One Platform. Two AI Engines. Here's Why That Matters.

findIQ remains an industrial AI platform — but is expanding it with Quinn, an LLM-based assistant. The diagnostic core remains deterministic and precise, while Quinn handles everything else: building knowledge from unstructured data, expert interviews, field documentation, and reporting. Two AI architectures, clearly distinct and deployed strategically — each used where it truly excels. What is changing is not what findIQ does, but how it is operated, and the time required for commissioning is being significantly reduced once again.

The conversation around AI in industrial service has split into two camps. On one side: generic large language models, powerful at text-based tasks, increasingly present at every trade fair and in every pitch deck. On the other: specialized, model-based AI systems built for precise, deterministic decisions in complex technical environments.

Both have real strengths. Both have real limits on their own.

From the start, findIQ chose the second path. The Industrial Knowledge Intelligence platform is built on a diagnostic core that resolves faults through structured knowledge models, not text generation. That decision hasn't changed. What's new is Quinn — an LLM-based assistant that joins the platform to handle exactly the tasks where language models genuinely excel.

The result is what we call the Dual AI approach: two AI architectures, one goal, each doing the work it was built for.

Why existing approaches fall short

Three approaches dominate the market today for keeping service knowledge accessible. None of them fully solve the problem.

Classic PDFs, wikis, and fault trees are widely used. They also become outdated faster than they can be maintained, are difficult to locate under field conditions, and offer no interactivity. The technician is left alone with a static document and no path forward.

LLM-based chatbots solve the discoverability problem on the surface. But they fail at a critical point: they lack industrial-specific expertise, are prone to hallucination, and provide no structured guidance. In critical infrastructure or on a custom machine, a plausible-sounding but incorrect answer is worse than no answer at all.

findIQ takes a third path. Experiential knowledge is captured in a structured way, linked in a dynamic model, and delivered to the technician step by step. The system continuously refines itself based on field data. The results are deterministic — because diagnostic work is deterministic by nature.

Quinn does not change what findIQ does. Quinn changes how it is used.

Introducing Quinn

Quinn is findIQ's new LLM-based assistant. It takes on the tasks that create real friction in day-to-day service operations but don't belong in the diagnostic core: structuring unstructured data, conducting expert interviews, documenting service cases, generating reports.

A customer survey conducted ahead of Quinn's launch identified the three most important use cases: support during fault diagnosis via voice input, documenting fault cases, and creating and editing maintenance routines. Those three use cases define Quinn's initial scope.

What Quinn does not do is touch the diagnostic logic. The heatmap, the structured knowledge base, the deterministic reasoning at the core of findIQ Intelligence — all of that remains exactly as it is. Quinn is an additional hand, not a replacement for the engine that drives it.

Two components, one platform

The Dual AI approach rests on a clear division of responsibility.

• findIQ Intelligence is the model-based AI at the core of the platform. It links symptoms, causes, and probabilities in a dynamic heat map, guides technicians through a diagnosis, and delivers precise, reproducible results. This is the component that has distinguished findIQ since its founding. It doesn't generate text. It reasons from structured knowledge to a correct answer.

• Quinn is the LLM-based layer that surrounds and supports that core. It accelerates knowledge building, makes the platform more accessible in the field through voice input, and turns service data into actionable reports. Responses from Quinn are grounded exclusively in verified source data and the diagnostic output of findIQ Intelligence — no hallucinations, no fabricated content. It is not a generic chatbot. It is an additional input and output layer on top of a curated, deterministic knowledge source.

How the platform works across the full knowledge lifecycle

The architecture covers three stages, from initial knowledge capture through field use to continuous improvement.

1. Building the knowledge base. Quinn accelerates the first phase significantly. The assistant automatically extracts unstructured data — maintenance videos, PDF documentation, service records — and integrates it with the implicit knowledge captured through expert interviews. Multiple contributors can work on a single knowledge source simultaneously. What previously required considerable manual effort becomes substantially faster.

2. Assistance in the field. In the field, the guided fault diagnosis does what it has always done: identify root causes directly, on average in six steps, with a hit rate of at least 80%. The platform works offline — a non-negotiable requirement in plants without stable network connectivity. Through Quinn, technicians can now interact via voice, document new problems on-site without additional effort, and receive responses grounded in verified diagnostic data.

3. Learning and adapting. The platform continuously improves through reinforcement learning and real user feedback, following a human-in-the-loop principle. Quinn supports this process by generating reports, recommending actions based on verified service data, and preparing workflows for automations and integrations.

The missing layer between data and decision

Most industrial operations today have two well-developed technology layers. ERP and MES systems process business data and support strategic decisions. SCADA, PLCs, and sensor systems deliver raw machine data and show what is happening in real time.

What sits between them — the layer that translates "what is happening?" into "what needs to be done?" — is almost always missing.

That is the knowledge layer, and it is precisely where findIQ operates. The Dual AI approach makes that layer more accessible than ever: less manual effort during knowledge building, more accessible operation in the field through voice interaction, and more substantive analysis of service data through AI-assisted reporting.

In summary, these are the key advantages of Quinn:

• Unlocking unstructured data: Maintenance videos, PDFs, and documentation are automatically processed and integrated into the knowledge base
• Scaling expert interviews: Quinn conducts interviews autonomously, significantly reducing the initial effort required to build the knowledge base
• Voice-driven operation in the field: Technicians can retrieve information in plain language instead of navigating through interfaces
• Documenting service cases: Directly on-site, without additional effort, based on the actual diagnosis result
• Creating reports: Quinn analyzes service data and prepares strategic insights

→ The result is faster population of the heatmap — and therefore quicker access to the company's own knowledge base.

Data security

Moving toward language models raises a legitimate question for industrial customers with sensitive machine and service data. For Quinn, findIQ uses established, business-grade model access: data is processed, not used for training language models. The underlying agreements include appropriate data protection guarantees and are designed for compliance with European data protection requirements.

The bottom line

findIQ remains on its Industrial AI course. The deterministic diagnostic core is unchanged. What expands is the reach of the platform; into faster knowledge building, more accessible field use, and more actionable service intelligence.

Two AI engines. Each doing the work it was built for. That's the Dual AI approach. And that's Quinn.

findIQ is the first Industrial Knowledge Intelligence platform — turning expert know-how into precise, scalable maintenance and troubleshooting guidance technicians can trust.

Frequently asked questions

What is the findIQ Dual AI approach?

The Dual AI approach combines two distinct AI architectures within a single platform. findIQ Intelligence is the deterministic, model-based AI at the core; built for precise fault diagnosis and guided troubleshooting. Quinn is the LLM-based assistant that handles knowledge building, documentation, and reporting. Each engine handles the tasks it was designed for. Neither replaces the other.

What is Quinn and what does it do?

Quinn is findIQ's new LLM-based assistant. It accelerates knowledge capture by extracting and structuring unstructured data from sources like maintenance videos and PDF documentation. In the field, it enables voice-based interaction with the platform. After service events, it supports documentation and reporting. Quinn's responses are grounded exclusively in verified source data and findIQ Intelligence diagnostic output, not generative text with a risk of hallucination.

Does Quinn replace findIQ's existing diagnostic core?

No. The diagnostic logic, the heat map, and the structured knowledge base are unchanged. Quinn changes how the platform is used, not what it does. Existing customer workflows continue to function exactly as before. Quinn adds new input and output options — it does not alter the underlying reasoning that drives diagnostic accuracy.

Why does precise, deterministic AI matter more than a general-purpose LLM for fault diagnosis?

In industrial service, a plausible-sounding but incorrect answer has real consequences: extended downtime, unnecessary part replacements, or safety risk. General-purpose language models generate probabilistic text — they are designed to sound correct, not to be correct with certainty. findIQ Intelligence uses deterministic, model-based AI that mirrors expert reasoning, delivering the same accurate result regardless of who is using the platform, what shift they are on, or how much experience they have.

How does findIQ address the gap between ERP/MES systems and real-time machine data?

Most industrial operations have strong data layers at the business level (ERP, MES) and the machine level (SCADA, PLCs, sensors). What is typically missing is the knowledge layer between them — the translator that converts real-time machine signals into actionable maintenance and troubleshooting guidance. findIQ fills that gap. The Dual AI approach makes the knowledge layer faster to build, easier to access in the field, and more useful for generating strategic service insights.

How does findIQ handle data security when using an LLM-based assistant?

For Quinn, findIQ uses established, business-grade model access from providers whose agreements explicitly prohibit using customer data for model training. Data is processed to generate responses and outputs, not retained or used for training. The framework is designed for compliance with European data protection requirements, making it appropriate for industrial customers with sensitive machine and service data.