The Unglamorous Uses of AI

Artificial Intelligence (AI) is often thought of as something from out of this world, with droids, spaceships or flashy computers. Usually people think of AI and cognitive computing as glamorous technologies used for cool things like self-driven cars, robots/ droids, or colonizing Mars. All that is true, but there are very meaningful, unglamorous and fundamentally transformational applications that are low hanging fruits. In reality artificial intelligence is changing the world in many mundane ways.

Let’s take the problem of adapting to changing consumer behavior in the Internet Age. Anyone selling anything online must ensure that their product is competitive — an overpriced offer is not only a lost business transaction, it leaves the customer with a perception that keeps them away from the website for some time. It was easier when there were was one discounts season. Now campaigns and dynamic pricing strategies start at a single click, and the idea of a discount season is only a myth.

Technology gives hope. Industry leaders like Amazon have invested hundreds of millions in developing proprietary technology to address this. As a result Amazon changes prices of its products multiple times a day. Others, like banks with multiple debt offerings use sophisticated models to price them centrally and assume that market will conform. Most players, however, use humans to solve the problem. Not only is that unglamorous, boring and tedious, it is also adhoc and limited in scope.

In other words, it is ripe for real AI.

AI scores over conventional technology in scalability, flexibility and richness of data extracted from any source. The biggest advantage however, is that AI technologies, like humans, can tell good data from bad and focus on the good part. They do so millions of times at once, and they do it fast. Then there are things like summarization of thousands of reviews on every product to distil exactly what consumers are saying or finding different patterns

Then take something like procurement. While it seems so drab on surface, the truth is different. Complexity of the procurement function is probably best judged by the complexity of its problems — tedious processes, fragmented data, difficult compliance to SOPs, very limited bandwidth of professionals leading to anecdotal research at best — essentially utter chaos.

It is a truism that trying to figure out cost controls and risks, while saddled with an outdated procurement / contract management system, makes for a dated player. Precious time and resources are consumed in dealing with doing things manually which result in limited options and sub-optimal decisions.

Now imagine 1,000 interns for the procurement team, powered by artificial intelligence. There are so many things a procurement manager can get done from this team:

• For every part, commodity or service, this team can read through all contracts across all locations and business units of the company to find the best procurement options.
• This team can also weigh these options with user feedback, public reviews, and other factors to further qualify.
• This team can also look at vendors not under contract currently. Some of these options may be better than existing contracts. In most cases this team will generate suggestions on who should be approached for contracts, how much money could be saved, and all the supporting data (See case example).
• Then, when new vendors are being brought on board the team could pull diligence packs them and generate supporting data for negotiations (See case example).

Suddenly, with the use of AI, a method emerges from this chaos. Of course, every decision is recorded, every data point stored and every assumption verified, which provides the foundation for healthier organizations.

As sexy as the technology is, applications of real AI may be completely unsexy. All we hope for now is that we can change the mix of an ecommerce or a procurement executive’s day from tedious browsing to proactively planning the next killer campaign. It is cutting down decision cycles and making startups as well as Fortune 500 players more nimble.