A BUSINESS-PRACTICAL WAY TO THINK ABOUT AI

A salesperson representing a company that deploys artificial intelligence solutions walks into a CTO’s office one day and starts talking about the company’s product. You’re the executive assistant observing the conversation. What do you see? Can you see an animated individual talking incessantly about how great his company and its products are, and another one nodding and looking knowledgeable while not seeing the relevance of any of that to his business?

This scenario plays out everyday somewhere in the world, and it’s not uncommon for such meetings to end with the prospect holding an even bigger bag of questions than before. Questions like “How appropriate is this product or service to my business, specifically?” are often left unanswered. Nothing like “It will reduce your manpower requirement by 32 FTEs” or “This will speed up your average response time by up to 23%” ever breaches the surface.

If you’re the business owner or executive responsible for the decision, questions like that might leave you wondering whether it’s worth the effort at all. More often than not, the salesperson will gloss over many of the challenges, which only makes the decision even harder.

To help put your train of thought on the right track, we’ve identified some key elements that any business eyeing AI deployment needs to think about no matter what its size. Hopefully, these points will help clarify your position on AI and whether it’s really viable or even necessary for your organization.

First things first.

QUANTIFY ITA lot of businesses fail to calculate the benefits of new, AI tech at work in a tangible way.

For example, if you’re responsible for customer support at your company, you need to ask how much AI chatbots will help reduce your issue resolution time.

If you operate an online store like Amazon.com, you need to know if a machine-learning-based inventory management system bring down your backorder levels or prevent the system from displaying out-of-stock items. Will customer ratings go up as a result? That’s the sort of tangible measurement that will help you develop your digital work.

It’s the same when adopting any new technology, like moving to a cloud computing environment. Moving to the cloud is a good thing for the most part, but unless you know exactly what workloads should be moved and how that will materially impact your revenue or other key metrics, it’s only going to be a trial and error exercise.

Ask metrics-related questions to help you pinpoint the areas that can positively impact your business. If you’re measuring a specific metric like backorder levels, the AI system you’re considering should ideally move the needle for that metric in the right direction, and considerably so.

Just as you quantified the benefits, you also need to consider the cons.

UNDERSTAND THE DOWNSIDEAI is a sensitive topic because of the perceived threat to jobs. What’s good for your business metrics might not be too great for your company when it comes to attracting new talent. If a lot of your business depends on bringing in the right people and your company is known for rapidly deploying efficient automated systems that result in job cuts, you might end up facing an HR crunch or labor unrest at some point. Here are some examples:

In early June 2018, casino workers in Las Vegas threatened a city-wide strike against companies like MGM Resorts International. One of their sticking points was the increased level of automation threatening their jobs.

Dock workers around the world regularly call for strikes because of the rampant automation efforts by freight companies and port authorities. In reference to the labor strike in Spain on June 29, 2017, the following comment was made by a leading port technology portal:

“Given the contemporary state of globalised business, it also means we inhabit a world of globalised unions, and with Spain seeking to gain Europe-wide support for its tangle with the EU, it is not impossible to imagine a much larger response to the burgeoning trend of AI automation in the near future.”

After a conversation with several human resources department heads, Deloitte vice chairman and London senior partner Angus Knowles-Cutler said,“The general conclusion was that people have got to (come to) grips with what the technology might do but not the implications for workforces.”

You need hard data to help you make the final decision, especially when facing strong opposition from the company’s stakeholders. It makes it easier to filter out unwanted investments that could hurt you versus those that can take your business to the next level in a positive way. In a future post, we'll explore an example of how one company implemented automation too quickly and too widely, and finally called the whole thing off.