“Once you’ve watched a billion‑dollar supply chain stall because someone missed a line in a spreadsheet, you stop thinking of AI as a buzzword and start seeing it as basic infrastructure,” said Kunal Barman, co‑founder and chief executive of Faction. “Operational AI is about making sure those bottlenecks don’t decide who wins and who loses in the industrial economy.”
From the outside, America’s manufacturing heartland looks newly ascendant. Factories are being reshored, megaprojects in energy and logistics are drawing record investment, and industrial distribution remains the unseen circulatory system that feeds it all. Inside those businesses, however, work often still runs on manual quotes, laggy ERPs, and unstructured email threads. It is in this gap that Barman has deliberately positioned his young company and, by extension, his own career.
Barman arrived in industrial AI through the capital stack that funds the physical economy. At BlackRock’s Financial Markets Advisory group, he worked on the creation of Saudi Arabia’s $53 billion National Infrastructure Fund, helping design investment frameworks across energy, logistics, and industrial development. Immersed in models of stadiums, power plants, and ports, he saw how modest operational frictions at the edge of the system could erode returns on a large scale.
That vantage point gave him a habit of tracing back macro outcomes to micro workflows. When he later evaluated AI companies as the youngest investor at QuantumLight, a $300 million venture fund founded by Revolut's Nik Storonsky, he brought that bias with him: a focus on tools that change how work gets done, not just how dashboards look.
His academic path reinforced the same pattern of thought. Studying Philosophy, Politics, and Economics at the University of Oxford, Barman developed an appetite for systems thinking: how rules, incentives, and information flows shape behavior. Academic prizes and scholarships recognized his performance, but it was the habit of moving between theory and practice that ultimately mattered more. It now shows up when he toggles seamlessly from a discussion about national productivity to the granular details of a distributor’s branch counter.
Faction emerged from Barman’s conviction that the physical economy’s limiting factor was no longer capital or even hardware, but the software layer coordinating everyday decisions across manufacturers and distributors. As co‑founder and CEO of Faction, he framed the company from day one as an “operational AI” platform: agents that automate work across sales, procurement, pricing, and finance, tuned specifically to industrial workflows.
Faction’s core products reflect that focus. One module ingests quotes and orders directly from emails, PDFs, spreadsheets, and text messages, extracting line items and metadata and writing them into the customer’s ERP with minimal human intervention. Another powers AI voice agents for inbound and outbound calls; answering questions about order status, capturing new orders, or handling routine collections conversations. A third cluster focuses on procurement and sourcing, including automating RFQs and conducting price-and-availability checks across multiple suppliers.
What ties these modules together is not just technology but a distinctive operating philosophy. Barman insists that Faction’s agents sit on top of existing ERP and CRM systems, respecting the decades of sunk costs and customization embedded within them. “We don’t tell a 40‑year‑old distributor to tear out the backbone that runs their business,” he said. “We connect to it, learn how their people use it, and then start taking work off their hands, piece by piece.” That incrementalism has become a hallmark of Faction’s approach.
The impact Faction markets to customers is expressed in numbers rather than abstractions. Across early deployments, Barman achieves a 75% or greater reduction in manual data entry by automating the movement from unstructured documents into structured ERP records. For front‑office teams, that means fewer hours spent retyping part numbers and more time spent on exceptions, complex deals, and customer relationships.
A second headline metric is more than 10% additional revenue per sales representative, driven by agents who pre-assemble quotes, surface cross-sell options, and ensure follow-ups are executed consistently. Beneath that figure lies a more subtle shift: the inside sales role becomes less about mechanical throughput and more about judgment and persuasion.
These outcomes are underpinned by proprietary data models that Barman pushed the company to build. Rather than relying on generic language models, Faction enriches each customer’s product catalog and business logic, enabling agents to reason about alternatives, margins, and inventory positions in a manner that aligns with the business's actual operations. In practice, this allows an agent to suggest a viable substitute when a part is out of stock or flag where a discount request would erode the margin below an agreed-upon threshold.
Inside Faction, Barman’s remit runs far beyond external storytelling. He leads enterprise sales cycles, often personally anchoring conversations with national distributors in sectors such as PVF, electrical, HVAC, packaging, and safety. Those meetings are less pitch than diagnosis: he maps where the quote-to-cash process slows down, where pricing discipline breaks down, and where teams are overwhelmed by manual work. The result is a shared picture of bottlenecks that operational AI can potentially alleviate.
At the same time, he directs product vision and market positioning. That dual role means decisions about which features to build or verticals to prioritize are informed directly by customer conversations rather than abstract roadmaps. It also allows him to maintain a coherent narrative across stakeholders: users hear the same rationale for a feature as investors do for a strategy, and engineers understand the field realities they are coding toward.
Equally central is his focus on building what he calls a “talent‑dense” organization. Barman is directly involved in hiring, culture‑setting, and leadership development, aiming to blend deep machine‑learning expertise with hard‑won industrial know‑how. This means recruiting individuals who are equally comfortable in a branch office in Ohio as in a demo environment, and rewarding those who take the time to observe how real orders flow through a customer’s system.
Barman’s credibility with industrial customers rests in part on earlier, more prosaic experience: working in the purchasing team of a medical distributor in the U.K. There, he saw how much of the operation depended on individuals’ tacit knowledge, and how much of their day was consumed by tasks that did not require that knowledge at all. “You had people spending hours chasing back‑orders and reconciling mismatched spreadsheets,” he recalled. “It was obvious that software should be doing more of that.”
That memory shapes how he frames Faction’s role. To purchasing managers and branch staff, he talks less about “agents” and more about ensuring that routine work just happens. To executives, he emphasizes that this is not about cutting heads so much as redeploying scarce talent to growth initiatives, strategic accounts, and complex problem‑solving. In both cases, the message is that operational AI should make frontline work less brittle and more meaningful, rather than turning people into supervisors of opaque systems.
This human‑centric framing also extends to how Faction rolls out its technology. Under Barman’s direction, implementations start with agents operating in “draft” mode, where every action is visible for human confirmation. Over time, as teams gain confidence and performance data accumulates, customers can selectively grant autonomy in well‑bounded areas.
Industry consultants warn that automating broken processes risks cementing bad habits in code. If product data is inconsistent or pricing logic opaque, even sophisticated agents can propagate errors at scale. Others raise concerns about over‑reliance on external vendors for critical workflows, particularly in sectors that prize tight operational control.
Barman’s answer is to treat those criticisms as design constraints rather than objections. Faction’s playbook includes data‑quality assessments before deep automation and enrichment of distributor data to make it usable; explicit encoding of pricing and credit policies; and clear logging of every agent action for audit. He also encourages customers to maintain a strong internal process ownership function, rather than outsourcing all workflow thinking to a vendor. In his view, the companies that realize the most value from operational AI will be those that pair strong internal stewardship with specialized external tools.
Barman argues that operational AI is changing what “value creation” means in the manufacturing heartland. Historically, productivity gains came from better machines, cheaper inputs, or more efficient logistics. Now, a growing share of incremental value lies in how information moves: how quickly a distributor can price a complex bill of materials, how accurately a manufacturer can forecast demand, how smoothly cash flows back through the system.
In that context, the achievements Barman has amassed in a short career form a coherent arc. Each step has been about tightening the link between capital, information, and operations. Each has pushed him closer to the messy, often invisible workflows that turn investments into outcomes.
Asked where this trajectory might lead by 2030, Barman offers a characteristically measured vision. “If we do this right, people won’t talk about ‘operational AI’ as a separate category,” he said. “They’ll talk about plants and distributors that simply run better and about teams whose days feel less like treading water and more like building something.” For America’s manufacturing heartland, now under pressure to deliver both competitiveness and quality jobs, that may be the quietest but most consequential standard of value creation.