers and adapt in real-time. According to Elena, AI systems will increasingly cater to individual needs and preferences, allowing businesses to adjust their offerings dynamically. Whether some users prioritize speed, comfort, or aesthetics, this shift toward customization addresses not only the challenges of changing consumer preferences but also the biases inherent in data interpretation, ensuring that AI-driven products are more finely tuned to each person’s unique requirements.
One of Elena’s most fascinating predictions is the rise of the "one-person million-dollar business," enabled by AI. She envisions a world where AI democratizes business creation, empowering small teams or even individual entrepreneurs to launch and scale ventures with minimal resources. Right now, MVP development costs can range from $10,000 to $150,000, depending on the complexity of the app or service, the location of the development team, and the chosen features. (Daria Taranets - Lunka) However, as AI continues to advance, it will reduce these barriers, allowing creative individuals to solve problems more efficiently. With AI’s ability to analyze real-time market information, social listening data, and simulate market conditions, entrepreneurs will be able to craft the perfect product tailored to their market, eliminating much of the trial and error that currently plagues business creation.
While Elena acknowledges that AI adoption has been largely concentrated in B2B applications, she argues that the future of AI must expand into B2C products to gain broader societal acceptance. For this to happen, AI needs to enhance everyday life and build trust with consumers. Elena calls for more dialogue with consumers to understand their fears, but just as importantly, to learn what advantages they hope to gain from AI. By developing solutions that improve day-to-day experiences and address real consumer needs, we can bridge the gap between technological innovation and human trust, ensuring that AI improves the quality of life for consumers in ways that feel intuitive, safe, and beneficial.
You can take a look at more of Elena’s work here:
Linkedin: https://www.linkedin.com/in/elena-vid/
Website: https://www.elenavid.com/