Vibhu Kapoor, Regional VP for Middle East, Africa & India at Epicor, has penned an op-ed, in which he explains that AI without ERP is a huge gamble for enterprises, and that why pairing them is the smartest choice. 
The ERP market is expected to top US$150 million in the United Arab Emirates this year, and US$111 million in Saudi Arabia. CAGRs of 4.24% for the UAE and 2.05% for Saudi Arabia paint a picture of GCC CXOs who consider ERP as, if not the beating heart of business, then certainly a vital organ. Now more than ever, these same CXOs are under pressure to implement impactful AI use cases.
The good news is that they need not make a hand-wringing choice between ERP and AI because ERP itself is fast becoming a critical component in the AI stack.
With data defining AI success, executives must answer key data-related questions. Where will it come from? How will it be shaped? Who will be allowed access? If the data is not clean, the models will not give the desired results. And if it is mishandled, regulators will be at the door. ERP was taking care of many of these concerns when AI was still considered sci-fi gimmickry.
Beneath the lake
ERP is built on a repository of the enterprise’s cleanest data. AI is famously reliant on data, and models are incredibly sensitive to the quality of that data. Data management and governance are built into ERP. Modern ERP platforms can gracefully integrate AI, giving it instant access to the standard of data it needs to glean accurate and actionable insights.
Normally concerns such as cybersecurity would accompany AI’s consumption of data. These are greatly diminished because the fundamentals have not been compromised — the ERP system already has the means to implement sound governance and cloud offerings even have robust cybersecurity measures in place.
One way ERP platforms implement an accessible-yet-secure ecosystem is through the data lake — a centralized store that can be either raw or clean, structured or semi-structured. Data lakes are made for AI, but they also come with a range of tools that allow organizations to define how and by whom data is accessed and what shape the data is in when it reaches the user. AI modelling algorithms can thereby be guaranteed only to work on data that is clean and appropriate. With a data lake, all the boxes on the AI wish-list are ticked, the main ones being:
- Quick to deploy
Once a data lake is in place, the organization will have enabled rapid consolidation of information that was previously the guarded jewels of different business units. These data silos would otherwise stand in the way of enterprise-wide collaboration on AI. Traditional data-warehousing can be a months-long project. The data lake approach takes weeks and gives way to a speedier time to value for the AI tools that will trawl its depths. This time to value emerges from an underlying competitiveness that is sustainable because of the organization’s attentiveness to the fundamentals of governance, security, and accessibility.
- Scalable
The start of the year is a time to think of growth. For regional businesses with ambitions of international forays, the scalability of its technology infrastructure is critical. One of the great attractions of the data lake is that it offers a low-cost storage solution. High-quality data cannot be of high value if does not come in large volumes. Storage solutions therefore must be capable of growing cheaply.
- All the tools you need
As soon as all its tributaries have flowed in, the data lake is ready to be tapped by thirsty machine-learning algorithms. The variety of the pool’s data makes it ideal for predictive analytics and other flavors of AI that thrive on access to many different types of information. Many of the best decisions an enterprise makes will come from the insights dredged from its diverse data lake.
- Secure and compliant
The data lake stands among the best approaches for data security and governance, granting complete control to the organization over everything it stores and how and when to make it available.
Regional companies must comply with many different regulations (such as the need for data residency) established by national agencies, but also with global regulations such as the EU’s General Data Protection Regulation (GDPR), and with industry regulations like the Payment Card Industry’s Data Security Standard (PCI DSS). Through the data lake, businesses can ensure trust in their brand continues throughout their onboarding of AI.
No moonshots
A private data lake gives an enterprise control over its future, which is why data lakes are the future of ERP and of AI. In short, the fastest route to AI value is through an ERP-resident AI suite that feeds on a well-designed data lake. Such a setup not only provides the speed, efficiency, and security needed to reap ongoing value from AI; it also provides the flexibility to integrate with other systems.
When AI has made its home in an ERP suite fed by a data lake, everything changes. The organization will have delivered more visibility to decision makers. It will have enabled unprecedented collaboration. It will have boosted efficiency and bolstered its supply chain. The ERP-resident approach to AI adoption is a departure from the risky, moonshot missions of those who expect overnight success. AI is not a cure-all; nor is it a sole serum for competitiveness. It is a tool that, if used strategically in cooperation with the right partners, can transform the business over time into a market leader.
