AI as an Enabler of Real Retail Outcomes: Nextuple Order Management Gurus Webinar
As part of the Partner Power Up series within Nextuple’s community-focused group, Order Management Gurus, our VP of Strategy and Chief Evangelist, Karthik Marudur, sat down with Laxman Mandayam, Head of Customer Solutions & Co-Founder, Nextuple, to discuss the future direction of AI within the context of retail and Order Management System (OMS).
The interview opens with Laxman challenging Karthik on the notion of ‘AI first,’ and what this means to OneStock.
What Do We Mean by ‘AI First’?
As Karthik explains, AI is moving fast in retail, with conversations rapidly shifting from hype to outcomes, as he starts by pointing out that the mantra of ‘AI first’ only matters if it delivers measurable operational impact, and if it’s grounded in trustworthy commerce truth.
He continues, “At OneStock, ‘AI first’ is described less as a marketing label and more as an operating model. Internally, the company is already using agents to remove friction from everyday work.”
As an example, Karthik pointed to tools like “OneBot,” which helps teams respond to RFPs, build assets faster, and resolve product questions more efficiently. “That internal discipline,” he explains, “directly shapes how OneStock designs customer-facing agent experiences: start with real problems, prove ROI, then scale.”
National Retail Federation 2026
Recapping OneStock’s recent participation in the event, Karthik points out, “That pragmatism was also a major takeaway from NRF 2026. The team saw a noticeable change in tone across the event floor: retailers and vendors are talking far more about implementable use cases than theoretical possibilities.
Shopping agents, in particular, are gathering momentum, though the broader ecosystem is still early, with many platforms not yet open enough to support seamless integrations at scale,” he adds.
The Practicalities of Agentic Commerce
The interview turns at this point to agentic commerce and putting it in a very practical frame. Agentic commerce, as defined in the webinar, is crucially about AI agents guiding customers through both sides of the customer journey: pre-purchase discovery and decision-making, and post-purchase service.
Expanding on this, Karthik continues, “Before purchase, the agent experience becomes a conversational shopping assistant that doesn’t just recommend products, it recommends executable options, based on the customer’s preferences and real-world constraints like current availability and delivery windows.
After purchase, the same approach turns customer service into a simpler, more unified experience. Instead of forcing shoppers to piece together status across multiple screens, an agent can summarize the order lifecycle, highlight line-item changes, and provide tracking updates in one coherent interaction.”
Agents’ Dependence on Operational Truth
But one of the most important themes in the conversation was a discussion on what dependency agents have on operational truth. Availability, delivery promise, and order lifecycle data aren’t “nice to haves,” rather they determine whether an agent is genuinely useful or dangerously confident. If inventory is stale, promise dates are inaccurate, or order events lag behind reality, the agent will still answer in a definitive tone. The result, Karthik says, is predictable: missed SLAs, failed pickups, increased cancellations, and customers who lose trust quickly because the system sounded certain and turned out to be wrong.
To solve that, OneStock is cementing its role as the provider of trusted, real-time commerce signals that agents can reliably consume. In the interview, Karthik describes how OneStock is enabling both retailer-owned agents (embedded into brand experiences) and external agents (such as ChatGPT) by exposing operationally reliable data through MCP tooling. The idea, he explains, is to make trusted OMS capabilities (inventory availability, fulfillment status, delivery options) ‘query-able’ in a standardized way, so agents can interact with the truth of the order lifecycle rather than loosely inferred estimates.
Operational Efficiency and Return on Investment
The fastest return on investment, however, isn’t necessarily in futuristic shopping journeys; it’s in operational efficiency. The interview highlighted how AI’s most immediate impact often comes from reducing workload and latency for customer service and store teams. When agents can quickly retrieve order context, explain exceptions, and surface the next best action, they compress resolution cycles and reduce the costly ‘swivel-chair’ work that happens across disconnected systems.
Karthik also points here to OneStock’s differentiated fulfillment capability: “competitive allocation.” Instead of assigning an eligible order to a single store and waiting, the model sends the order to multiple eligible stores and allows the fastest to claim it – while applying guardrails to reclaim orders that stall and deprioritize stores that repeatedly decline. In the example referenced, Karthik cites an average claim time of just 13 minutes where competitive allocation comes into play, coupled with a meaningful reduction in cancellations, firmly positioning speed-to-commit as a lever for improving delivery outcomes.
From Individual Agents to Connected Agents
Looking ahead, the interview closes on Karthik’s forward-looking view of where this all goes next: not just individual agents, but connected agents, collaborating behind the scenes across systems. “That’s the direction OneStock is leaning into; tools and standards that let multiple specialized agents work together across the commerce lifecycle, while still being anchored to the same core requirement: operational truth.”
Because in agentic commerce, connection alone isn’t enough. If an agent can’t access reliable availability, promise, and lifecycle status in real time, it can’t be trusted to act – and trust is the currency that ultimately determines whether these experiences scale, he concludes.
