Steve Murphy: Returning the engine to search

Key Takeaways

Epicor is redefining its approach to ERP by integrating generative AI, focusing on becoming a 'Cognitive ERP' that emphasizes action over mere record-keeping.

CEO Steve Murphy highlights the importance of accuracy and relevance in AI outputs, aiming to minimize the risk of 'hallucination' in AI-generated information for better decision-making in supply chain processes.

Epicor's historical strength in manufacturing and supply chain positions it uniquely to leverage AI in addressing contemporary challenges, with a vision to significantly grow its market presence without compromising its tailored service to specific industries.

Epicor CEO Steve Murphy on putting the engine back into search, and solving the supply chain challenge.

Being all things to all people. Does it ever work? Talking to Epicor – namely its CEO, Steve Murphy – it’s a question that comes to mind, with a very simple answer as the response.

First though, let’s talk about the art of finding answers. For the last two decades, search engines have been the usual first port of call for all of us. But with the rise of conversational AI – assistants, chatbots, copilots and what have you – we are beginning to understand the algorithms behind search that have always been there, whilst also being inspired to ask more from search’s sorcery. 

AI has made us realize we don’t – nor can’t – just have the usual offering of all things to all people. Instead, we want to know the right thing as individuals. That means an accurate end result, not an answer that exists simply because its output is necessitated through code. 

This is especially true in an end-user context, whereby an increasing expectation can be found for both record and action; the act of searching records for answers and the affiliated act of generation. That expectation, as we all know, is being heeded and encouraged by the ERP vendors as they tout their GenAI wares to customers. But how many of those same companies can claim to have provided one of the world’s original “search engines”?

I’m talking about Epicor, and Steve Murphy talks to me in the company’s Dublin, California base. As he reminds ERP Today, generative AI still poses the risk of “hallucination,” so if you ask some agents who the CEO is, a mostly correct answer is returned – whilst sometimes a wholly inaccurate one is generated instead.

Murphy explains: “Part of getting our AI right is making sure there are checks and balances around what the ERP system tells you is true and what you should do. So five years from now, you can ask the ERP anything. It’s not just a system of record, it’s a system of action and it’s accurate.”

The exec is confident that Epicor’s AI poses little to no risk of hallucination, which is good news for the end user. In our interview it becomes clear Murphy is confident full stop in the company’s vision of “Cognitive ERP”, a symbiosis between AI and ERP that aims to redefine processes and operations in the supply chain for its manufacturing-heavy customer base.

In some ways, Epicor is making its next pivot. But in others, the tracks have long been laid down, ones intrinsic to Epicor’s DNA so that its move to AI is more of a natural next step coming hand-in-hand with its supply chain prowess.

Pivotal to understanding this is remembering Epicor’s search engine prestige, stemming back to before all things internet and GenAI.

 

Back to the beginning: From print to cloud

Steve Murphy, CEO Epicor SoftwareI’ve seen companies fail because of cloud promises they can’t keep.

The name Epicor dates back to an end-of-the-century rebrand, with the company forming from the financial and accounting software vendor, Platinum Software Corporation, which was founded in its headquarters-to-date, Austin, Texas. Platinum had started developing manufacturer-specific software packages in 1998, and by 2005 Epicor was offering what it called the industry’s first manufacturing solution based on a completely service-oriented architecture.

2011 saw Epicor merge with Activant, a leader in supply chain connectivity and the automotive market. In 1984, the California-headquartered company had brought its electronic parts catalog from print to digital. Listing over 8.8 million automobile parts from engine to exhaust, Activant’s Automotive Electronic Catalog was one of the first digital inventories out there, if not the world’s first for the US automotive sector.

From print to digital, Epicor likewise transformed from on-prem to cloud. Murphy was appointed as CEO in 2017 to solidify the company’s burgeoning SaaS model, joining Epicor from a tenure as president of OpenText, where he oversaw enterprise information management (EIM) solutions across cloud and on-prem. 

“Seven years ago, we [Epicor] knew we had great products and a lot of great people,” Murphy reflects. “But we had to quickly migrate to a best-in-class cloud architecture.

“Now we have [that] architecture and it has dramatically changed our profitability, our growth rates, the scale we’re at and our success.”

April of this year saw the company reach $1bn in annual recurring revenue, with Murphy labeling the subscriptions within Epicor’s cloud revenue as the “crown jewels.” From prior experience, the exec has seen how important it is for customers joining the cloud to understand what they’re going to get when purchasing from Epicor – the buy of a customizable product on a cloud subscription. 

“You may have to adjust your work process a little bit so that it does everything you want in the end, which will be better than what you have [currently.] I’ve seen software companies fail and go out of business because they do not understand this concept or they make promises they can’t keep in the sales cycle and they disappoint their customers,” Murphy says.

The CEO also puts Epicor’s success down to the company being steeped in manufacturing, distribution and supply chain management for over half a century.

 

Murphy’s lore

As a professional, Murphy himself is steeped in that same authority. The CEO’s career began in the 1990s as an engineer for Procter & Gamble (P&G), where he witnessed first-hand the importance of software and mechanical systems for the success of major factories, distributorships and warehouses.

“I was like 22, 23 when I saw how traumatic it was when system failure brought these plants down,” he recollects. “I was working side-by-side with the operators, the technicians, the hourly workers that keep the operation running, and it made a huge impact on me as far as how important that was and how frustrated they were when for no fault of their own, the production system was brought down. 

“In many cases, if the software failed or had a glitch it was because it wasn’t built properly. It really affected everything from the performance of the plant, to the bonuses people earned. So that made a difference as far as what we are really trying to do here when we build a software product and bring it to market,” Murphy continues. “We want something that they’re going to love, that they’re going to think is great and which makes their job easier.”

An MBA from Harvard followed for Murphy, before a role as logistics consultant for Accenture. Murphy remembers one US high tech customer from this time which was sourcing “everything out of Southeast Asia” while an equally-American competitor had clusters of production in the US with only some overseas outsourcing. 

“The question was, if [your competitor can] meet demand based on what comes in on the website on a 24-hour basis, then can you compete with them? And the answer I came to was not unless you change your supply chain.

Steve Murphy, CEO Epicor Software.We’re not everything to everyone – and we’re OK with that.

“The importance of the physical supply chain and the ability to meet demand when prices are at a premium was a huge deal in high tech and it went counter to a lot of the prior wisdom. And then as long as you were cheap enough on the components and everything else, you’d be able to sell it well with high tech [and how] quickly it ages [into] obsolescence. 

“That was a big lesson learned for me around the differences between conventional wisdom, what people tell you should work and what actually does work.”

This period saw Murphy realize the power of ERP, finding it a “confluence of all the things I learned and like.” Working on software implementations thereafter, at the likes of Oracle and more, taught Murphy to never underestimate “how much work it might take to make an adjustment to the software, how it works, or how it presents itself to be easier to the user.”

 

From Cloud to Cognitive

On the AI software front, Murphy knows GenAI can’t be all things to all people, as he recognizes that isn’t what best serves the Epicor customer base. As he succinctly puts it, “We’re not everything to everyone – and we’re OK with that.”

Murphy sees Epicor’s strengths as serving users in manufacturing, distribution, retail and automotive industries who are searching for actions, not just answers. This can often mean a worker on the shop or factory floor using an AI agent for inventory, this agent being able to flag an item that a customer is willing to pay a premium on to receive sooner compared to other clients’ timescales.

Without AI, Murphy argues, this agility may not be as feasible. As he describes it, current actions can mean a product gets sent deep into a 400,000-square-foot warehouse. Once there, it either doesn’t get retrieved or goes intermodal, locked up on a ship or truck to not be received for another six weeks. The same inertia occurs when a defect issue comes up by surprise and the product needs to be inspected. The later this happens in the process, the more costly and loss-making.

“That issue [can instead] get pushed to a production worker or supervisor, who decides to divert the item and prepares the trucks and forklifts. So they’re going to be able to make a change in the process and make more money.”

Murphy explains: “The biggest difference now is the information is available from the point of demand, where someone wants it and they’re willing to pay for it as it’s important to them.”

This is good for the business – and also good for the supply chain. Murphy elaborates on AI’s potential in supply chain, with the flagging of incorrectly assigned inventory and even environmental factors, such as a tornado, which may affect distribution.

“The intelligent agents are that smart,” Murphy says, smart being the keyword. With Dublin being Epicor’s M&A hub, there was nice continuity around our interview with Epicor newly acquiring Smart Software, a Massachusetts-founded provider of demand planning, forecasting and inventory optimization solutions. Smart Software boasts a statistical forecasting solution providing probabilistic AI models which deliver analysis on ‘what-if’ scenarios. These can help users mitigate risk by adjusting stock levels to enable them to meet demand, and suggest actions should something like tornado season be on the horizon.

The offering ties back to Murphy’s comments on the benefits of having a system of action over a system of record, in which a “Cognitive” ERP is provided to users. Discussing this in Dublin with Epicor CTO Vaibhav Vohra, ERP Today delves deeper into the C-Suite’s vision for Cognitive ERP.

For Vohra, the Epicor mission is, “empowering essential workers with superpowers to give them ten times the skills and insights.”

He continues, “What I mean by that is for AI or any advanced technology to be successful, it has to be ten times better than whatever the human does, right? Essentially AI is going to cost more money. The value will not totally be delivered because of this, so you have to invest far more than you ever think.”

 

Handshakes, Gears and Sparks

The CTO sees three elements of the Cognitive journey, starting with “Handshakes” to redefine how humans and machines talk to each other. Next are the “Gears”, essentially low-code tools which make automation and insightful analytics possible on the shop floor. Finally, the “Sparks”, which he dubs insights to the power of ten.

An example is given of one Epicor customer, Visa Cash App RB Formula One Team, based in Faenza, Italy. 

“So we have an AI that actually allows [Visa Cash App RB] to send through all their suppliers and come up with the right trade-offs,” says Vohra. “And then gives you the insight to say ‘you should use this supplier.’” 

Epicor is used to track in real time approximately 14,000 components, so the team will know when they need to make or buy a part. When the make or buy decision happens, Epicor Prism allows the team to quickly determine the best vendor to deliver the part at the right price, the fastest, and if that beats the time/cost of making it themselves, they buy it from that vendor.

In other words, the catalog from 1984 has come a long way, souped up and raring to go in a post-pandemic world. The CTO goes on to explain how Sparks are industry-specific insights powered by AI.

“It could be not just that example of supplier insight, but also a product recommendation for your customer. Steve mentioned turning a system of records into a system of actions – those are the Sparks,” explains Vohra.

The CTO also elaborates on the Gears element of Epicor’s cognitive ERP, which can help end users learn ERP faster. Citing the forecast that four million jobs are going to be created in US manufacturing by 2030, Vohra notes that 50 percent are estimated to go unfilled due to skills shortages.

“It can take someone a year to two to learn ERP. What if we could bring it down? The Gears are meant to create this easy way to level up through an ERP system,” he says.

It’s an alluring proposition, especially when one considers the rise of AI comes with an increasingly apparent skills shortage, whatever the sector. ERP’s role is a crucial part of this discussion, for as Vohra states: “ERP is the vehicle for AI, right?”

 

Epicor’s epic visions 

Steve Murphy, CEO Epicor Software.If anyone wants to copy us – go ahead, good luck!

The message from Epicor’s C-Suite is confident and unified: AI is the solution for today’s supply chain challenges, and Epicor is equipped for the artificial age and the future of supply chain. This is down to an inherent agility as shown by the company’s journey from print to cloud, against which a move into AI looks easy in comparison.

On supply chains, Vohra notes that if Epicor already serves so many customers along the same chain, then the network can be fortified in future by virtue of those clients being in the same cloud, on the same platform with the same AI tools, interconnected without the disruption that comes with software siloes.

Shorter-term, Steve Murphy sees the company in five years’ time as a global leader in ERP, and doing this without needing to become all things to all people.

“We’ll be more than twice the size we are now,” he predicts. “So $2bn in ARR, maybe two and a half to three in total revenue.”

Murphy has evident pride in Epicor, calling his decision to join as CEO as the “best career move of my life; I’ve really enjoyed building a company of enduring value.” This value comes from his pride in “having great people, functionally excellent products [and] deep industry expertise” in manufacturing and supply chain. 

Other vendors, he notes as a comparison, “are aiming really high” in going after huge customers, while offering “not very good” products for SMBs.

In comparison, Epicor, he notes, offers an “industrial-strength product for the high-to-upper SMB market where our close rates are high.”

“And if anyone wants to copy us – then go ahead, and good luck,” he says with a smile. “It ain’t that easy.”

In other words, you can either be everything to everyone – or, like Epicor, be true to your customer. That’s where all the answers – and luck – can be found in today’s AI age.