As artificial intelligence begins to reshape digital commerce, the role of ERP and inventory systems is undergoing a fundamental shift. The rise of agentic commerce—where AI-driven agents assist or even automate purchasing decisions—means that enterprise systems must provide accurate, real-time operational data not only for internal teams but also for machine-driven marketplaces and digital platforms.
For many organizations, this shift places new pressure on inventory infrastructure. Systems originally designed for periodic updates and human workflows must now support real-time visibility, machine-readable product data, and reliable fulfillment signals across multiple channels. In this environment, inventory accuracy is no longer just a back-office operational concern—it has become a frontline driver of customer experience, brand trust, and revenue performance.
In this written Q&A with ERP News, Riikka Söderlund shares her perspective on how agentic commerce is reshaping ERP and inventory management strategies. She discusses the growing importance of real-time data visibility, the architectural role of cloud infrastructure and composable systems, and why ERP leaders must rethink inventory systems as part of a broader digital commerce infrastructure. The conversation also explores how organizations can modernize inventory intelligence without replacing their entire ERP stack, and what capabilities will define successful ERP platforms in the next phase of AI-driven commerce.
Q: Agentic commerce is gaining momentum as digital demand and physical fulfillment converge. From Katana’s perspective, what fundamentally changes for ERP and inventory systems in this new model?
A: Inventory management is no longer just an internal tracking tool or function; this new model is shifting toward leveraging inventory data as external decision infrastructure. AI agents are increasingly taking over the consumer journey, replacing default website browsing and customer experiences by taking the user from discovery to checkout in the same platform. Systems must provide real-time, machine-visible product data to remain discoverable within these agentic shopping experiences. This shift is stress-testing inventory systems. ERPs need to evolve their approach and guarantee accurate product availability and fulfillment data to keep up with agentic queries and transactions.
Q: Many organizations still treat inventory accuracy as a back-office concern. Why is it becoming a front-line performance metric for revenue, customer experience, and brand trust?
A: Inventory accuracy now determines brand visibility. Inventory accuracy is now directly affecting revenue, discoverability and brand trustworthiness. From a revenue standpoint, inventory opacity ties up working capital in excess safety stock, creates stockouts that turn away legitimate orders and damage marketplace rankings. For most small and mid-sized brands, this means someone is spending up to 20 hours per week on operational firefighting instead of growth. On the customer experience side, inaccurate available-to-promise data leads to overselling, post-purchase cancellations and broken promotions, eroding loyalty and forcing margin-eating make-goods. Most critically, in the era of AI shopping agents, if transactions fail, agents route buyers elsewhere. Companies are realizing they are already paying for poor visibility in lost sales and trust, and that investing in a real-time system of record is now essential to stay competitive.
Q: AI tools are only as effective as the data they rely on. What core data layers must be in place for agentic and AI-driven inventory decisions to actually work?
A: Organizations need a few core foundations: real-time visibility across all locations, consistent product and order IDs across systems, and clear records of every inventory movement. Having a data stack anchored in a true system of record is crucial to maintaining machine-speed as AI drives consumer behavior and queries inventory. On top of that foundation, businesses also need event-driven, real-time aggregation so every system updates instantly rather than in batches. Complete audit trails and state history are also essential to ensure trust, traceability and reliable decision-making at machine speed. AI systems can trust the data they access, which is essential because agents won’t compensate for uncertainty the way humans might.
Q: How can manufacturers and SMBs modernize inventory visibility without replacing their entire ERP stack or disrupting daily operations?
A: The most practical approach is to implement a dedicated inventory truth layer that focuses on real-time data and composability rather than all-in-one control. Instead of replacing an entire ERP, brands can adopt a system designed to act as a foundation for real-time visibility, providing a single source of record by aggregating data from various channels and locations without transforming it into stale numbers. Modernization also means shifting away from legacy batch processes to event-driven architecture, ensuring available-to-promise counts are accurate across every sales channel simultaneously. By leveraging emerging commerce protocols like UCP and ACP, prioritizing API-first microservices and adopting a composable approach, businesses can connect their entire ecosystem without degrading system performance and swap out specific parts of their tech stack as they grow. The most effective way to avoid disruption is to implement these systems before they lead to load-bearing chaos.
Q: In your experience, where do companies most often struggle when trying to operationalize real-time inventory data across sales, operations, and finance?
A: Most companies struggle to operationalize real-time inventory because they’re still relying on manual processes, fragmented data or legacy systems that were never built for the speed of modern commerce. Even fast-growing brands often run on spreadsheets updated once a day, creating what amounts to “inventory opacity.” Teams spend more time reconciling discrepancies than driving growth and often carry excess safety stock because they don’t trust their own numbers. Technically, many systems still rely on batch updates instead of event-driven architecture, cap integrations making it nearly impossible to maintain a true single source of truth. As a result, companies struggle to calculate accurate available-to-promise inventory across channels, leading to overselling, missed revenue or stalled sales conversations. Now, with AI shopping agents querying inventory in real time, the bar is even higher. Brands need accurate inventory infrastructure or risk being deprioritized altogether.
Q: Tariffs, labor shortages, and supply volatility continue to pressure margins. How does improved inventory intelligence help organizations remain resilient in this environment?
A: Improved inventory intelligence provides a foundation for operational resilience by transforming inventory from a financial blind spot into a real-time asset. This allows organizations to reduce safety stock, avoid costly rush shipments and fulfill orders confidently during uncertainty. Real-time visibility also helps organizations remain resilient against supply volatility by providing a high-trust available-to-promise calculation. When a company lacks trust in their data, they build up safety stock buffers, which ties up working capital that could be put toward other things. Companies with real-time visibility know exactly what they have in stock and can meet consumer demand accordingly, reducing the need for conservative buffers and positioning themselves competitively on the market.
Q: How should ERP leaders think about the balance between automation and human control as agentic workflows begin to execute decisions autonomously?
A: Automated workflows are best suited for handling data with speed and scale, while humans should remain in charge of policy, rules, exception handling and strategy. More agent customers means ERP leaders need to be equipped with agent workforce to match the speed of AI shopping. Traditional ERP systems were built on the assumption that humans are forgiving and can tolerate 15–30-minute sync delays. However, AI agents operate in milliseconds. The goal isn’t perfection. Strategically automating just 90% of the operational burden, leaving only the exceptional cases manual, can make an impact and allow teams to focus on areas of growth instead of constant inventory monitoring. A winning balance keeps autonomy governed and maintained.
Q: What role does cloud architecture play in enabling real-time visibility and scalable agentic workflows compared to legacy on-premise systems?
A: While legacy systems were built for human-centric timelines, modern cloud infrastructure is designed for the high-frequency, high-trust demands of agentic commerce. It gives ERPs real-time data and the API-first connectivity needed for many agentic and commerce protocols, which on-premise systems might struggle to support. Many older systems are monolithic, meaning their performance degrades as more connections are added. Embracing an approach designed on microservices means systems can handle unlimited integrations without performance loss. Cloud infrastructure provides the foundation necessary to remain discoverable and reliable in an environment where machines, not just humans, are making purchasing decisions.
Q: Looking ahead to the next 2–3 years, how do you see ERP priorities shifting as agentic commerce moves from experimentation to execution?
A: I’m watching brands to see who is building for agent-driven commerce today and who’s retrofitting human-driven systems with AI tools. I think we’ll see a massive consolidation as architectural debt becomes unavoidable. Systems that rely on those monolithic systems will hit integration caps and struggle to achieve the data connectivity needed for modern inventory management. ERP priorities will likely shift from feature lists to architectural readiness. Leaders will begin embedding AI and automation governance into workflows so systems can support real-time data access, unlimited integrations and composable infrastructure. In the agentic commerce era, I also see reliability becoming the new SEO. ERPs will be judge on their ability to maintain a brand’s visibility within AI platforms.
Q: For organizations just starting this journey, what is the first practical step they should take to prepare their ERP and inventory infrastructure for agentic commerce?
A: The most practical first step for businesses is to identify one operational source for inventory that is cloud-based and API-accessible. Audit a sample of SKUs across locations and compare to see if system orders and stock movements line up. Track the hours spent on manual reconciliation and lost sales caused by conservative stock buffers. In many cases, you will find you are already spending more on inefficiency than you would on a system designed to fix it. If accuracy is inconsistent or the answer isn’t immediate, that’s the signal for organizations to prioritize real-time visibility. That single, reliable inventory layer is the foundation that everything in agentic commerce relies on.
ERP News Editorial Team
The ERPNews Editorial Team covers global developments in ERP (Enterprise Resource Planning), enterprise software, cloud platforms, AI, automation, and digital transformation, providing independent news and editorial analysis for senior business and technology leaders. Our reporting focuses on market signals, strategic shifts, and enterprise impact across the ERP and enterprise technology ecosystem.
For editorial inquiries, please contact:
📩 [email protected]
