Fast, Cheap, and Weird


First, the bad news: Your data isn’t going to save you.

I remember the first time I heard the phrase “monetize the data.” It was about 20 years ago, when a colleague on my corporate innovation team was pitching a new internal venture. After explaining the benefits to the company of the proposed venture, my coworker concluded the pitch with a hand-wavy final argument. He said with an almost theatrical flair, “And then…we’ll monetize the data.” The icing on the cake! 

The Myth of Unique Data

I’ve heard it many times since, and you probably have too. Unfortunately, the promise of “gold in them data hills” rarely materializes in a meaningful way. The reason? Just like everyone wants to believe their child is the brightest in the class, everyone wants to believe their data is uniquely amazing. The truth is that most data is either not as unique as the owners think it is, or not valuable outside the company that generated it. If it were, we’d see more robust marketplaces for data in the wild.

Some believe that with widespread adoption of AI, the moment has finally arrived: big companies are going to win the AI game because they have unique data to train proprietary models. For most companies, and in most instances, this simply isn’t true for the same reasons that data monetization has always been hard. Your data probably isn’t as special as you think it is. Yes, data is the “new oil,” but it’s bubbling up in everyone’s backyard. 

The good news: Big companies can win the AI game, but probably not in the way you think. 

Elliott Parker, CEO, High Alpha Innovation

Three Ways AI Can Transform Your Business

First, let’s establish a framework for thinking about the ways AI can be used by big companies:

  1. To improve efficiency: Many companies have already found ways to improve efficiency in their operations through AI by automating or speeding up tasks. This includes applications like accelerating research, training sales staff, or getting an assist in content creation.
  2. To improve the customer experience: Many companies are actively running experiments to improve the experience for their customers. This is nothing new in the online world; online retailers, for example, have long used sophisticated algorithms to improve purchase recommendations. Thanks to AI, the call center prompts that we all hate (“Press 1 for…”) are already feeling outdated as companies find ways to use AI to rethink how they engage with customers.
  3. To build entirely new, AI-enabled business models: This one is the hardest, because the solutions aren’t obvious, and for most companies, represent a significant leap from what they do today. But this is where the biggest opportunities will be found.

The Early Internet Era

You may remember the early days of the Internet, when magazines were “printed” online as a digital manifestation of the physical experience, with the same layout as the print version, sometimes with an animated page turn that replicated the real world experience. It was hard initially to imagine other formats for “magazines.” This “skeuomorphic” phenomenon arises every time a new technology becomes prevalent: we seek to apply the technology to what we already know. Another example: early movies looked like stage plays, before filmmakers realized the new technology enabled entirely novel ways of changing scenes and following actors’ movement through an environment. We’re in the same phase of experimentation with AI. It’s hard to imagine what the novel models and applications will look like. The only way to figure it out is to run a lot of experiments: fast, cheap, and weird.

AI isn’t just any new technology. It’s revolutionary, because of how it gets to the core of what humans do best – think.

Of course, AI isn’t just any new technology. It’s revolutionary, because of how it gets to the core of what humans do best – think. In many cases even the most basic applications, to improve efficiency or the customer experience, require such a radical overhaul of internal company processes and resources that building something new, outside the corporation, may be required. Building internally is more likely to lead to skeuomorphic solutions–movies that look like plays. Reimagining is hard when there’s an underlying pressure to preserve what already exists.

Case Study: Revisto and AI

Recently, we launched a startup with a large pharmaceutical company to use AI to speed up rounds of legal and regulatory review for marketing materials. The marketing review process can be incredibly expensive inside of regulated companies, and the pharmaceutical company we partnered with had been trying unsuccessfully to shorten the process–without compromising safety–for years. The company’s goal in using AI in this case is efficiency improvement, but to get there we jointly decided it best to launch a startup, called Revisto, with a new AI-enabled business model. Why? Because of incentives, startups can move quickly to figure out what works, without worrying about preserving what came before. In this case, the pharmaceutical company took a minority stake in the startup and signed up as the first customer. The corporation is getting a working solution — within months, not years — and learning from how the startup is applying AI to this use case, all while accruing the benefits of equity ownership. 

This is what rapid experimentation looks like. Startups are experiment engines. They’re built to try things quickly, until they figure out what works. In contrast, big companies are organized to execute at scale and to avoid mistakes. 

The Urgency of AI Adoption

If you want to figure out how to apply AI to dramatically improve efficiency, transform customer experiences, or–more importantly–uncover entirely new business models and growth trajectories, you need to make a lot of cheap mistakes. Go launch new ventures. And do it outside the corporation, where things move exponentially faster and breaking paradigms is built into incentives. You don’t have time to waste! 

Take what you learn from the startups you build and apply the lessons to your amazing, already-scaled business. Your big companies can win the AI race, but not because of unique data. You can win through deep market understanding, access to customers, and ability to apply insights at scale. But first you need to acquire the insights, and the best way to do that is by running experiments, often in the form of new startups.

Venture capitalists and would-be competitors are investing billions of dollars right now to create your company’s AI-enabled replacement, so get moving.

Elliott Parker is CEO of High Alpha Innovation.