Speakers:
Katie MacQuivey, Partner, AI, Transformation & Customer Strategy | Sumeet Mahajan, Partner, AI & Data | Adam Wengert, Partner, Cloud Services
KATIE MACQUIVEY: Too often, companies don’t even have the baseline of how they measure their current state, so it’s very difficult to capture that ROI. One main suggestion is to actually look at your current systems, your data, and do that cleansing and hygiene so that you have a good view of your current state of, whether it’s growth with your customers, employee satisfaction, productivity, any sort of financial impact.
That sets the layer that you can then do the comparison on. A lot of times companies are buying 10,000 licenses or 500 licenses, and they just look at the cost of the license. What they really need to look at is, where were you at, and then how much unlock did those licenses get you?
SUMEET MAHAJAN: Many PoCs [proofs of concept] never scale to pilot and many pilots don’t scale to production not only because of all the readiness factors that I was talking about, but primarily because there’s no good way to show, where did I make a difference.
The way to get around this is tying it back to what am I trying to solve at a business strategy level, and intentionally embedding some framework from the very beginning that allows you to capture, what is it that I set out to improve and whether the AI or the data or XYZ investment is actually moving the needle or not.
So just like we do process maps on current state and future state, we have to intentionally embed how are we measuring value today versus how will we measure value once this new solution or new model is in place.
ADAM WENGERT: You need to have a really clear signal about what’s going to impact the business. Being able to do that and then clarify all the way back to the board, “These are the three things that we focused on, these are the outcomes that we expected, and here’s what we achieved,” that’s a huge part of just then being able to say, “And now let me put dollars to that.” Time saved across the board is probably not a great metric, but being able to demonstrate, say, a work product.
I’ve mentioned RFPs earlier, time to RFP, or even just a business process that you can quantify and directly call out, “Hey, we just cut that in half, and we think we can see a way to get it down to 25% of what it was.” That’s a great example of being able to get very specific with the board without losing the fact that you need that signal about what the heck are you doing with AI and why.