IFS has expanded its industrial AI ambitions through a strategic partnership with Anthropic, positioning the enterprise vendor to accelerate deployment of safe, domain-specific AI across manufacturing, aerospace, energy, and field-service environments. Announced alongside a joint session titled “The Secrets to Industrial AI: In Conversation with Anthropic,” the collaboration aligns Anthropic’s safety-driven AI platform with IFS’s industry-specific applications and Nexus Black engineering capabilities.
The partnership arrives at a moment when industrial organizations are attempting to operationalize AI beyond back-office tasks. While generative AI has made measurable strides in software engineering, financial modeling, and digital workflows, mission-critical environments such as plants, depots, field operations and maintenance ecosystems that still contend with fragmented data, legacy diagrams, tribal knowledge and non-digitized processes. According to Anthropic’s head of enterprise, Garvan Doyle, the challenge is not building better models but safely transferring AI into real-world, high-stakes contexts.
IFS CEO of Nexus Black, Kriti Sharma, framed the collaboration as a pragmatic counterpoint to AI hype cycles. “When you’re operating in aerospace, defense, energy, and heavy manufacturing, you need a partner that understands the operational consequences of AI failure,” she says.
Doyle reinforced the sentiment, noting that Anthropic selected IFS due to its field credibility, domain reach, and experience operationalizing technology in complex industrial settings.
Building AI for the Real World
The partnership centers on deploying Anthropic’s latest Claude models into IFS’s industrial workflows, which includes asset performance, maintenance engineering, quality, scheduling and field service. Anthropic’s research breakthroughs in early 2025 enable models to generalize technical knowledge and reason across unstructured operational data, including diagrams, documentation, equipment manuals and sensor logs.
Doyle emphasized industrial AI requires more than a strong model: “The magic is in the full solution stack—model intelligence, the experience layer for technicians, and the evaluation discipline that defines guardrails,” he says.
Doyle also highlighted the growing importance of “context engineering” and evaluations frameworks, which allow domain experts to define safe operating boundaries without constraining AI usefulness.
IFS’s Nexus Black engineering group has been central to rapid co-development, using AI-assisted engineering techniques to iterate “ten to one hundred times faster,” according to Doyle. Both organizations have shifted to forward-deployed development models by building with real customers in factories, plants and field operations rather than relying on theoretical workflows.
From Demonstrations to Deployable Products
In the session, Sharma showcased early products already in customer pilots, including technician-assist capabilities, automated inspection intelligence, and reasoning agents for maintenance planning. Doyle responded to skepticism about speed by pointing to two forces: exponential model scaling and real-world iteration loops that compress delivery timelines.
Both leaders also stressed industrial AI remains under-explored compared to knowledge-work automation. With new scientific advances allowing models to generalize beyond digital simulations, AI is capable of tackling economically valuable physical-world tasks. The companies position this partnership as a catalyst for closing that gap.
What This Means for ERP Today Insiders
There is a shift toward operational decisions with AI-augmented workflows. For technology executives managing asset-intensive environments, the IFS–Anthropic integration signals a shift toward AI copilots that can parse engineering drawings, maintenance logs, and compliance documentation in seconds. Early adopters such as including European utilities piloting similar reasoning models have reported inspection-cycle reductions of 25 to 30% and measurable gains in technician productivity.
Clearer evaluation criteria are needed when selecting industrial AI partners. The market for reasoning-based AI in enterprise operations is expanding at double-digit annual growth, with vendors like SAP, Oracle, IBM, and ServiceNow also embedding domain copilots. What distinguishes IFS’s approach is its emphasis on verification frameworks and field-tested guardrails—criteria CIOs and CTOs should prioritize when procurement teams assess vendor maturity, model safety, and readiness for regulated environments.
There need to be smoother integration paths into existing ERP and operational systems. Companies such as Suzuki Motor and Rolls-Royce, which have already embedded IFS automation and analytics, provide case studies on how to layer new AI capabilities onto established ERP architectures without disrupting maintenance, inventory, or scheduling processes. Their success underscores best practices: start with narrow reasoning tasks, integrate evaluation loops with SMEs and address change-management barriers early to avoid slowdowns during plant-level adoption.
