Embedded AI in Industry: IFS's Strategy for ERP and Supply Chain
As industrial AI initiatives multiply, manufacturing companies often swing between two extremes: waiting for "the right solution," or spreading themselves thin across too many parallel AI projects with no measurable return on investment. At the Tech For Industry Show, Sébastien Delauré (IFS) presented his position: fully embed AI into business processes rather than treating it as a separate topic.
IFS, a European ERP Vendor Specialized in Industry
IFS presents itself as a European leader in enterprise software, of Swedish origin, present in more than 90 countries with over 7,000 employees, generating revenue of 1.023 billion in 2024. The vendor deliberately focuses on a limited number of sectors in order to offer ready-to-use functionality: aerospace and defense, construction, engineering, service, energy and resource management, manufacturing, transport and telecommunications.
One Unified Platform: IFS Cloud, Enhanced by IFS IA
The IFS portfolio is built on a single platform, IFS Cloud, which covers Field Service Management, ERP and Asset Management. This foundation is enhanced by IFS.IA, the layer that embeds artificial intelligence into these processes, along with an ecosystem of partner solutions: IFS Assist (ticketing), Copperleaf (asset investment planning), POCA(digitizing shop-floor work instructions), Seven Bridges (AI applied to transport management) and SOFON (advanced warehouse management).
Three Pillars of Industrial AI, from Simplest to Most Advanced
Pillar 1: Ready-to-Use Native Use Cases
IFS offers around 60 AI use cases natively built into the ERP, Field Service Management and Asset Management, organized around six areas: content generation, recommendation (via a built-in copilot that relies on the customer's own data, unlike an external general-purpose copilot), contextual knowledge, forecasting and simulation, constraint-based optimization (production planning and technician scheduling, factoring in geographic proximity, SLAs, skills and customer time windows), and maintenance-side anomaly detection to improve asset availability.
Pillar 2: Agentic AI, or "Digital Collaborators"
The second level concerns autonomous AI agents, called "digital collaborators," deployed via the IFS Loops platform. Their role: automate complex, repetitive workflows by interfacing with the existing application ecosystem (Teams, Slack, Jira, Microsoft tools, SAP) through standard connectors. These agents run continuously, but remain part of a decision loop where the human manager keeps the final say — for example, to approve a supplier follow-up.
IFS offers preconfigured digital collaborators, mainly around the supply chain: customer order management (Customer Order Manager), materials and inventory replenishment (Material/Inventory Replenisher), supplier order reconciliation (Supply Order Manager) and invoice matching. More recent agents now cover Field Service (planning, dispatch, knowledge base).
According to IFS, companies that have deployed this type of agent see an average gain of around 70% on the processes concerned, with the gain varying depending on the starting point and the business process. Processing non-digitized customer orders, for example, is estimated by external consulting studies to account for around 40% of the time spent reconciling these orders with the ERP.
Pillar 3: Custom Projects, Co-Built with the Customer
The third pillar covers specific use cases, built jointly with the customer in a "start-up mode," drawing on the best tools on the market rather than IFS technology alone. Defining the use case can take just a few hours, with initial results generally available within 2 to 3 weeks, and measurable benefits within 3 to 12 weeks, ahead of any potential large-scale rollout.
To accelerate these projects, IFS offers several preconfigured accelerators: Nexus Resolve (AI-assisted diagnostics for field technicians, via equipment scan and photo), its predictive and Field Service variants, Master Data (cleaning and consolidating data reference sets), Nexus Comply (regulatory compliance for aerospace, notably tracking manufacturer service bulletins), and an optimization module for manufacturing.
Three Recommendations for Starting an Industrial AI Project
IFS offers three practical pieces of advice for companies looking to get started:
- Target a concrete, repetitive, high-volume process, with a realistic rather than an ambitious goal.
- Measure a tangible gain on a target panel of users, rather than running a POC driven solely by IT with no clear ROI indicator.
- Scale up deployment, site by site for multinational groups, once the use case has been validated and industrialized.
Explainability and Governance: The Conditions for Trust
One point comes up repeatedly in the presentation: however high-performing an AI may be, it becomes risky if it cannot justify its actions. IFS therefore highlights its platform's ability to trace, audit and measure all actions carried out by digital agents, with safeguards and systematic human interaction to validate or override decisions made.
Key Takeaways
The approach presented by IFS illustrates a fundamental trend in industrial AI: rather than multiplying isolated initiatives, the goal is to embed AI directly into existing business processes — ERP, supply chain, maintenance — through a progressive approach (native use cases, autonomous agents, then custom projects) and a clear governance framework that ensures traceability and explainability.
Watch the full session, along with the other Tech For Industry Show replays, on our dedicated page.
