Updated with comments from Charles Phillips, CEO of Infor
The buzz at Infor’s annual user conference (and a side analyst summit) was around Coleman, the AI platform Infor announced to use “powerful machine learning to improve processes such as inventory management, transportation routing, and predictive maintenance”. The name Coleman is a nice touch – honors Katherine Coleman Johnson, a NASA pioneer depicted in the movie “Hidden Figures”. But I left with many questions about the technology, as I have done with the spate of such announcements from most ERP vendors.
I was more impressed with a few other things at the conference
– I liked a talk by Steve Feilmeier, the CFO of Koch Industries which has recently invested $ 2.5 billion in Infor. Koch is one of the largest private companies in the world with a broad portfolio of investments. He said the investment in Infor was driven by “paranoia”. He cited how Uber had impacted the car rental business, how evolution of seed genomics is affecting farming – in effect, STEM is transforming every one of their portfolio companies. Infor brings them a portfolio of skills and IP which can help the digital transformation of many of these companies. At a tactical level, they will adopt Infor HR, financial and asset management software. But Infor also brings a lot of smart human capital, I have previously written about, including Hook and Loop Digital and Dynamic Science Labs which can help in much more creative transformations.
– I also liked other customer examples I heard of as in Infor’s role at one of the largest mixed commercial and residential developments in the world and at the Commonwealth Games at City of Gold Goast (Australia) helping with among other things utility billing and permit applications. That reminded me of the One Microsoft strategy where a single vendor can bring a wide range of capabilities to bear. Like Microsoft, though, Infor executives are afraid of the tag “systems integrator” and want to continue more with a “build once, sell often” model (and one reason I have not written about One Microsoft in a couple of years)
– I like the progress Infor has made integrating and modernizing two recent acquisitions, Birst, a BI platform and GT Nexus, a network of global logistics, trade and related finance
CEO Charles Phillips summarized the progress his team had made since they arrived six years ago, in the slide below.
They have turned around what was an unwieldy and uninspiring portfolio of products, but I think they could have delivered so much more on the bottom layer – the vertical strategy they embarked on. By now they could have had dominant products in retail merchandising, utility billing, healthcare EMR, hospitality CRS – each screaming for modern architectures and choices beyond players like SAP, Oracle and Epic. Instead, Infor has some basic vertical functionality as with healthcare features in Lawson’s finance and HR modules. That to me is horizontal functionality with a thin veneer of vertical.
In contrast to the products, Charles and team have not been as successful in turning around the customer base (in fairness, other legacy ERP vendors have not done much better). Initiatives like UpgradeX launched years ago should have delivered considerably more transitions than the 8,500 customers it reports in its cloud. The reality is the majority of the Infor customer base of 90,000 (per its web site) is slow moving, and you cannot ignore that when you consider the innovation talk from its management team. Which brings me back to Coleman, the AI platform. Charles: On the 8,500 customers, keep in mind these are business critical applications and our larger enterprise customers. Getting that many in 3 years is a stunning number. And we did while maintaining strong margins and growth. There aren’t too many companies that have done anything at this scale. Whether we have 9,000 or 90,000 total customers isn’t the issue. It’s how many absolute numbers of customers is reasonable to move given the mission critical nature of these apps and how many projects can the ecosystem handle in a year. If you hadn’t heard the 90,000 number, by any measure you consider over 8,000 individual companies in 3 years a success. Each customer needs to assess the new product, analyze customizations, get funding, get approval, review integrations and reports, and evaluate cloud security, and migrate terabytes of data. This is not moving your email. It just doesn’t happen in days. And it’s the larger more complex companies that were more attractive economically so that’s who we are moving first. That’s larger than the customer base of most companies in our industry. It’s absolute numbers not just percentages.
My view is ERP vendors, not just Infor, are out-marketing each other with AI buzzwords and monikers (and players like Salesforce with its own impressive moniker, Einstein). If they were serious about automation, they would already be doing lots more with robotics, drones, wearables and 3D printing in shop floors and warehouses to help customers move to newer advanced manufacturing and logistics models (as I describe in Silicon Collar)
AI has been evolving for decades and the current excitement is we have oodles of data and compute power to teach machines. Infor has been doing custom analytical projects optimizing prices and inventories for some customers. But it is a small sample. The faulty assumption is ERP vendors have access to lots of data. On-premise implementations or single tenant clouds do NOT have much easily accessible data, certainly not diverse enough data that you can get across pools of customers. Then you have privacy and permission issues which further reduce the usable data.
I have a feeling Coleman for a while will mostly deliver on what the press releasesays “uses natural language processing and image recognition to chat, hear, talk, and recognize images to help people use technology more efficiently.” Infor has a strong relationship with Amazon as its cloud partner. It can use its Lex, Polly, and Rekognition AI services. But so can anyone else.
In Infor’s case, you have decent volumes of data in pockets – in the GT Nexus network or collected via ION in customers which have adopted that middleware. So, it will be interesting to see the use cases they do announce for Coleman. There was not much in the way of specifics at Inforum. Given the slow-moving customer base, the caution may be justified. Charles: We also have data in the cloud via Birst. Birst is a pioneer of native cloud BI. They created the concept of networked BI – the ability to create semantic consistency across federated data sources at the source. They ‘ve always had a strategy of propagating data to the cloud from on premise or other cloud services. They then help customers with a wide range of analytical services from visualization, and discovery on the front end to automated data refinement at scale across complex, disparate schemas. So they are used to getting data to the cloud as part of every project. Customers have multiple apps and databases across multiple locations in different lines of business. Moving that data to a central location in the cloud is standard procedures for Birst customers.
Overall I left with more questions than answers – one reason I prefer more dedicated analyst summits. User conferences are for customers – they deserve the executive attention.
(Cross-posted @ Deal Architect)