Can algorithms re-plan our enterprise supply chain in real time?

Transformation w/Tony Saldanha

Tony Saldanha

Sponsored Content

In early 2015, when I was VP at Procter & Gamble’s multi-billion dollar Global Business Services and IT operation, I was faced with an ironic question. What do you do when you’re supposed to be best in class in the industry but you don’t want to rest on your laurels because we live in a disruptive era. How do you create the next generation of capabilities? And is it even possible to translate all the digital capabilities out there into a next gen set or do you risk wasting time and resources trying to create a future that’s outside your grasp?

Next generation

At P&G, we set up a small team called Next Generation Services (NGS) and reached out externally to learn about 10x idea possibilities from more than a hundred organizations—consultancies, peer companies, venture capitalists, start-ups, educational institutions, and futurists. The exercise was a roller-coaster ride that was both thrilling and terrifying, with periodically stomach-churning jolts of new digital reality that had us questioning whether the digital revolution wasn’t even more urgent than we had thought.

Digital Assistant

One such experience occurred in April 2015. I was trying to schedule a meeting over email with AJ Brustein, the CEO of a start-up named Wonolo. I sent AJ an email message suggesting that a P&G colleague and I would be open to a phone call. AJ responded, “Sounds great. Copying Amy as well. Thanks.” He copied Amy Ingram, who I assumed was his assistant. This was on April 10, a Friday. To clarify availability, I replied to him and Amy, “Thanks AJ. I am out next week, but perhaps we can connect the following week?” and copied my admin assistant Kim on the message. Later that day, we got a message from Amy: “Happy to get something on AJ’s calendar. Does Monday, Apr 20 at 11:00am PDT work? Alternatively, AJ is available on Monday, Apr 20 at 4:00pm PDT or Tuesday, Apr 21 at 10:00am. I’ll include the dial-in on the invite.” Kim chose a time over email, and the meeting was all set. It was just a routine day at the office. However, later that day, a P&G colleague copied on the messages asked me to check out Amy’s digital signature on her email. It said, “Amy Ingram | Personal Assistant to AJ. Brustein” and below that, “—artificial intelligence that schedules meetings.” Amy was a robot!

Turing Test

I was flabbergasted! This was a “Turing test” type moment for me—the test, named after Alan Turing in 1950, being a challenge of a machine’s capability to exhibit intelligent behavior that’s indistinguishable from a human’s. We dissected Amy’s responses carefully. Her messages were in perfect business language. “She” had clearly “read” and “understood” my email on April 10 that I wouldn’t be available the next week and had therefore suggested times on April 20 and 21.
If a robot could manage the most personal type of executive service, then why couldn’t AI run so-called “judgment-based” financial decisions with suppliers and customers on accounts receivables and payables? Why could we not supplement our buyers in the purchases function with an AI “buddy” that could digest the latest information on suppliers, materials, pricing trends, and payments and trigger advice and decisions? Could we not redefine the traditional corporate systems user experience from Stone Age to Siri? Could we perhaps forecast and proactively self-heal most IT systems outages across the P&G globe? Could algorithms re-plan our supply chain in real time?

The game is on! 

Over the next few years we would go on to run more than 25 disruptive projects in these areas. Each had to be a 10x idea (I.e. ten times the improvement, not just 10%). And each had to meet a financial threshold that was several tens of millions of dollars in financial return. It was hard work but it was fun! The extremely successful work of NGS proved that it was indeed possible to create the future of shared services. Or to be precise, we had to find it because it was already out there, but not evenly distributed. There were clear use cases already there, as Amy the robotic admin demonstrated. 
Editor's note: This article is a specially modified extract for the AIIA network from Tony’s new book “Why Digital Transformations Fail”.  The book has been rated #1 on Amazon’s Hot New Releases List for organizational change. See more here....