Automate with AI and grow your team

A popular misconception about artificial intelligence (AI) is that it will replace humans. Just like machines on the factory floor, these robots-in-the-cloud will do the jobs of everyday people. While it now seems clear that manufacturing robots do eliminate human jobs, the same is not true for AI. In fact, quite the opposite.


Otto and automation

Otto is a major German ecommerce company, basically the equivalent of Amazon. They partnered with an outside firm and automated reordering with AI. Now, 200,000 orders are placed every month to ensure products stay in stock.

As The Economist reports, rather than reducing the team, in fact they grew their team. Why?

Palantir holds the answer

Palantir is the "unicorn company" that applies AI and AI-like technology to massive problems, especially in the defense and intelligence space. They have always and will always pair analysts with software.

AI software is powerful enough to solve problems that humans couldn't possibly solve–like, predicting sales on 200,000 different products and reordering them as needed, every day, or sifting through billions of phone records to find the terrorist cell–but the AI lacks context. Humans provide this context, and will be needed for the foreseeable future

Not all AI is the same

A form of AI powers autonomous cars–and this is a solvable problem. A stop sign is a stop sign; a pedestrian is a pedestrian. We are within sight of solving this problem completely.

But other applications–like pricing, or forecasting–are not truly solvable. Because they change. For example, in pricing you have substitutes, or different versions of a product. This can be difficult for AI to understand on its own even though it might be obvious to a human. In reordering, you might have a new version of a product–again, something that might be obvious to a human but not obvious to an AI.

AI enhancing humans

The AI deployed by Otto drove massive business improvement–2.5 million fewer returned items, which corresponds to much happier customers and better retention. This grows top and bottom line revenues, and as a result Otto grew its team. Not to mention, the system to manage this requires management itself.

The biggest AI pricing team is also automated

Uber is famous (or perhaps infamous) for it's surge pricing, and yet even though it has been automated since day 1 the team now has almost 100 people (85ish, as of this writing). For something that is AI-driven, that might surprise you. But the fact is, AI needs oversight.

When you think about the problem Uber's pricing team has to solve, your head might hurt. They set prices not only in hundreds of cities, every minute of the day–but also, in each neighborhood or sub-neighborhood of each city. And they do that across a dozen or so Uber "products". Even if you had 1,000,000 people working on this problem, they couldn't solve it. But AI and humans can.

None of this matters, unless...

If you don't have a problem you need AI to solve, then none of this matters. You can continue to export data into excel, make some charts, and estimate what you need to do. Excel (or R, or many other apps) do a good job of analyses like linear regression and more, which for many cases are good enough.

But if you have a complex business that breaks those tools–where you don't know where to start getting a handle on it–then AI may be for you. And if you're a practitioner–that is, not a manager but a do-er–fear not! There is plenty of work for humans in this new AI world.

Get the Ultimate Pricing Book


Related Posts