AI shortcuts won’t replace enterprise fundamentals…
Neha Ralhan has some bad news for companies looking to ‘vibe an ERP’ rather than buy a new ERP: It’s not going to happen. But vibe coding will play a role in ERP going forward.
The Gartner senior principal analyst says she’s getting plenty of enquiries from Australia and New Zealand about whether they can use vibe coding to avoid a large transformation with current or potential new ERP vendors.
“If anything, ERP is going to become more prominent, more important and more central to your organisation.”
“The answer is no,” Ralhan told iStart.
Despite the hype surrounding AI-generated applications, Ralhan argues that replacing ERP with vibe-coded software is not a realistic enterprise strategy. Compliance, governance, auditability and operational risk matter.
“It’s not a viable option. It’s not an enterprise option and it’s not an option for organisations that are highly regulated – and every organisation nowadays is highly regulated,” she says.
Part of the problem is that many organisations are mistaking impressive AI demonstrations for production-ready enterprise systems. Experience is resetting the perception about the cost to build software.
“People think of AI as a silver bullet,” Ralhan says. “But there’s a lot that needs to be done right and correctly in order for us to use these tools.”
While perception about build costs may be valid, Gartner says there is significant misalignment in expectations around the gap between demonstrating vibe-coded assets and deploying them in production, and the difference between running personal tools and delivering enterprise-grade software.
ERP is getting bigger
But that doesn’t mean enterprises should dismiss vibe coding altogether. In fact, Ralhan believes organisations that embrace it in a controlled way may ultimately gain an advantage, accelerating ERP innovation, improving delivery quality and, as a result business capability, reducing technical debt and maintaining greater control over increasingly AI-driven ERP landscapes.
“Organisations that are able to incorporate vibe coding in a very organised, governed, sensible way are going to come out on top.”
A new report from Gartner, Extend Your ERP with Vibe Coding, co-authored by Ralhan with Johan Jartelius, says application leaders should establish vibe coding as a core capability within a program of AI-generative delivery practices to ensure ERP robustness and flexibility.
In fact, Ralhan says AI may actually increase ERP’s importance, rather than reduce it.
The end of one-size-fits-all ERP
Historically, ERP has been viewed primarily as a system of record, the place where organisations store financial, operational and transactional data.
With the evolution to Enterprise Resource Execution (ERX) – Gartner’s emerging framework touted as the next iteration of ERP – ERP becomes a system of execution.
“The time of thinking of ERP as shorthand for a finance system is over. If an organisation is mature and brave enough to go on the journey, it can provide a system that actually fits the changing direction of an organisation.”
In this model, ERP becomes the operational engine powering workflows, decisions and increasingly autonomous processes and increased agility.
“If anything, ERP is going to become more prominent, more important and more central to your organisation,” Ralhan says.
“With that we need to start looking at different elements that can bring it about, and vibe coding is one of those capabilities.”
As the era of the monolithic ERP one-size-fits-all off-the-shelf system comes to an end – something Ralhan is clear is happening – vibe coding can help organisations tailor software around their unique requirements. Rather than replacing ERP, AI-generated development can extend it, allowing business teams and employees to build targeted solutions for specific needs.
“It shouldn’t be the centrepiece of your ERP, but it should be additive,” Ralhan says of vibe coding. “Treat it as another tool.”
The goal, she says is not to rebuild SAP, NetSuite or Dynamics using prompts but instead to create differentiating capabilities around core enterprise systems.
Shadow AI is the real risk
If vibe coding is going to be widely adopted, governance becomes even more critical.
Ralhan warns that many organisations are already facing a new form of shadow IT, one driven by generative AI rather than unsanctioned cloud software.
“Shadow AI is here,” she says. “And AI is hard enough to govern as it is, let alone when an organisation isn’t aware of what’s happening.”
The concern is not just security but the risk that employees begin building ERP-connected applications outside approved controls, creating competing versions of the businesses truth.
“ERP is always considered the version of truth,” she says. “When we start overlaying employee-led applications, the version of truth may become diluted, which can have devastating ramifications.”
For that reason, she says organisations should establish boundaries early.
“It’s always better to have stronger boundaries and then loosen them up as opposed to going the other way around because once the cat’s out of the bag, it’s hard to get it back in.”
Flexibility and disposable apps
Rather than rejecting vibe coding, or embracing it blindly, Ralhan recommends organisations develop a more flexible ERP strategy, assessing build-buy-blend approaches and experimenting with AI-driven capabilities through tightly governed pilots.
“We need to have a more flexible understanding of what the ERP strategy is, and at least assess it. You might revert back to your mega vendor, and that’s absolutely fine, but also look at what the options are. Is it more of a build-buy-blend approach or is it about focusing on extending what you have at the moment through other means.”
Disposable applications – applications that work for a limited amount of time, much like an Excel spreadsheet used briefly to do just one thing and then deleted – are also an option. Rather than launching major development projects for every business requirement, companies could use AI-assisted development tools to quickly create targeted applications that solve a specific problem, automate a process or support a temporary business need. When that need disappears, the application can simply be retired. The approach allows organisations to experiment faster without cluttering their core ERP environment with customisations that become difficult and expensive to maintain. In this model, ERP remains the foundation, while AI-generated applications become lightweight extensions that can be created, used and discarded as needed.
“If you’re able to be flexible with the ERP strategy and you’re able to sell that to the board, your organisation is going to be in a really strong position.”
