Despite my incessant buzzword bashing, I'll concede this much: it's important to grapple with next-gen tech via experts who actually know what they are talking about.
We got an earful on day one of the Constellation Research Connected Enterprise 2019 event. Example: most CXOs are not falling over themselves to launch quantum computing projects in 2019, but they do need to be aware of possible threats to RSA encryption:
How quantum computing could (someday) break 2048-bit RSA encryption https://t.co/o5EfaqgcZN
"New study shows quantum tech will catch up with today’s encryption standards sooner than expected" pic.twitter.com/yfOgi9lXoj
— Jon Reed (@jonerp) November 5, 2019
Still, next-gen tech needs to be held to the fire of project results. Blockchain is a case in point. My upcoming podcast with blockchain panel moderator (and critic) Steve Wilson of Constellation will get into that in a big way. That's precisely why a day one CCE '19 highlight was "The Road to Real World AI and IoT Results." Moderated by Constellation's "Data to Decisions" lead Doug Henschen, the panelists shared AI lessons, as they bring tech to bear on logistics problems.
3M on AI - how does a 100+ year manufacturing company stay relevant?
Panelist Jennifer Austin, Manufacturing & Supply Chain Analytics Solutions Implementation Leader at 3M, told attendees why 3M is pursuing several AI-related initiatives. Start with the disruptions in the manufacturing sector:
We're looking at how, as a hundred-year-old plus manufacturing company: how do we stay relevant? How do we keep our products [in line] with consumer changes?
Joking about an earlier panel debate, Austin quipped:
I was also glad to hear that manufacturing is not dead.
As for those AI initiatives, one is a global ERP data standardization project:
As some of the speakers spoke of this morning, we struggle with our data, and we have a lot of self-sufficient organizations across the world. And so we don't have a standard way to represent our data. So we've been on a long journey to do that through our global ERP.
The next AI project? Smart products, such as 3M's smart air filter. The third? Manufacturing and supply chain pursuits, including an Industry 4.0 push:
The third [AI area] that I'm most focused on right now is in our manufacturing and supply organization. One aspect is Industry 4.0, which we're referring to as "digital factory."
We have over two hundred sites around the world, so we're trying to make sure that we have those all fully sensored, and that we're using the data that comes off of those sensors in a meaningful way - to help us with things like capacity optimization, planning and cost reduction, and quality improvement for our customers.
Another aspect of the intelligent supply chain pursuit? Inventory optimization and other "customer value" projects.
The second portion of that manufacturing effort is connected to supply chain, so it's more transactional. That's where we're doing more of the machine learning activity right now. It's focused on things such as optimizing your inventory, by automatically determining what your saved stock should be. It's about minimizing and leaning out your value stream so you can deliver faster to our customers.
This is not a tiptoe into the AI kiddie pool:
We're starting to introduce some exciting new algorithms that are homegrown from our data scientists using, of course, open source models to scale that across the entire operation. So it's something we started out about 18 months ago. [At first], we didn't really think it was real, but it is very much real - and driving results for our business.
Unibap AB - pursuing Industry 5.0 on earth, and in space
Next up? Frederick Bruhn from Unibap AB. Unibap is what you'd call a forward-thinking outfit. In a nutshell, they commercialize so-called "intelligent automation" - both on earth and in space. They've adopted the phrase Industry 5.0 to emphasize the shift from connected manufacturing (Industry 4.0) to intelligent automation.
AI-in-space sounds like a science fiction popcorn movie special. But as Bruhn told us, it's a reality today, and not as different from "on earth" as we might assume:
For us, automation is both in the factories of tomorrow, and in space. Because if you have a mining operation on the ground, or if you have a mining operation on the moon for instance, for us it's the same. So we actually build the server hardware for space, and on the ground, and we do have software to go with that.
One of the cases for Unibap: replacing humans in real-time production lines for painting and coating, assembly, welding, and drilling. No, there aren't any mining operations on the moon yet, but Bruhn says that will happen in about fifteen years. In the meantime, Unibap is supplying computers to customers like NASA, "for intelligent data processing in space."
Royal Dutch Shell on AI - serving customers better is the goal
Deval Pandya of Royal Dutch Shell told us that Shell already has predictive maintenance models in operation, "giving us insights which you can act on to make business decisions and operational decisions, that is generating immense value for us."
The renewable energy space is another AI playing field for Royal Dutch Shell, including solar batteries, and a project to optimize when to charge or discharge batteries. Many of these "AI" and/or IoT projects, despite their focus on automation and "smart" machines, ultimately come back to serving customers better. Pandya:
We've been driving this culture of customer-centricity, and Shell is one of the largest in energy retail. There is a lot of information, and we're just starting to extract value out of it.
Getting AI projects right - talent and culture over tech
On diginomica, we've criticized digital transformation efforts that lack buy-in and total organizational commitment. Yet there is a need for small wins. In that context, how do you get AI projects right? Austin told Henschen: no matter how sexy the tech is portrayed, it's just a tool.
I think that we have less of an AI strategy, than a commitment to delivering for our customers and our shareholders. So it's all about growth and innovation. AI/machine learning has become a tool that we're now more comfortable with. It's becoming a primary driver for helping us deliver on what our agenda is.
Pandya hit a similar note. Royal Dutch Shell has combined their digital technologies into a digital center of excellence:
A big portion is AI or machine learning, but a lot of it goes hand-in-hand. So in IoT, we are using a lot of this IoT data, and then applying AI to it.
I don't care how good your tech is, or how good your implementation partner is, you're still going to face adversity, your digital moment of truth. Henschen asked the panel: what is your biggest sticking point: talent, culture or technology platforms?
Halliburton's Dr. Satyam Priyadarshy says it's the talent. But for Halliburton, it's more of a training problem than a talent problem:
I call it talent transformation. Because we can't go and hire data scientists, right? A lot of us face the same challenge... We compete with Silicon Valley talent as well. The burning talent question for Halliburton is: can they transform the talent they have? The oil and gas industry has one of the most talented workforces scientifically, from geophysicists to geologists, right? So the question is: can they be turned and trained into data scientists? That has been very highly successful; we have been globally training people.
Two companies on the panel, Halliburton and Shell, use hackathons as a means to spark new hires, or upskill. As Priyadarshy shared, their hackathons are a crash course for developers on industry issues:
Our hackathons, or what we call boot camp workshops, are very contextualized and customized. Everybody can go and take a class on Coursera on AI, right? But how do you apply to oil and gas industry problems - that remains a challenge.
So Halliburton designs these boot camps to get geologists and drillers immersed in AI and IoT:
We have a big workforce of drillers; they are actually on the field. We are sitting in the office. So we have to understand their mindset.
For Pandya, culture comes first, then talent, then tech: "culture sets the stage for everything else." But Pandya makes a critical point: if your workers don't feel free to fail, then your culture isn't ready for digital change.
This new technology is changing fundamentally the way we do business, the way we make decisions. And so it is a different mindset... The culture of failing fast and learning from failures is something which we have championed across Shell. It's okay to fail. And that's a huge, huge change in mindset, because when you are putting billions of dollars of investments [at stake], failure is usually not an option.
Most of the panelists are investing in some type of AI/IoT COE (Center of Excellence). Give me a COE over a POC (Proof of Concept) anyday. A COE reflects a grittier commitment - and a recognition of the skills transitions needed. True, not all companies are able, or willing, to build data science teams, but it's instructive to see how approaches like COEs are holding up across projects.
A couple of panelists emphasized choosing the right implementation partner/advisory - that wisdom remains a constant. This panel was a welcome reminder that enterprise tech is at different maturity levels. It's our collective job to push beyond the marketing bombast and determine where we stand. Blockchain and quantum computing remain futuristic in an enterprise context, albeit with very different issues to conquer, whereas IoT, and now AI, have some live use cases to consider. Granted, none of the panelists offered up hard ROI numbers, but that's also a question that wasn't explored, and probably should have been.
Any discussion that comes back to data-powered business models must also return to issues of security, privacy, and governance. That wasn't a focus of this panel, but it was addressed in other Connected Enterprise sessions. My upcoming podcast with Steve Wilson on the persistent problem of identity will dig further.