In just a few months, the volume knob on Artificial Intelligence was cranked up to 11. But for all that deafening noise, most CIOs are finding AI doesn’t yet live up to the hype. If you’re assessing AI for ERP, here’s guidance on how best to look before you leap.
AI is all the talk right now, but as with any breakout technology for enterprise applications such as ERP, it takes time to understand the potential benefits and challenges, infuse the new tech or capability into existing code and processes, and/or deploy new ones that leverage the capability while reducing the new risks it creates.
As Pat Brans stated in a recent CIO article, “Most CIOs have begun exploring generative AI to make sure they stay relevant. But many are finding that the technology on the market doesn’t yet live up to the hype.”
Expect AI to be transformational but it’s early days for ERP
We spoke with several Rimini Street clients and prospects regarding how they use generative AI for ERP. What we found aligns with Mr. Brans’ thinking. There is a lot of enthusiasm but also a lot of “tire kicking”. While there appears to be tremendous interest in AI, it’s difficult to find ERP-specific use cases that are operational, much less proven.
Virtually everyone we spoke with expected AI to be a game changer for their organization, but they were still in the planning or analysis phase for where and how AI could improve enterprise application outcomes. McKinsey estimates that, “…generative AI could add the equivalent of $2.6 trillion to $4.4 trillion annually…this would increase the impact of all artificial intelligence by 15 to 40 percent.”
What we didn’t find was any ERP leaders who were sitting on the sidelines waiting for the technology to mature. This is in direct contrast to other big ERP changes like the SaaS ERP model, or like moving from the mainframe to a client/server model, where it took years for most organizations to embrace the new technology.
For ERP, generative AI initiatives are occurring at the edges of the core system
For enterprise applications like ERP, the code has matured to a point where it’s generally a good fit for conducting back-office operations. The mechanics of recording business transactions (e.g., purchases, payments, tracking inventories) have been optimized. Conversely, customer experience and knowledge handling are key areas where early adopters are seeing results that improve business outcomes. Analytics is another one, and all happen at the edge of core ERP.
It will take a while for the ERP vendors to catch up
More than any breakout information technology since ERP’s inception, AI has the potential to change how business works. Additionally, it appears to be doing so more rapidly than previous technologies given the avid interest organizations are showing. Changing the fundamentals of doing business will eventually force enterprise software vendors to rewrite their core ERP code rather than providing tools to build features around the edges as they are initially doing with generative AI. For now, however, most ERP customers are safe to keep running their deployed ERP systems while they innovate at the edges.
Although ERP vendors like SAP and Oracle are introducing AI development tools, it might take some time for them to figure out how to package and price and leverage AI in their software products. Early adopters of interim offerings run the risk of getting caught in the quagmire of rapidly evolving pricing models and availability as AI offerings mature. For example, SAP S/4 on premises and private cloud customers risk being left behind. According to IT Jungle, “All new ERP capabilities, sustainability and carbon accounting solutions, and new AI innovations, ‘will only be available in the cloud and delivered via RISE and GROW with SAP.’ (quoting SAP CEO Christian Klein from the company’s Q2 2023 earnings call).”
Two big challenges stand out for ERP vendors and customers
1. ERP vendors risk losing the initiative as they attempt to infuse common AI functionality into their flagship products.
As generative AI use cases evolve for ERP and users become savvy about using AI development tools to create new code that improves how business works, a spaghetti soup of AI apps that individualize ERP deployments will result. Vendors will need to rethink their definition of standard functionality and core code (an enormous and time-consuming task). Brian Sommer states in a recent Diginomica article, “…vendors don’t have any firm ideas yet on what they will develop (using generative AI), when it will ship, what privacy issues they’ll expose your firm (and its data) to, how they’ll sell and price it, etc.”
It’s reminiscent of the best-of-breed quagmire that ERP was introduced to correct, only worse since the AI code will be unique to each organization. The rapid adoption of AI will put ERP vendors at risk of losing the ability to control the definition of the next generation of ERP.
2. Replacing customers’ unique, fit-for-purpose AI solutions with the enterprise software vendor’s AI offerings could become too pricey and complex.
For ERP customers, generative AI tools allow for creating the ultimate in rapidly deployed custom code that surrounds their ERP system. This will exacerbate the customization factor. It will be hard to build a solid business case to eliminate the custom code and replace it with an upgrade to their ERP vendors’ next generation code set, particularly when the functionality offered by the vendors will have been designed with a one-to-many solution rather than the one-to-one of custom AI built by each organization.
AI can be a good fit for ERP but beware the risks
As with any emerging technology, generative AI carries some significant risks, some of which, according to PwC’s Managing the Risks of Generative AI, include:
- Risks to privacy
- Cybersecurity vulnerabilities
- Regulatory compliance over data used and AI-generated outcomes
- Third-party relationships – who owns support and maintenance, error resolution, testing, etc.
- Legal obligations for data and solutions
- Intellectual property rights
The path to success with AI is still being paved and the unknown could be costly or disruptive in a bad way.
Recommendations:
- Take the time to develop a risk assessment and risk mitigation plan for each AI initiative. Establish conditions for when to pull the plug or accelerate further.
- As you take the AI journey, assess how adding AI components into your ERP portfolio will impact your ability to provide cohesive, effective support and services.
- Use a unified approach to make running and supporting the portfolio simpler and more scalable by reducing the number of vendors, products, and support service layers involved. To add to the value proposition, outsource the unified support and services but plan this move carefully.
