Turning the “Hype Cycle” into a Hyper Cycle.
People don’t buy technology, they buy what the technology enables them to do. It’s a testament of human optimism that everyone—from CEO’s to consumers, want the new-new thing, in this case GenAI, based on the assumption that it will allow them to do something they really need. This has caused analysts, watching from the sidelines, to publish a rather pessimistic description of a new technology’s lifecycle, like the classic Gartner Hype Cycle. Based on a analysis of technologies tracked along the Gartner Hype Cycle only about 22% of them will reach adoption. A large part of that is due to the high failure rates for these technologies face when they are initially trialed and tested.
Rand recently published a research report on why GenAI projects fail, with some estimates placing the number of misfires as high as 80%. Rand identified five leading root causes for the failure of AI projects:
The organization focuses more on using the latest and greatest technology rather than on solving real problems that are important to their intended users.
Stakeholders often misunderstand — or miscommunicate — the problem they want to solve using AI.
The organization lacks the necessary data to adequately train an effective AI model.
The lack of adequate infrastructure to manage the data and deploy completed AI models..
Finally, AI projects fail because the technology is applied to problems that are too difficult for AI to solve.
Mapped on Gartner’s hype-cycle these triggers first heighten expectations and extend the downward slope of the cycle toward disillusionment. For organizations whose expectations of AI exceed its capabilities, their individual hype cycle may in fact prove too great an inhibitor, resulting in their delayed adoption of the technology, or potential abandonment altogether. In the case of GenAI, one contributing factor that has elevated its Peak (and risks an equally deep Trough) is how easy it has been to create a cool demo. Using free tools and with no particular training, anyone can create a demo that supports whatever narrative they have in mind, bypassing data integration, security, compliance, performance engineering, etc.
I have to admit reading the Rand report, it sounded very familiar. Indeed, looking back at previous technological miracles: the internet, mobile, social media, etc., I am reminded of the popular definition of insanity. GenAI is just the latest example of organizations doing the same thing and expecting different results. The issue Rand identifies are fundamentally the same root causes organizations have faced rolling out every new technology since the introduction of the PC.
I believe there is a better path to achieving productivity–and profitability with new technologies. And it should come as no surprise to anyone who’s been following along what the answer is. Design.
Think back to the mobile craze. Everyone had to have a mobile app—even if they had no idea why or for what, everyone was racing to get something into the AppStore. Design was eventually called in after the initial wave of apps crashed and burned. With a customer/user centric POV, design was able to start with the people who had the phones, who were untethered, nomadic, but wanting access to services and information assuming they were scaled to their immediate context, and relevant to their ever changing set and setting. Design created solutions that connected people via their phones.
I believe there is a lesson here.
Developing technology within a human-centered, design thinking process aligns stakeholders, technologists, and customers around a realistic set of expectations that are focused on solving the right problem(s) for your customers. Early design prototypes provide a tangible articulation of the desired outcomes and serve as a common point of reference when evaluating tradeoffs and improvements.
Starting with Rand's five root causes let’s explore the role design can play in effectively mitigating these risks for GenAI, the way it has done for other emerging technologies in the past.
Solving Real Problems
Design’s core value is advocating for the user. Human-Centered Design helps ensure you're solving the right problems in the right way. Prototypes can serve as development roadmaps, evolving with technology and validated by customers for feasibility and viability. For GenAI, this includes finding the right mix of automation and user control to enhance the experience.
Stakeholder Expectations:
Design Thinking starts with users, identifying key problems and creating prototypes to align expectations. Prototypes clarify feasibility, viability, and desirability, providing a clear picture for everyone. At SnapLogic, we've successfully used this process to elicit use cases, create proofs of concept, and identify new development opportunities.
Data Readiness
As SnapLogic’s Global Head of Design, design has played a critical role in developing our AgentCreator, making it easy and transparent to consolidate data from diverse sources, and designing agents to autonomously execute complex processes in response to data changes.
Adequate Infrastructure
User needs can provide guardrails to ensure you have both the capacity and the means to operate your GenAI solution with scalability and growth. By sizing your solution to match your customer’s problems, rather than projected technological trends, design can provides critical input to your growth planning.
Right-Sized Problems
No design ever ships whole and complete the first time. Design excels at breaking problems into manageable parts, allowing for incremental success while still delivering maximum customer value. Applied to GenAI, design can help ensure your solutions' address user needs while balancing your development costs and timelines.
Racing to be the first in their market, organizations are failing to deliver meaningful GenAI solutions. They are repeating the same pattern of investing heavily for the sake of the technology rather than on creating solutions to solve users’ problems. But if we are honest, the real failure is one of learning; these organizations have yet to learn the lesson that design reduces their costs, eliminates risks, and accelerates growth. All things being equal, organizations that include design from the beginning can do more with the same technology than their competitors who don’t.
Remember: People don’t buy technology: they buy what the technology enables them to do.
