Some Nascent AI Startups and Hype
Note To Readers: The observations in this blog post are based on meetings with nascent AI startups I have met for the first time. They are not established or funded companies. These observations have little or no congruence with the more mature companies I advise and have invested in. Therefore, please consider these observations as those from the wild field of dreams inhabited by zero- and early-stage startups.
I recently attended several AI startup meet-ups in New York and Massachusetts. I was inspired and amazed, yet also conflicted, all at the same time. Encountering a lot of hype from leaders of these companies led to this blog post.
The Gartner Group knows how to separate the hype from the reality. Gartner’s Hype Cycle was developed to separate myth from commercial reality. It conveys the phases of maturity and adoption for new technologies, and shows how they function both as solutions to real business problems and opportunities to exploit new market opportunities.
Think of the hype cycle as a sort of “technology life cycle” divided into five key phases:
Innovation Trigger: A new technology elicits early proof-of-concept stories and media interest, resulting in often outsized company claims and lots of PR. At this point usable products exist, but commercial viability is unproven.
Peak of Inflated Expectations: The original laudatory stories in mainstream media and/or trade press are followed—sometimes immediately—by failure stories. As a result, while there is an early adopter market, many potential customers defer actual purchases and become late adopters. This group may also simply reject the new tech altogether.
Trough of Disillusionment: Failure to live up to the hype results in waning project interest. A shakeout of related vendors can occur, while others might fail outright. Angel and VC funding continues only if the surviving vendors produce a community of satisfied early adopters, otherwise the cycle starts over again after some intervening years.
Slope of Enlightenment: Use cases become more widely understood and more enterprise adoption occurs. Vendors produce new features and more complete solutions, which are themselves regarded as next-generation product improvements.
Plateau of Productivity: Mainstream adoption starts to take off. “Bake-offs” become more common, and more clearly defined criterion for vendor assessments become widely available. (I think of this as the “Life is Good Phase,” for obvious reasons.)
The Early-Stage AI Business Today
Through my recent conversations AI startup leaders, it’s clear they face consistent challenges with their data sets, time-to-results, and funding.
AI startups follow clear and well-defined paths initially:
A team is formed and completes basic market and technology research
The AI technology is developed—often offshore to save $$$
The team secures one or more “beta” sites or proof-of-concept (POC) projects
These “beta” sites or POCs are with potential customers that vary from small, medium or large in size and challenged with their data sets, and results take longer than planned.
Still, based on their situational analysis, AI startup founders think they have a “good business”.
Here is the honest reality:
Good market and technology research doesn’t always lead to great products
Good market and technology research, combined with good products and good businesses, don’t always lead to products that are easily marketed or sold
Good market and technology research, combined with good products and good businesses, don’t always lead to businesses which scale
Good products don't always make great businesses, and
Good businesses don’t always lead to great exits
Almost all early-stage companies I’ve encountered lately in the AI space underestimate the complexities, costs, support structure requirements, and great humans-in-the-loop realities of sales and marketing. They struggle to execute the company’s go-to-market (GTM) plan. Transitioning from good to great is challenging for all companies, but for AI startups hype to substance appears to be an absolute built-in part of the journey.
As a former CTO I often share my perspective with other first time CTOs. And though it pains me to say it, I most often have to advise them that at the end of the day the technology doesn't matter. Yes, it has to work well, must be high quality, etc. But more importantly is the market opportunity and GTM strategies to build an impactful business. CTOs need to understand and contribute toward that success - their engineering perspective might actually provide insight that is helpful to the sales and marketing functions.