Originally posted on Medium
———
Every major company is now saying they have Generative AI and are doing AI. The hype battle is long underway, and the marketing avalanche is in flight.
Industry professionals know it’s merely the beginning of something much bigger that will take time to unfold; meanwhile, most of the market thinks it’s just another thing.
Photo by Christina Deravedisian on Unsplash
Professionals are all looking at it from differing angles and interests, and one can often tell the angle by the role they have at the company. The CFO smells stock success. The Product Team is excited by unimagined experiences or customer success. Sales and Marketing is thrilled to have something new everyone wants.
But every data industry hype cycle has to fall flat. The promises made are simply too high¹ to deliver on! There are too few resources and is too little actual organizational buy-in to meet the vision sold to customers.
Companies bought “Business Intelligence” and “data-driven decision-making” and they got charts and dashboards that 5–10% of their organization uses on a regular basis. But the money had already been paid for licenses and staff; therefore, after all the hype long died there were still people left over hard at work paving the way to deliver on the original promise — modernized data pipelines. This takes time, about 25 years. And the work is ongoing. But now the tooling is closer to the previous hype cycle promise².
In the meantime, Data Storytelling rose and fell as the siren song for data-driven decision-making. It is the grand attempt to focus more on humanizing the data into a concept any lay person can relate to. It’s built on timeless storytelling fundamentals. Storytelling is wonderful and useful in principle and in practice, but this too got so watered down in industry marketing hype and application that the phrase means practically anything to anyone in corporate environments now³.
Now we are at least halfway into the hype of AI, and the result will be the same. AI will rise in the data industry as every company deploys their own version of LLM augmented features that surely will help your entire organization make better data decisions, faster. It has to work!
But it won’t. At least not yet.
The problem isn’t that the goals are generally bad in all these movements. The heart is good but the expectations are all wrong.
Expectations are set that it’s a quick-fix solution. Contracts are signed with many words that mean different things to different people, lack specificity, and the reasonable sense of the commitment required to reach these visions.
Don’t get me wrong — AI definitely has new and unique capabilities to offer. Every day we’re seeing new ways to speed up workflows, to automate content, and summarize or analyze insights from piles of window context. Soon we’ll see the less immediately obvious applications emerging, and everyone will have a slightly helpful digital AI assistant in their pocket. This is exciting, and I’m looking forward to applications that are genuinely improving our lives and goals, stepping beyond the gimmicks or initial chatbots.
My message to businesses is to look beyond the quick fix gimmicks you’re tempted to implement. Do get the minor wins, but don’t think “that is what AI can do” or that it should just solve all your problems. Your core challenges, like company culture and internal alignment, are not going to be changed by generative AI.
Without organizational alignment every initiative fails. Sure, alignment kickoff tasks feel like success in the making — meetings, hired guns, and strategic goals — but every company has its culture. Culture doesn’t change through automation. It is the hardest to influence in any direction regardless of the tech thrown at it, and making strides towards a data fluent organizational culture is no different.
Can AI help make those data-driven cultural strides? Yes, definitely! It’s a tool ripe with possibility. But you can’t skip the hard human steps with a tool. You can’t skip the need for belief from every employee that they are impacting the big picture, that the small decisions matter. You can’t smooth over a lack of vision or unclear execution towards goals. You can’t get AI to write your strategic internal emails or take top-account renewal calls.
You can get AI to accelerate the opportunities for account-based success managers to apply a relational touch alongside project execution. You can use AI to surface unforeseen internal issues, or empower frontline staff to think differently about the small decisions. You can serve up data, but cold data doesn’t change culture. It needs a human touch to shift culture.
The most exciting opportunities, to me, are in using AI to move decision making downhill. A more distributed network of ownership and responsibility is a stronger network, and AI can help us consider loosening rigid, one-size-fits-all structures. For example, we could give teachers more leeway in personalized student learning, or students more ownership over learning that serves a purpose for families, friends or their community⁴. Or Citizens could have greater awareness of city initiatives or community needs, and be given lucrative opportunities to try out subsidized side hustles.
You may have heard the saying “Culture eats strategy for lunch.” Culture changes from the ground, with the people doing the work. If you can empower the ground-level with better data, dignity, and ownership then data-driven decision-making will be easy because its not an initiative — it’s culture.
Those futures are possible and have already started, at least in targeted doses. AI is a tool that can legitimately help accelerate progress arming people with insightful context to impact real change. It’s good work.
Still, my sense is we’re about one to two years away from the trough of disillusionment on the AI hype cycle⁵. This is ok. This is normal. Keep pressing ahead and take advantage of the short term gains. Walk through the doors that are opened, and then keep going. Change is happening. It will just take a bit longer than expected.
Footnotes
Granted, I am a product person, and product teams tend to feel this most viscerally because we want to genuinely accomplish the “what-could-be” of future products.
This article Dashboards are Dead: 3 Years Later by Taylor Brownlow is a good quick retrospective on the industry controversy of such statements.
I don’t fault specific roles for this, because its simply the nature of these cycles. My disappointment mostly stems from implementations falling far short of the potential.
This is nothing new. Project-Based Learning has been practiced in pockets, but at least in mainstream USA, where I’m most familiar, it hasn’t spread much to the public school system.
Only now are major corporations starting to invest in and roll out AI initiatives, and it will take some time before the lack of impact becomes apparent. If you want the in-depth take, check out Gartner.