When assessing the value of new tech — what comes to mind? Assessing the value of Generative AI has gotten a lot of buzz for its novelty, unique applications, and potential impact on the business world. The “buzz” has contributed to the reported size of the generative AI market: $8 billion in 2021, with a CAGR of 34.6% through 2030. But the true impact of generative AI — is just that, buzz so far — the potential to create value, but not actual value.
What are companies really spending that $8 billion on?
Is the spend truly going toward AI? Or is it more data engineering plus a bit of machine learning? It’s hard to tell right now, as the hype and mystery of “generative AI” inflates valuations and feeds scores of headlines.
With generative AI engrossed in its own hype cycle, companies risk getting caught up in the thrill of a new invention, impulsively investing serious dollars and time. But like any shiny new invention, companies shouldn’t rush to adopt generative AI without considering how to extract real value. This is the critical difference between innovation and invention.
Buzz does not equal value
Generative AI is a form of artificial intelligence that creates net new content, including text, images, and speech. Think of ChatGPT, a model that interacts conversationally with users to generate new data from simple requests. This generative aspect signifies a transformative step: Previously, AI and machine learning (ML) could only analyze or act on existing data.
The generation of new content
The promise of generating net new content has companies salivating at the chance to apply the technology across their processes and systems. We’re already seeing generative AI used to:
- Develop original content (writing, pictures, video).
- Create large amounts of synthetic data or data about data, which can train other machine learning models or test new products and services.
- Plow through large data sets to highlight patterns.
- Personalize user experiences and content within a product experience or a digitally enabled service.
- Automate repetitive tasks, such as data entry or image annotation.
Although generative AI could significantly impact the business world, the technology’s specific benefits will vary depending on the business, industry, and application.
Innovation goes beyond mere invention
While generative AI has excited the collective imagination, the companies primed for success in the next wave of the digital economy won’t chase the shiniest new technology without putting customer or business value first. They understand that innovation is really about doing something in a new way that generates value — even if that something new is done using older tools.
For example, it might seem appealing to incorporate machine learning into a product recommendation engine to output recommendations to users. It’s the more novel invention, after all. But a decision tree can produce accurate product recommendations about as effectively while being faster to build and cheaper to maintain.
What is the lifecycle of generative AI?
Generative AI is still early in its lifecycle — the shiny new invention phase — and hasn’t had much time to generate significant real-world commercial success. People are reluctant to make consequential decisions based on data they cannot verify. This natural (and healthy) skepticism increases when a person or company doesn’t understand how that technology generates data.
Data collected and transformed
How data is collected and transformed to be used by AI impacts the quality and value AI can achieve for enterprises. This consideration requires a significant investment to factor into the ROI. Most businesses are already struggling with their current systems buried under mountains of valuable information that is hard to work with, so we cannot overlook this fact.
Successful organizations in 2023 will innovate while being mindful of this reality. They won’t build technology for technology’s sake — they’ll understand their hypotheses, make modest investments before making larger ones always with an eye toward the desired outcome.
Discovering Generative AI’s true value
Insights into what customers find valuable will outpace cool technology over time. When customers give up something they value — like money or time — they demand more value in return.
Successful companies will meet their customers’ needs with a three-pronged approach: assessing their target market or markets to determine the “why” behind the technology, testing their hypotheses in a lean approach (modest investment), and ultimately understanding where the compelling and durable value lies.
Determining the ‘why’ behind the technology
Start by considering your target market to determine the “why” behind the technology. Create a set of opportunity-hypotheses. Think of as many as you can, and don’t be afraid to ask a wide range of people for ideas — you will winnow down your list later. These opportunity-hypotheses should include who would benefit, how they benefit, and might include who would pay and why.
Evaluate and rank the list of opportunities against criteria like:
- How well positioned is your business to provide these value propositions?
- What are your brand and customer expectations?
- How big could the opportunity be?
Before you test these hypotheses, discuss the threshold to hit that’ll convince you to invest more in any single opportunity or combination. This is key to resisting the temptation of confirmation bias — seeing only the results that confirm the hypothesis you want to be true. Because this is an exploration, you may get a very unexpected result. An unexpected result can lead to an unexpected insight that can lead to even bigger opportunities.
Test your hypotheses in a low fidelity, lean way
How can we test our concepts without building them? What small investment in a test would convince us to want to do another round of investment?
The answers to these questions lie with your customers, not in your meeting room. You must get out of the building to test your hypotheses.
I highly recommend using user researchers during this stage. Their techniques for asking open questions without leading the interviewee to sense the answer you’re hoping for are invaluable to getting verifiable and repeatable results.
Paper prototyping and remote user testing
I’m also a huge fan of paper prototyping and remote user testing. The tooling to support these methods has come a long way. These options drastically reduce the cost of hypothesis testing, can be recorded or observed live, and allow you to quickly pivot the test script or the hypothesis.
When most leaders hear “user research,” they envision a long, expensive, and murky process. The best user researchers complete small batches of testing (5-8 users) and interpret the results with others before doing another round.
Done right, this type of testing is collaborative and participatory by stakeholders. The potential impact for later investor conversations is huge as executives can cite specific examples of potential clients talking about their context and what they value and would pay for.
Understand where the value lies
Once you complete your tests, it’s time to interpret your results. I often hear leaders say they want to be “data-driven,” — and I was one of those leaders. When I started observing user tests, I first noticed that the responses were qualitative and inconclusive, which felt insufficient. But then I realized that patterns would quickly emerge from those results.
I learned that interpretation is vital to the process and ripe for confirmation bias, unspoken assumptions, and opinions weighted by a person’s place in the hierarchy. I now seek to become “data-informed,” and the process itself is “a hunt for customer insights.”
So what makes for a truly valuable result in testing?
There are multiple possibilities. One clearly confirms that customers would value and pay your company for the solution you sketched or a close variant. These are rare, and you should be on the lookout for teams trying to tell you what they think you want to hear.
A likelier and better result is that your testing reveals that customers would generally find value in the solution, and you gain insight into why and what they value. This additional color is critical to all future decisions and gives your company a more significant competitive advantage, even if others are pursuing the same solution.
These additional opinion insights also offer options to pivot if you discover better or cheaper alternative approaches to deliver the same or richer value.
Completing these three phases is vital if companies want to build digital products with true business value and positively advance digital transformation.
Leverage real business value instead of hype
The age of needing to be the first-mover has been discredited. We’re still years away from the widespread adoption of generative AI — and we need that time to develop the talent capable of driving value-adding adoption. Costs will come down — the talent pool will deepen, and generative AI will move from hype to functionality.
In the meantime, we’ll see many companies claim to use AI — leveraging the hype — while actual adoption remains peripheral to the core of their products/services. The temptation to hop on the bandwagon will intensify as the market matures.
But those who determine how to leverage technologies like generative AI to create real value will set themselves apart and be best positioned as the next wave of the digital economy crests.
Featured Image Credit: Tara Winstead; Pexels; Thank you!
