Artificial intelligence is racing into the enterprise core, but the winners won’t simply be the fastest. They’ll be the ones that earn trust. That’s the argument from Soma Somasundaram, President & CTO at Infor, who believes the defining feature of the next wave of enterprise AI won’t be raw automation; it will be transparency that people can interrogate, audit, and ultimately rely on.At the recently concluded Infor Velocity 2025 summit, that thesis was on full display. The conversations were less about flashy demos and more about dependable outcomes: agentic systems collaborating with humans, decisions that can be traced and explained, and architectures that make modernization feasible for companies that don’t have infinite budgets or patience.From hype to habit: trust is the adoption curve“AI today is where the cloud was a decade ago, everyone wants it, not everyone knows how to use it,” Somasundaram says. The gap he points to isn’t a lack of models; it’s a lack of confidence. If a recommendation engine can’t show its work,what data it used, which signals mattered, how risks were weighed—then automation stalls at the edge of critical workflows.His framing is pragmatic rather than purist. “Transparency is the bridge between experimentation and trust,” he notes. In practical terms, that means giving a procurement manager the why alongside the what: not just “pick supplier B,” but the performance history, delivery adherence, pricing, and compliance factors that led to the verdict. The Infor agents are role-based and micro-vertical, built on Infor’s Industry Process Catalogs and Value Maps, and draw on Infor and non-Infor data so buyers, planners, and finance users can approve, override, or escalate, with the why attached. “This isn’t about controlling AI,” he adds. “It’s about making it trustworthy enough to delegate. Legacy gravity and the modernization trapIndia’s adoption story carries a structural twist. The country has world-class engineering talent—and a vast mid-market still running on aging ERP cores. For decades, bolt-ons and bespoke integrations masked the problem. But AI magnifies technical debt. Fragmented data and brittle interfaces quickly turn “intelligent” automation into sandcastles. Somasundaram’s counsel is blunt: treat core modernization as a business necessity, not an IT indulgence. Stabilize what exists, carve out high-value processes with clear success metrics, and then move those to cloud-native, multi-tenant architectures where security, upgrades, and extensibility are baked in. The goal isn’t a grand rewrite; it’s compounding momentum from a sequence of visible wins. Delay, he warns, simply transfers risk to the next breach or ransomware incident.India’s leapfrog moment - if we design for itThe paradox of India’s market is familiar: large enterprises can absorb multi-year hybrid journeys; the growth engine, small and midsize firms cannot. They don’t want a best-of-breed jigsaw; they want plug-and-play intelligence that “just works,” respects regulations, and doesn’t blow up the budget. Cultural economics matters too: where labor is cheaper, automation faces skepticism. But, Somasundaram argues, the long-term gains—fewer errors, shorter cycle times, cleaner data—are decisive. “India’s greatest opportunity is to skip stages. Just as it jumped from landlines to mobile, it can leap from legacy to explainable, intelligent operations.”That’s our thesis. For India, Infor is launching an AWS region in Mumbai for local data residency. We already have hundreds of Indian customers running in Frankfurt; as regulations evolve, local hosting will matter more. If AI recommends a supplier, it should tell you why—delivery history, quality, price dispersion, compliance flags,” he notes. He add, "Strategically, in the industries we serve, we want the natural question to be, “Why not Infor?” The talent is here, the use cases are here, and the leapfrog potential is real—just as India skipped landlines for mobile, enterprises can skip outdated architectures and adopt cloud-native, multi-tenant systems built for the next decade". Innovation with a seatbelt: compliance as a continuous verbRegulation is no longer a side quest. From Europe’s AI Act to India’s DPDP law, the governance surface area keeps widening. The trap, Somasundaram says, is to make compliance with the opening move. “If you front-load compliance and sideline innovation, you stall.” His playbook: innovate first, then validate in structured rollouts with early adopters where every step is logged for security and explainability. He points to disciplined adherence to OWASP and ISO standards, isolation of customer data by design, and “human-in-the-loop” as a non-negotiable. The message to boards: trust in AI isn’t a declaration, it’s instrumentation.Sovereignty is part of that trust calculus. Cloud partnerships that support national data-residency needs, while preserving modern tooling are maturing. The north star is simple: give regulated industries credible paths to keep sensitive data in-country without sacrificing the benefits of contemporary AI infrastructure.What good looks like: process intelligence that pays for itselfIf trust is the adoption curve, process intelligence is where the payoff lands. The most compelling scenarios at Velocity didn’t fetishize models; they celebrated measurable outcomes. Agent-assisted procurement that shrinks cycle times and surfaces supplier risk early. Process mining that auto-summarizes bottlenecks instead of burying teams under event logs. Finance and shop-floor users who get recommendations they can approve, reject, or escalate with every decision auditable.Somasundaram’s litmus test is refreshingly enterprise-native: does the system reduce rework, prevent leakage, and cut time-to-cash? Can a domain user see why it acts and decide whether to act on it? When the answers are yes, trust compounds, and adoption becomes a habit, not a pilot.There’s a welcome sobriety to this moment. After a year of generative AI exuberance, leaders are demanding systems that can stand up in front of auditors, not just on stage. The frontier is shifting from “Can we automate it?” to “Can we explain it, and still ship on Friday?” That bar is higher. It’s also healthier.And while no vendor has a monopoly on wisdom, Somasundaram’s through-line is a useful compass: design for explainability, modernize the core, and aim AI at the processes that move the P&L. If enterprises do that, speed shows up anyway, because trusted systems get used.
Published On Oct 24, 2025 at 09:30 AM IST
Join the community of 2M+ industry professionals.
Subscribe to Newsletter to get latest insights & analysis in your inbox.
Rewrite this article in English only.
- Maintain authoritative, neutral journalistic tone
- Use professional editorial news style with proper attribution
- Preserve key facts, quotes, and statistical data
- Ensure proper paragraph structure and flow
- Do not add editorial commentary or phrases like “Here’s the rewritten article”
- Output only the rewritten content
- Maintain similar word count (±10%)
- Include relevant context for better reader understanding
