Artificial Intelligence: Be Goldilocks

Every story has a moral you just need to be clever enough to find it.
— Lewis Carroll, Alice in Wonderland.

Road tested learnings from running Seldn, an artificial intelligence (AI) startup. Read about the genesis of this series here.

The Goldilocks Offering

Not too low; not too high; your product’s value needs to be just right.

Enterprise sales are an enigma wrapped in a puzzle. They are tricky even when you know they are tricky. I came into Seldn with an unfair advantage: I knew first hand that an enterprise product needs to be way better than just a nice-to-have. Working closely with enterprise startups, as a founding member of Shell’s venture capital arm, had taught me that much. A nice-to-have doesn’t give a startup enough enterprise pull to survive the lengthy sales cycle. Large-enterprise sale cycles (catered to Fortune 500’s) take anywhere between 3 months to 2 years. For a new startup in a new market, it hobbles closer to 6 months — assuming you have a stellar product and team. In a resource-strapped lean startup, with just a ‘nice to have’ product, you simply cannot generate enough client traction. It might spike initially when your customer resources (time, attention & budget) are plenty but quickly weans when they start facing any real business challenges (e.g., a single bad quarter). So the coveted hockey-stick sales trajectory remains elusive. Knowing this, we built a product that taps directly into $100M–1B of annual enterprise value. Here is what happened:

Surprisingly, offerings that tie in directly to massive enterprise value do not take off either. Reason: most big corporations will not trust a new startup with extremely high-value decisions. You’ll see this reflected in two ways. 1). Pilots or field tests take an incredibly long time to deploy, test & feedback. 2). Your customers aren’t tapping into the full-suite of functionality or testing the full range of prediction capability. There is simply too much at risk to test them all at once.

So, too small of a value proposition doesn’t work. Too large of a value-proposition doesn’t work either. Here’s the key lesson: your first product needs to be a goldilocks offering, i.e., build an enterprise product that is way better than a nice-to-have and at the same time, doesn’t map on to 10’s of millions of dollars. Your product needs to be easily tested. i.e., the risk of tapping into the benefits of your product needs to be minimal. When it comes to a successful product for enterprise the mnemonic is “low-risk high-reward”. This is precisely the reason why products that make tasks easier and more convenient take off. E.g., automated report generation or easy dashboards that saves hours of analyses.

Building a product that offers a middle range of value seems contrarian — especially to a tech founder. It definitely runs counter to the founder’s mentality of building a billion dollar startup. You want to tap into value based pricing, your product has to be at least 10x better than status quo, and you want to showcase the sheer power of what’s possible with AI. Plus, investors won’t be interested in anything less than a billion dollar trajectory.

So how to build the right offering?

A successful AI offering needs to have tiered layers of value where the product offerings can be quickly tested in live, low-risk situations. Otherwise, the road to gaining trust in actually using the AI insights will be long and will be riddled with what I call the enterprise catch 22’s.

Read more about the enterprise catch 22 in my next post.

Like what you read? Give Radhika Dirks a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.