XSelect a record to create:
IIoT Spotlight Podcast EP027b: How to Monetize IoT Data & Use AI to Accelerate Businesses: An Interview with Peter Bourne of BrightWolf.
Published on 02/21/2018 | Strategy
Peter Bourne is the CEO of BrightWolf, strategic advior for ProAxion Incorporated, and an Entrepreneur in Residence for Blackstone Entrepreneurs Network. Prior to Bright Wolf, he spent 15 years in Director, CEO, COO, EVP and GM roles at both public and private companies, from early stage through revenue to profitability and exits.
Peter’s career covers a breadth of industries experiencing digital transformation, spanning all company operations including P&L, sales and marketing, and R&D. He has spent 20 years in CEO, COO, EVP and GM roles at both public and private companies, helping to commercialize new technologies, and developing companies and products for success.
Bright Wolf has almost a decade of experience as a system integration and technology partner for Fortune 1000 industrial equipment providers seeking digital transformation through adaptable connected product solutions. Bright Wolf's pre-built components, open system architecture, and expertise with industrial controls, protocols, and embedded systems enable rapid delivery of flexible IoT products and services that generate business outcomes for customers.
IoTOne.com post: https://www.iotone.com/podcasts
Welcome back to the Industrial IoT spotlight. I'm joined today by Peter Bourne. Peter is the CEO of Bright Wolf. It is an interesting company -- they are an enterprise software company that also does very deep system integration support for their customers. they get very involved in the business, not just on the technology side of the IoT solutions. Peter is also the strategic advisor for ProAxion Incorporated, as well as an Entrepreneur in Residence for Blackstone Entrepreneurs Network. And Peter you have a really deep background in in a range of technology companies. Peter thanks so much for taking the time to speak with us today.
Thank you it's great to be here.
You say very clearly “build value via infrastructure” and so my understanding is that you're helping companies by providing this 70 to 80 percent of the common infrastructure so that people are not wasting resources reinventing technology that's out there. And then focus on building the customizable value added components of the system.
What does your technology suite look like? If you break it down and then you can focus maybe on some of the areas where you really differentiate and other areas where you might work with partners.
We bring three core components let's say in the technology space to the table.
One is what the cloud people call the edge, but what our customers view as their product and the real world, there are many “system integrators” for IoT who don't really bring in embedded systems, and industrial control, and industrial connectivity and protocols experience to the table.
Most of our customers’ projects fall out right there where it can be quite difficult to actually get the data from the machine or device in the first place. We bring some industrial protocol and some edge analytics and some edge intelligence runtime and configuration tools, basically a tool kit to the table at the edge to facilitate the activity for our customers.
Two, all IoT projects require data management capabilities. We have some quite deep IP as it relates to time series data management, in the form of some software that we've written, that is adaptable across cloud platforms or, even in some cases, on-premises deployments. So it's pluggable to Google Cloud, to AWS,
and to Microsoft, both in terms of the storage and in terms of taking advantage of the internal services that they provide and some of the things that they've been starting to add over the last year or two to their IoT suite. So that's the second class of IP program.
The third class is a reference application and it's really all of the common backend components you're going to need to build your IoT application. We bring this reference system to the table; we use it to set up and build the IoT infrastructure for them. Either us or somebody else can build the UIs, mobile and web UIs, to go on top of that, and at the end of the engagement we provide a transition service to our customers that where we give them the source code to that reference application and to the reference system. If you want to be a software enabled, a data centric organization, we are going to make a transition path to that for you. So unlike a product company that just wants to sell you products or a consulting company that just wants to sell you hours, we actually want to get your system up and running and then get you in the driver's seat of that system going forward.
I know the answer is going to be it depends but when you do that handoff of the source code, to what extent are you finding that companies are able to take that and with their internal teams really evolve the product and own it, and to what extent do companies either require you or maybe another partner to be involved on a very regular basis? I'm sure that these products, as they build these systems, they're going to be constantly updated as the business environment changes and as they learn more about the requirements of the system.
That's right. So we take a step by step approach to handing off the source code. There are big systems are at the heart of our customers’ product offering so they are supercritical so we break it out into the constituent components of the system, and sort of hand those off as fast as our customer is capable and willing to take them. The common elements of that, we continue to version – we version control and continue to provide them updates too. Our involvement in the project decreases over time but never quite goes to zero. We want them moving as fast as they possibly can be. For them to move at the pace they want to move, they build that capability internally, they hire the right people. Our involvement moves to more sort of monitoring the architecture and then providing them updates to the baseline code as our roadmap evolves.
Next question. This is a bit it's a little bit tangential but I'm interested in your thoughts on this because I think this is probably something that you've put some thought into.
IoT exchanges: everybody is kind of talking about the value of data, and data is the new oil and so forth, and companies are doing probably a pretty good job now of collecting data and storing the data, and to an extent analyzing some of that data and making use for very spot uses in their operations or for their customers. But they're not doing a very good job of transacting that data and providing value to the marketplace of other organizations that might extract value from the data and actually monetizing it.
I've talked to some companies recently that are working on the data exchange and the data marketplace problem. What are your thoughts on this? Do you think that this is feasible? Is there a timeline where you think that corporates will be transacting the data, will have the technology, the business and the policy frameworks? Or do you think that this is, in seeable future, really going to be a lot of one off relationships between partners as opposed to more of a market marketplace solution?
That's a great question I have and I don't have a crystal ball on that. I have an opinion like everybody but happy to be proven wrong about this. You said something about the business and the policy framework. Data exchanges are a worthy technology challenge but not an eminently solvable problem at the technology level.
Let me tell you the coolest story that blends two massive customers that we are engaged with and their sense on this route of getting my opinion to tell you anecdotally what we're hearing. Both of the companies buy equipment and then operate manufacturing or production facilities. One is in oil and gas and the other one is in agriculture. Their equipment providers want the data from the equipment, so that they can do a better job of operating their facilities.
The discussion that they have with their equipment providers goes something like this, the equipment provider says: we want the data off the machines, but the prices on the hardware that we sold to you in the first place do not alone support the business; we've had a lot of pressure on the prices are on the initial price of the equipment. A large portion of our business health is supported by the parts and services business that we run after you've made the initial purchase. If I give you all the data from the machines, I'll be out of the parts and service business and then I will be able to sell you equipment anymore. For me to give you the data that you need to have a better operator we need to have a commercial agreement for that; it’s not something I can just hand to you because it makes my business not viable over time.
I think that's the challenge. The people I talked to who are running offshore oil rigs in that example estimated that this was going to take somewhere between five and seven years to work itself out, to where they were able to manage that dialogue with their suppliers, and make everybody in the ecosystem healthy because nobody wants the supplier to go out of business. So their estimate was this is going take a while for us to figure out a new business model around that.
That makes sense. It's going to be a series of experimentation and then eventually companies will figure out something now don't begin to be adopted.
Yeah there'll be some success stories in the next year or two about how it's working great. I think there'll be more in the framework that you suggested where it's a tighter ecosystem where it's not a full open exchange there'll be more failures and successes for a while. And then industry by industry, answers to that will emerge. Finance is an interesting place -- there is some comfort now with buying products and sharing data on especially around risk in the finance markets. Ten years ago, there's no way you would have said that that was going to happen.
You mentioned that this is a solvable technology problem. What are the technology problems now that you're really interested where were you guys are working on it, or other companies are working on, and you think there's a real differentiation opportunity, if a company is able to solve a particular technical challenge in this space?
The poster child for that right now is AI and Machine Learning. So the ability to take enormous volume, velocity, and variety of data that comes from connected devices, and in a compute way, makes sense out of that. There is a lot of discussion around that and some very public case studies, but it's early in that. I think there's a great opportunity for people to be able to differentiate themselves, both in terms of being an equipment manufacturer, and/or all these data exchanges that you're talking about, or companies like us and product companies as well, to be able to differentiate themselves in terms of how they approach that problem.
One of the insights that we've had over the last six months is many of our customers have aging field technician or courses and the generation that's coming up behind them to replace them doesn't have the know how yet, which is normal. Everybody always inserts a millennial joke there but I think you know there's always a learning curve. Particularly now, AI may have the capacity to accelerate that learning curve and to ease the transition of the aging out of their current field tech force and the know how that they have. But to do that you have to insert the fuel tech force into the learning loop; you have to make it human based. That's an interesting challenge – it’s a technology challenge, it's a human factor challenge, and it's a huge problem facing many of our mechanical industrial oriented customers.
More deeply inside the system, there are two places where there's a lot of innovation happening, AWS and Microsoft are both great forces for moving undifferentiated heavy lifting forward. It doesn't get as much press; it's not as exciting as AI now, but it is no less necessary. In fact you can't do AI until you have a proper data management infrastructure an approach in place. Garbage in garbage out. So there's a lot of work being done, down in the basement, down in the boiler room in these IoT systems to figure out what IoT really comprises of in the next system of the intelligence or system of record inside the enterprise that is actually connected to the product. As I mentioned the data, the velocity, the variety, and volume there are way higher. So there's a whole new order of magnitude challenges that need to get solved there we are. We have IP in that regard, others do as well. It gets way less press but successful IoT projects require it. So there's two places where the innovation happening.
I was just interviewing a candidate a couple hours ago and I asked her, why are you leaving your current job. She's working for a good company in a good position; I'm sure her salary is fine. So she said I think I'm going to be outsourced to an AI within a few years. She's doing kind of back data management for business data, and that that was an insightful answer that she's really thinking, “if I keep specializing in when I'm specializing in now, it's on the downslope”.
In AI, I guess there's kind of two high level two approaches. One is the more the more general approach that a couple of the larger companies like IBM’s Watson are approaching, and then there's a lot of companies that are building more specific algorithms for spot problems. What's your take on this? Do you think that most of that is going to be provided by these more spot focused companies, or do you think that at some point a couple larger companies are going to kind of crack into a more general solution to the problem and, based on the volume of data, they'll be kind of having scalable solutions that apply that to many of the potential cases?
Awesome question. I have several thoughts there. As it applies to horizontally thinking about scalable AI solutions, I think there's a great opportunity there for quite a large business to be established. From what I can see today, I think folks like AWS and Microsoft are in the best seats for that because it would take substantial funding and a substantial reach to be able to do the more generalized thing that you mentioned there, so I think it's going to take a while. It took AWS five years to build Aurora, and these are people that are cranking out over a thousand new services a year. These are deep problems and so you know you can take at a whole other magnitude and it is going to be at least that hard to solve a general AI applicability problem. So to your question, the idea that these are going to be solved in more specific cases seems more apparent to me, at least in the near term.
As I mentioned it it's really early in that space; there's a ton of hype around it right now. IoT itself, interesting enough, was at the top of the Gartner hype cycle curve for three years which, according to a guy I talked to at Gartner about that, was as long as the Internet itself was at the top of the hype cycle. I don't know how long AI is going to be there, if it's not there already, it'll be there for a while also. Then as soon as, of course, it comes off the peak, we'll start to see actual real solutions as it always happens. For us, our approach mentioned algorithms and those kinds of things -- I think in the algorithm space, there's a much more near term and tangible opportunity to generate useful and more broadly applicable approaches there for training AI models and those kinds of things and for analyzing data. There are several companies out there already making quite a bit of headway in that. That’s probably one of, if not the, next wave and this is solving more broadly applicable business logic in the form of algorithms.
Our approach is to integrate with those; we are not data scientists and we're not trying to be. We think it's important but that's an enterprise system in and of itself, as it matures, I think it needs to be connected to your IoT system in ways that make it useful.