Dominic Jainy is a seasoned IT professional who has spent much of his career navigating the complex intersection of artificial intelligence, machine learning, and blockchain. With a particular interest in how these transformative technologies can be woven into the fabric of traditional business operations, he offers a grounded perspective on the current AI gold rush within the enterprise resource planning (ERP) sector. In this discussion, we explore the critical necessity of a solid commerce foundation, moving beyond the superficial allure of software demonstrations to focus on data governance, operational dependability, and the strategic alignment required to make AI a truly productive asset rather than a source of operational friction.
The conversation covers the recurring pitfalls of implementing high-level AI features on top of shaky data architectures, the growing financial pressure on executives to modernize, and the vital distinction between simple system integration and true ERP governance. We also delve into the evolving role of Microsoft Copilot and how businesses can prepare their internal structures to ensure that intelligent automation supports—rather than undermines—their core business rules.
Many companies are currently dazzled by high-tech demonstrations of AI-powered search and automated content, but what is the danger of prioritizing these visible features over the underlying commerce architecture?
The primary danger is that these demonstrations act like a high-end coat of paint on a house with a crumbling foundation. We see manufacturers and distributors get incredibly excited when they see an AI-driven search bar or a tool that automatically generates product descriptions, but they often fail to ask if those tools are pulling from a source of truth. In a Business Central environment, the commerce experience isn’t just a simple catalog; it is a complex web of negotiated pricing, inventory commitments, and credit limits. If the AI is built on a separate, replicated platform rather than the ERP itself, it may confidently offer a customer a price that was overridden two hours ago or suggest a product that is currently restricted for their specific account. You end up with a very impressive, modern interface that is effectively lying to your customers because it lacks access to the governed business rules that reside in the ERP.
Given that 68% of finance leaders expect their digital transformation expenses to increase this year, why does higher spending not always translate into better AI outcomes for eCommerce?
It really comes down to where that capital is being deployed. The Grant Thornton 2026 CFO survey shows that while the appetite for spending is at an all-time high, the actual success of these investments is wildly uneven. Gartner has observed that the organizations seeing real returns are those investing up to four times more into foundational areas—like data quality, governance, and change management—compared to those who are struggling. Many companies make the mistake of thinking that AI is a “set it and forget it” feature you can simply buy off the shelf. In reality, AI is an amplifier; if you have messy data and inconsistent pricing rules, AI will simply help you scale those errors faster than ever before. Executives need to stop viewing AI as a standalone feature and start seeing it as a component of their complete commerce operation that requires a rigorous, data-first strategy.
You often emphasize the difference between a system being “connected” to Business Central and being “governed” by it. Why is this distinction so pivotal for AI readiness?
This is the central architectural question that determines whether your AI will be a help or a headache. When a system is merely “connected,” it usually means data is being synchronized or copied from Business Central into a secondary platform, creating a “shadow” version of your business logic. If your pricing is copied and maintained in a separate eCommerce tool, you’ve essentially created two versions of the truth that must be constantly reconciled. Governance, on the other hand, means that Business Central remains the brain that makes the final decision on every transaction in real-time. When AI begins to initiate actions or give advice to customers, it must do so using the governed logic of the ERP. If the AI uses a copy of the data, it might miss an updated credit limit or a newly applied order-validation rule, leading to a “failed” order that an employee then has to manually fix in the back office.
How is the introduction of Microsoft Copilot and autonomous agents within Business Central changing the way companies should evaluate their eCommerce solutions today?
Copilot represents a shift away from the idea that all AI value must come from your eCommerce platform provider. Microsoft is building an ecosystem where AI assistants and agents are baked directly into the Business Central environment to spark creativity and eliminate tedious tasks. This means that a forward-thinking company shouldn’t lock themselves into a commerce solution that tries to own all the AI logic. Instead, they should look for an architecture that preserves their flexibility to use Microsoft’s broader Copilot tools alongside their commerce-specific AI. The real question is whether your eCommerce solution will help you leverage these new tools without weakening your central governance. If your commerce solution pulls critical logic away from the ERP, it makes it much harder for a Copilot agent inside Business Central to accurately assist your team or your customers.
What are some of the practical, real-world consequences that occur when an AI tool attempts to function using stale or replicated commerce data?
The consequences usually manifest as “faster exceptions”—problems that happen so quickly your team can’t keep up with the corrections. For example, an AI recommendation tool might suggest a high-margin product to a customer, but because it doesn’t recognize an account-specific restriction living in the ERP, the customer tries to buy something they aren’t allowed to have. Or perhaps an AI assistant confirms that an order can ship today based on an inventory sync from an hour ago, but it fails to see a massive bulk order that just hit the ERP five minutes later. We also see AI-generated quotes that use copied pricing which no longer reflects a customer’s current agreement. In all these cases, the customer feels misled, and your internal staff is forced to spend hours investigating and reconciling the discrepancy between what the AI promised and what the ERP allows.
When evaluating a potential eCommerce provider, what kind of specific scenarios should a company use during a demonstration to test the platform’s true AI reliability?
You have to get past the generic retail scenarios and push the provider into the “weeds” of your specific Business Central setup. I always recommend asking a provider to use a realistic customer profile that has negotiated pricing, multiple ship-to locations, and strict credit controls. Ask them to show you exactly where each rule is evaluated—is it happening in the ERP, or is it a copy in their platform? Then, ask what happens to the AI’s suggestions the very moment you change a price or a credit limit within Business Central. If the AI doesn’t reflect that change immediately, you know you are looking at a system that relies on stale data. This helps the evaluation team distinguish between an attractive user interface and a dependable commerce process that can actually be trusted to handle complex B2B transactions.
Beyond the technology itself, what organizational responsibilities must be addressed to ensure a company is actually ready to adopt AI in their commerce operations?
Technology is only half the battle; the other half is internal clarity and ownership. AI cannot fix a business process that is fundamentally broken or undocumented. If your pricing rules are inconsistent or your product records are incomplete, the AI will only highlight those failures. Business leaders must identify who truly owns the commerce information and how changes are approved before they even think about turning on an AI agent. This involves cleaning up inventory practices and ensuring customer classifications are accurate and up to date. Preparing your data and your people in this way isn’t a delay to your AI adoption; it is the essential groundwork that makes responsible, scalable adoption possible in the first place.
What is your forecast for AI-enabled eCommerce?
I believe we are moving toward a “Headless Governance” model where the interface becomes increasingly invisible and the logic becomes increasingly central. In the next few years, the most valuable AI features won’t be the ones we see in flashy search bars, but the autonomous agents working behind the scenes to reconcile thousands of data points across the supply chain in real-time. We will see a shift where the ERP isn’t just a database, but a dynamic engine that powers AI interactions across every touchpoint—voice, chat, and automated procurement systems. However, this future belongs only to the companies that have the discipline to prioritize their data foundation today. Those who ignore governance in favor of quick-fix AI features will find themselves trapped in a cycle of constant architectural rebuilding as the technology continues to outpace their shaky infrastructure.