CIOs have a tough balance to strike: On one hand, they’re tasked with maintaining a large number of applications – research from Salesforce shows that in 2023 organizations were using 1,061 different applications – in varying stages of age, all the while maintaining interoperability and security and reducing overall spend.
On the other hand, they must look to the future state of the business with an eye toward innovation and investment in new technologies like artificial intelligence (AI). While savvy CIOs bring both business and technology acumen to the table, the most successful follow a business-driven IT roadmap, not one handed to them by their ERP vendor. Especially when it comes to AI.
AI requires a shift in mindset
Being in control of your IT roadmap is a key tenet of what Gartner calls composable ERP, an approach of “innovating around the edges” which often requires a mindset shift away from monolithic systems and instead toward assembling a mix of people, vendors, solutions, and technologies to drive business outcomes. And nothing necessitates this shift more than AI.
AI is a generation-defining paradigm shift in the way the world works and lives. The technology has made tidal waves in society, as more than 180 million ChatGPT users tap the fastest growing app for everything from writing term papers to debugging code. And, as explained in Rethinking ERP Reimplementation in the Age of AI, AI is causing significant impact on enterprises worldwide.
While vendors wield the promise of AI as a forcing function for reimplementation, customers who comply with vendor-dictated AI roadmaps likely face four significant challenges:
Challenge 1: Roadmap limitations & delays
How do SAP and Oracle stack up in terms of AI features and functions? In this nascent field, do they have the right technologists, engineers, and product developers to support continuing growth? Are they on the bleeding edge of this technology or are they simply following the pack?
While they certainly could become powerful AI players, successful organizations need flexibility and should be able to select from AI industry leaders for technologies—beyond their ERP ecosystems—that meet business needs today, adopt technology from industry AI leaders that can easily plug into multiple databases across your entire enterprise. Why limit your enterprise’s innovative potential to the speed of a big ERP vendor?
Will Henshall, a writer for Time magazine, reports that AI progress over the past 10 years has been nothing short of staggering. His article notes that over the past decade, AI’s performance has exceeded that of humans when it comes to speech recognition, image recognition, reading comprehension, language understanding, and common-sense completion.
With such rapid development underway, your enterprise must have the flexibility to choose the right AI vendor to deliver the right AI solution at the right time in order to drive the best business outcomes. And while SAP and Oracle could emerge as major AI players, there’s a lot of greenfield out there. Your organization must direct a business-driven IT roadmap to stay ahead of the curve.
Challenge 2: Leaving on-premises data behind
For AI algorithms to be successful, they need a massive amount of historical data to draw from. As Gene Marks, a contributor to Forbes wrote, “For AI to do its job it needs to use data.” Remember the “garbage in, garbage out” adage: The more clean data available to an AI algorithm, the more predictive and fine-tuned the results will be.
Henshall’s article in Time echoes the importance of data for training AI: More than half of the AI models Henshall analyzed since 2020 have training sets of 100 million or more data points. “In general, a larger number of data points means that AI systems have more information with which to build an accurate model of the relationship between the variables in the data, which improves performance,” he writes.
With the high price of cloud storage, customers reimplementing on the vendor’s SaaS cloud might not take all their on-premises historical data with them. We often see organizations migrating only a few years’ worth of data, potentially leaving 10 or more years of data behind—the very data that’s the lifeblood of AI.
There is no denying the fact that with more historical, clean data, the more accurate predictive analytics and data correlation can be. The value of the ERP in AI is the data that it contains, and that already exists today within the on-premises systems. It’s best to ingest the relevant, clean, and accurate data from ERP and other systems into a centralized external AI model for best results.
Challenge 3: ERP vendors’ AI setups only look at data in the system
Vendor-embedded AI typically can only work with ERP data. But there are many data stores across an organization that are independent of the ERP system that should be included in any enterprise AI implementation. So, leaving AI to a single monolithic ERP vendor makes little sense. The good news is that there’s a better way.
You can adopt technology from industry AI leaders today that can easily plug into multiple databases across your entire enterprise This flexibility speaks to the power of having a composable ERP, especially one with a robust data orchestration layer. Making your data accessible across your organization will not only benefit your employees but also unlock new potential for more powerful AI algorithm use inside your organization.
Challenge 4: Loss of license ownership risks cost increases & shrinkflation
In addition to leaving your customizations and data behind, reimplementing on-premises ERP functionally to the subscription cloud could mean leaving your leverage of software license perpetual entitlement behind, which can lead to out-of-control costs and shrinkflation.
According to recent financial estimates from Deloitte, many companies that have moved to cloud have incurred complex software licensing issues and costs that can reach as much as 24 percent of total information enterprise technology spend. Even after initial TCO analysis, “many organizations still encounter a cost explosion when the actual migration begins, in part because they were unaware of the licensing requirements for cloud, which can include licensing transfer, purchasing, and visibility issues,” Deloitte says.
Turns out shrinkflation—the tactic of reducing the size of a product and either keeping the price the same or increasing it—is not only taking a bite out of your candy bar, but also taking a bite out of your cloud. Research by Vertice finds that more than 24% of businesses have been hit by SaaS shrinkflation during the past 12 months, where cloud vendors are charging the same price for reduced functionality.
Examples of SaaS shrinkflation include non-cumulative pricing, reduced discounting, and feature bundling/unbundling. Vertice advises that to be in a strong negotiating position, you should start due diligence 6-8 months before renewal. But ultimately, to secure the best possible price you need leverage. And without the leverage of software license ownership, considerable cost and shrinkflation risks persist.
Ready or not, the AI revolution is here
I think Bill Gates was spot on when he stated: “The development of AI is as fundamental as the creation of the microprocessor, the personal computer, the Internet, and the mobile phone. It will change the way people work, learn, travel, get health care, and communicate with each other. Entire industries will reorient around it.”
The volume and velocity of innovative AI technologies is happening at breakneck speed—a pace that many ERP vendors will likely be unable to keep up with. That’s why it’s imperative for organizations to focus on business-driven IT roadmaps, innovating around the edges of their ERP, and solve the challenges that ERP vendor-led AI roadmaps present. Timing is of paramount importance; successful organizations must act quickly to innovate around the edges and outpace the competition.
Learn more: Discover how Rimini Street can help you reallocate resources to further innovation, gain competitive advantage, and accelerate growth.
