How a San Mateo customer-support AI startup became the tip of the spear for IFS's €15B industrial agentic AI push — and whether the pivot from "CX ops" to "Digital Workers" actually holds together.
IFS Loops is the rebranded, reoriented remnant of TheLoops — a San Mateo, California customer-support intelligence startup founded in 2020 by Somya Kapoor, a former SAP and ServiceNow product executive. TheLoops raised a modest ~$8.75M seed round in 2021 (Dell Technologies Capital, Tidal Ventures, Westwave Capital) and spent four years building a narrow, well-reviewed product: real-time case context, QA automation, and escalation prediction for SaaS support and success teams.
On June 26, 2025, TheLoops was acquired by IFS — a Swedish enterprise software company selling ERP, EAM (enterprise asset management), and field service software to asset-heavy industries, itself valued at roughly €15 billion as of April 2025 (EQT and Hg as co-control shareholders, TA Associates, ADIA, and CPP Investments as minority holders) and reporting more than €1B in ARR growing 30%+ year-on-year. IFS did not fold TheLoops into an existing product line — it renamed the entire company "IFS Loops," kept Kapoor as CEO, and repositioned the platform almost overnight from a SaaS customer-support tool into an "industrial-grade agentic AI platform" for manufacturing, energy and utilities, telecom, aerospace and MRO, and construction.
Co-founder of TheLoops; retained as CEO post-acquisition. Former SAP and ServiceNow enterprise software leader. Public face of the "Digital Workers" narrative — frequent conference speaker (Ai4, HumanX, IFS Industrial X Unleashed) on why most agentic AI pilots fail.
IFS Loops now reports into IFS's broader product and AI organization. IFS's own leadership (CEO Mark Moffat and team) has publicly framed the acquisition as the foundation of an "Industrial AI Workforce" strategy spanning IFS Cloud, EAM, and field service.
The original engineering and product organization built the case-intelligence and QA-automation stack now being re-marketed as "Digital Worker" infrastructure. Team continuity is a genuine asset — this is not a from-scratch industrial AI build.
EQT (majority control) and Hg Capital, with TA Associates, ADIA, and CPP Investments as minority investors following the April 2025 stake sale that valued IFS at €15B. A PE-controlled cap table with growth-at-scale expectations sits above every product decision Loops makes.
IFS Loops now sits at the intersection of two very different, fast-moving markets: the crowded, VC-flooded AI customer-support agent category (Decagon, Sierra, Forethought, Intercom Fin, Ada) that TheLoops originally competed in, and the emerging, incumbent-dominated enterprise agentic AI / industrial AI category (SAP Joule, ServiceNow, Salesforce Agentforce, Microsoft Copilot) that IFS is now asking it to win. Both markets are converging on the same underlying pattern — "LLM plus tools in a loop," orchestrated with governance and audit trails — but the buyers, budgets, and proof requirements are entirely different.
Analysts describe a $1.2T+ opportunity as Microsoft, Salesforce, ServiceNow, SAP, Oracle, and Palantir race to embed autonomous agents into core enterprise workflows. Salesforce Agentforce alone reached $1.4B in ARR with 114% growth in its first year.
IFS's core market — ERP, EAM, and field service for manufacturing, energy, aerospace and defense, and construction — is large, sticky, and historically slow to adopt AI relative to horizontal SaaS. IFS's own 30%+ ARR growth suggests the AI narrative is already pulling incremental spend.
Decagon and Sierra are both well-funded unicorns; Intercom's Fin and Forethought compete on helpdesk augmentation. This is the market Loops actually has five years of production data and customer proof in — and it's the market IFS's messaging has quietly de-emphasized since the acquisition.
| Company | Model | Moat | Overlap With IFS Loops | Threat |
|---|---|---|---|---|
| SAP Joule | ERP-native agent layer inside S/4HANA | Deepest ERP data model in the industry; massive installed base | Industrial/ERP-embedded agentic AI, same buyer (ops & finance leaders) | HIGH |
| ServiceNow (Now Assist) | Workflow-native governance platform + AI Control Tower | ITSM/enterprise workflow incumbency; governance-first trust story | Governed agent orchestration, ops workflow automation | HIGH |
| Salesforce Agentforce | CRM-native autonomous agents | $1.4B ARR, 114% growth, huge CRM install base | Customer-facing agent workflows — direct hit on Loops' CX heritage | HIGH |
| Decagon / Sierra | Pure-play autonomous customer-support agents | Deep VC funding, unicorn valuations, fast iteration on CX-only use cases | Directly competes with Loops' legacy Agent Assist / AI Autopilot layer | HIGH |
| Intercom Fin / Forethought / Ada | Helpdesk-augmentation AI agents | Native integration with existing support stacks (Zendesk, Intercom) | Overlaps with Auto QA Analyst, Knowledge Generation, CX Analyst | MEDIUM |
| Microsoft Copilot / Copilot Studio | Productivity-suite + Dynamics-embedded agents | M365 ubiquity, Azure OpenAI relationship, low-friction distribution | Enterprise agent building; also one of Loops' own integration partners | MEDIUM |
| C3.ai / Palantir Foundry | Broad industrial/enterprise AI + data platforms | Long OT/IT integration history, federal and industrial reference accounts | Competes for the "industrial AI platform" narrative and budget line | MEDIUM |
| Aveva / Rockwell / Honeywell Forge | OT-layer industrial software adding AI copilots | Decades of plant-floor and process-engineering trust | Adjacent — asset/process layer rather than workflow/ops layer | LOW–MEDIUM |
SAP Joule and ServiceNow's Now Assist mean the two most obvious channels for reaching industrial and enterprise-ops buyers are also building competing agent layers natively into systems those buyers already run. IFS Loops has to win on being IFS-native — a much smaller aperture than "industrial AI" broadly.
The idea that ERP systems of record and agentic "systems of action" need to converge is gaining real traction — IFS is betting its acquisition on being early to that framing. But SAP, Oracle, and Microsoft are making the identical argument with far larger R&D budgets and distribution.
As agentic pilots mature, buyers are shifting focus from "can it act autonomously" to "can I trust, audit, and roll back what it did." ServiceNow's AI Control Tower and SAP's compliance-anchored Joule agents are explicitly selling governance. IFS Loops' SOC 2/HIPAA/GDPR posture is necessary but not sufficient here.
Decagon and Sierra have raised at unicorn valuations specifically to win the support-agent category Loops originally built for. A mid-sized, newly-acquired business unit competing on both fronts at once is structurally under-resourced against either pure-play.
Winning in enterprise/industrial agentic AI in 2026 requires strength across four axes simultaneously:
Agents are only as good as the operational data model beneath them. ERP-native players (SAP, IFS itself) have a structural advantage over bolt-on tools.
Mission-critical workflows in aerospace, energy, and manufacturing require approval chains, rollback, and audit trails — not just accuracy.
Buyers in these categories are conservative. Published, verifiable case studies with hard metrics move deals; roadmap slides don't.
Whoever already owns the buyer relationship (the ERP, the ITSM tool, the CRM) has a durable wedge that a point-solution AI vendor lacks.
The product line is genuinely two layers stitched together: the legacy CX Ops layer TheLoops built and sold for five years, and the new Industrial Digital Worker layer IFS is building on top of the same platform. Both are live on theloops.io today, under the same navigation, priced and sold through what appear to be two different motions.
Surfaces case context, suggested next actions, and product/customer signals inside the agent's existing workspace (Zendesk, Salesforce, ServiceNow, Jira, HubSpot, Slack, Teams). This is the original TheLoops product core.
Replaces manual sample-based QA review. G2 reviewers report QA coverage moving from ~1% to 100% within 90 days of deployment at reference accounts.
Rounds out the CX suite — AI-powered self-service, automatic detection and filling of knowledge-base gaps, and continuous sentiment/escalation-risk scoring across the support queue.
Templated Digital Workers — customer order manager, field technician, supplier coordinator — that autonomously manage multi-step tasks: dispatch, supplier coordination, inventory replenishment, across ERP, EAM, and field-service systems. 10 workers and roughly 50 agentic skills at launch, with a stated roadmap to 100+.
Lets customers create, monitor, and deploy their own agents for complex workflows using natural language — reducing dependence on IFS professional services for customization. Launched around Q1–April 2026.
A genuinely distinct GTM motion: rather than selling only to IFS's own customer base, IFS Loops offers itself as embeddable infrastructure for other ISVs — "accelerate your agentic roadmap" without building an agent platform from scratch.
| Segment | Buyer | Problem Solved | Representative Customers |
|---|---|---|---|
| SaaS support & success orgs | VP Support / CX, Head of Support Ops | Manual QA, inconsistent case handling, slow escalation detection | Bloomreach, Outreach, KOHO, Sauce Labs, Tricentis, Gainsight |
| Manufacturing operations | VP Operations, Plant IT / Digital Transformation lead | Fragmented operational, service, and quality data across systems | Not yet publicly disclosed with named case studies |
| Energy, utilities & resources | Head of Field Service / Asset Management | Regulated, complex asset operations; compliance and response speed | Not yet publicly disclosed with named case studies |
| Telecommunications | Network Operations leadership | High-volume, event-driven network and service signals | Not yet publicly disclosed with named case studies |
| Aerospace & MRO | Maintenance & Compliance leadership | Precision maintenance, inspection, and documentation workflows | Not yet publicly disclosed with named case studies |
Reported at reference CX accounts post-implementation — one of the few hard, third-party-verifiable (G2) metrics available for the platform.
Cited improvement within 90 days of rollout, moving support teams from sample-based to full-coverage quality review.
Customer satisfaction improvement reported by reviewers — modest but directionally consistent with the QA and MTTR gains.
Christina Augustine, COO of Bloomreach, is the platform's most visible customer reference, citing impact on both frontline support and success functions simultaneously.
All CX-layer metrics predate or are independent of the IFS acquisition and reflect the original TheLoops product. No equivalently specific, named metrics have been published yet for the industrial Digital Worker layer.
The starkest strategic fact about IFS Loops is that its go-to-market now has to serve two structurally different buyers through one website, one brand, and — as far as public information shows — largely the same underlying platform.
Fast-moving, product-led, comfortable buying and deploying point solutions in weeks. Cares about MTTR, CSAT, QA coverage, and agent productivity. This is the buyer TheLoops built five years of credibility with.
Risk-averse, compliance-driven, multi-quarter procurement cycles, deeply embedded in incumbent ERP/EAM relationships. Cares about auditability, uptime, integration with existing OT/IT systems, and vendor stability. This is the buyer IFS is asking Loops to win.
These two segments do not just have different needs — they evaluate vendors on almost opposite criteria. Buyer A rewards speed and iteration; Buyer B rewards governance and proof of scale. Selling to both from a single undifferentiated homepage risks under-serving both: CX buyers may see "industrial-grade" framing and wonder if the product has drifted from their use case, while industrial buyers may find a support-QA case study (Bloomreach) unpersuasive for a plant-floor dispatch decision.
IFS Loops inherited a genuinely well-reviewed, metric-backed CX product and a capable, retained founding team — that is a real asset most acquirers don't get. The critique below focuses on the gap between the "industrial agentic AI platform" IFS is marketing and what is publicly demonstrable today.
Vision: IFS Loops becomes the trusted, auditable agentic execution layer across IFS's enterprise installed base — and, via OEM, across other ISVs' products — proving out one shared "Digital Worker" abstraction across both support-ops and industrial-ops workflows rather than running two disconnected businesses under one name.
North Star Metric: Autonomous Task Completion Rate — the percentage of Digital Worker actions executed end-to-end without human intervention across live, paying industrial accounts (not pilots). This single metric forces the organization to prioritize proof over positioning. Target by end of 2027: 40%+ of in-scope workflow steps completed autonomously, up from an unreported (likely low single-digit) baseline today.
The Strategic Pivot: From "one brand marketing two products" to "one platform, two proven motions" — CX Ops and Industrial Ops — each with its own named case studies, its own success metrics, and a clearly articulated shared technical core.
Formally segment messaging, case studies, and buyer journeys into two tracks — Loops for Support Ops (the proven CX motion) and Loops for Industrial Ops (the new IFS-driven motion) — while keeping one shared underlying agent platform and skill framework. Buyers self-select into the track built for them instead of wading through undifferentiated "Digital Worker" messaging.
Commit to publishing 3–5 rigorous, named industrial case studies — manufacturing, energy/utilities, and aerospace/MRO — within two quarters, each with hard, verifiable metrics (dispatch time reduction, first-time-fix rate, compliance audit pass rate) held to the same evidentiary bar as the existing Bloomreach CX reference. No case study, no claim in the marketing copy.
Invest disproportionately in Agent Studio and OEM self-serve tooling so that other ISVs building on IFS Loops become a distribution multiplier beyond IFS's own sales force — the one strategic advantage neither SAP Joule nor Salesforce Agentforce is structurally positioned to offer at the same openness.
| Risk | Severity | Likelihood | Mitigation |
|---|---|---|---|
| Industrial pivot fails to produce proof before PE-driven growth targets force a re-prioritization | HIGH | MEDIUM | Fast-track 3–5 named case studies (Bet 2) within two quarters; report interim leading indicators (pilot-to-paid conversion) to IFS leadership monthly. |
| SAP Joule or ServiceNow out-message Loops on governance before the whitepaper ships | HIGH | HIGH | Prioritize governance transparency (Initiative 3) ahead of new feature marketing; borrow credibility from IFS's own compliance posture in EAM. |
| CX customers churn due to perceived deprioritization as industrial messaging dominates | MEDIUM | MEDIUM | Ship the CX-track segmentation (Bet 1) first; maintain visible CX product investment and roadmap communication. |
| A shipped "skill" causes a compliance or safety incident in a regulated industrial account | HIGH | LOW–MEDIUM | Formal review gate (Initiative 5) before any skill ships into regulated workflows; conservative autonomy defaults with human-in-the-loop escalation. |
| Kapoor or founding team depart before the pivot's proof points materialize | MEDIUM | LOW | Retention packages tied to OKR milestones (O1–O3); document institutional and technical knowledge outside any single individual. |
The 50-to-100+ skill sprint is impressive velocity but a liability in regulated workflows. Slow the shipping cadence for anything touching compliance-sensitive industrial accounts until the review gate (Initiative 5) exists.
The Bloomreach 37% efficiency gain and the MTTR/QA/CSAT figures are real and valuable — for the CX audience. Presenting them near "industrial-grade" copy without a clear label invites exactly the credibility gap this report identifies.
Either publish the technical substance behind the claim or retire the phrase. Vague model claims are a bigger reputational risk in industrial/regulated buyers than in the SaaS market Loops came from.
Loops cannot out-spend or out-distribute the ERP incumbents. Its differentiated angle is proof, focus, and OEM openness — competing on marketing volume plays to IFS Loops' weakness, not its strength.
IFS Loops is a genuinely interesting acquisition built on a genuinely solid product. TheLoops' CX ops platform earned real, verifiable customer results — a 45% MTTR reduction, full QA coverage, a named COO testimonial — the kind of proof most AI startups spend years chasing and never get. IFS was smart to keep the founding team and CEO rather than absorb and dissolve the business.
But the "industrial agentic AI platform" IFS is now marketing to manufacturing, energy, telecom, and aerospace buyers is, as of today, a repositioning exercise running well ahead of its evidence. Every hard number on the site belongs to the old business. The new business is described entirely in adjectives — "industrial-grade," "governed," "context aware" — with no named customer, no published metric, and no disclosed technical substantiation for its most specific claims. That's a normal and forgivable state for a pivot fourteen months old. It becomes a real strategic risk if IFS's growth expectations for the €15B business outrun the two-to-three quarters it will realistically take Loops to generate credible industrial proof.
The path forward is not complicated, even if it is not easy: stop trying to sell one undifferentiated story to two very different buyers, publish the proof the industrial claim needs, and lean into the one structural advantage — IFS's existing enterprise distribution and the OEM channel — that neither Decagon nor SAP Joule can easily replicate.
Sources: theloops.io, ifs.com, IFS press releases and investor updates, PRNewswire, G2 reviews, Crunchbase, PitchBook, Verdantix, ERP Today, TA Associates, EQT, White & Case, Futurum Group, HumanX 2026 agenda, Tech Talks Daily podcast. Analysis as of August 2026. Metrics are company- or reviewer-stated as cited; no independent audit was performed.