Published on the 13/06/2024 | Written by Heather Wright

Compelling functionality propelling market…

Adair Durie admits that nine months ago he was a generative AI sceptic when it came to the benefits for ERP.

“If you had talked to me about AI nine months ago, I wouldn’t have rolled my eyes, but maybe internally I would have been,” the sales director for FujiFilm MicroChannel whose ERP line-up includes Microsoft and Sage, says.

“The advances we’ve seen in the past few months, are making it much more obvious why you need to go through the digital transformation process.”

“But now it is being delivered in ways where the benefits are understandable to business people, rather than it just being marketing, tech or developers getting excited.”

Vendors have been rushing to add generative AI to ERPs to provide customers with better insights and to help draft communications and reports and to generate insights.

A new report from Forrester, How Generative AI will Transform ERP, says the technology is starting to change ERP’s potential.

AI, of course, has been part of ERP for a long time, but generative AI use cases are just beginning to emerge.

Liz Herbert, Forrester VP and principal analyst and author of the report, says those use cases are real and available, but newer and in some cases still in beta mode.

“ERP customers are curious but real adoption remains early, with limited customer success stories at this time,” she told iStart.

And while there has been some AI washing, with vendors breathlessly touting their wares, Herbert says there is a significant amount of new, real AI already released ‘with a lot more to come in future’.

She says key use cases already seen include dunning/collections emails, narrative reporting and summarisation.

“We will likely see many more including more use of generative AI to do financial and operation planning and analysis, which is a relatively immature part of the ERP market.”

They’re all areas that have Durie’s customers excited, with the addition of ‘compelling functionality’ driving new demand for ERP sales.

He notes one of the earliest examples he saw last year, within Microsoft’s Business Central, which saw genAI automatically creating marketing text for an ecommerce website, based on different tones, using attributes from the Business Central database.

“That was a massive efficiency saver that was used by a number of our customers, predominantly in the B2B ecommerce space.”

The technology is also making in-roads in bank reconciliations. “When a customer pays multiple invoices, AI is able to say those three invoices match that one payment. So it’s generating lines on a bank reconciliation. And that scenario is live and happening.”

He says one offering, yet to be released in the Australian and New Zealand market, is a Microsoft Dynamics feature. It can already generate warnings and tell users about the potential impact on their business of changing delivery dates for raw materials, for example.

“But now, based on that information, you can ask the system what alternative actions you can take, and the system can do that.”

For Durie, that’s a key area where the power of GenAI will come to bear.

“Things like accuracy and confidence in numbers and confidence in decision making – I can see AI is going to be improving that immensely.”

Also just launched is an AI version of Zone & Co’s Zone Capture product – the highly used AP automation tool for NetSuite. The new version includes AI/machine learning capability, enabling it to handle foreign language invoice translation into NetSuite data fields, for example.

Forrester’s report cautions buyers to watch how functions evolve, saying some will go away or change drastically.

“While some of the interesting AI use cases focus on solving today’s ERP challenges, keep in mind that some of these challenges may not exist at all in the future. For example, the consolidation and closing process is changing. The assurance process is changing. Payments are becoming ever more electronic, thus reducing the need for capabilities like OCR and invoice matching in the long term.

“Make sure you aren’t using AI to blindly address processes without examining whether those processes will remain necessary in an AI-first world.”

Durie says it’s still too early to put any figures on genAI in ERP’s true value.

“We’re probably a few months too early still.”

At the moment, he notes, it’s not life changing – yet – and instead highlights that it is a journey that we’re on.

What he can say, however, is that people are now buying because of the technology.

“It’s definitely resonated with the market,” he says.

“Typically, we see large numbers of business owners, c-levels who say ‘why do I need an new ERP system, my old one still works’. And they’re right. Even some of the really old ones are still solid.

“But with the advances we’ve seen in the past few months, even those who [traditionally were reticent to upgrade] are seeing that their competitors now have access to these technologies and insights that their system can’t do. It is making it much more obvious why you need to go through the digital transformation process.”

Herbert warns that companies need to have the basics in place in order to reap the benefits of genAI in ERP.

That includes investing in the ERP stables of data quality and change management.

“GenAI is not going to magically solve your ERP rollout challenges,” Herbert says. “It is not going to magically fix adoption or data quality issues, especially in the near term.”

It does, however, put a significant burden on data quality from the outset, creating additional pressure to have the right data strategy in play.

Durie agrees, noting that going to any new system requires data cleansing.

“Any digital transformation process on ERP always includes data cleansing. You’ve got your customers, suppliers, inventory, bills of materials – all that level – but it’s also how you view your business, cost and profit centres, posting groups…”

Herbert says that companies also need to watch the costs involved with genAI and know that they are still rising.

“Costs and pricing are still evolving and have historically been a barrier to AI success, which we saw with IBM Watson.

She notes that for the most part, costs remain unclear. While some AI is becoming part of the ERP costs customers already pay, others are starting to be monetised separately, either through new modules or through overt packaging and pricing changes.

She notes that sophisticated AI tends to be expensive – and those costs have to be paid somewhere.

“You will likely incur additional costs for service providers, partner add-ons and training.”

She’s urging businesses to experiment and learn.

“It’s important to get started with experimenting and laying the groundwork,” she says.

“Focus on data hygiene, governance and policy – all of which are prerequisites. And if core systems are old, modernising them will best position you for AI success in future.”

And when it comes to all that marketing buzz, Herbert says make sure you press suppliers for real examples and ROI.