The customer data, i.e. the knowledge of what the FINDIQ software processes, is stored on Hetzner servers in the European Union (video) - if Software-as-a-Service hosting is selected.

• findIQ provides the method, artificial intelligence and software for processing the data, but at no time has ownership of the customer data, unless otherwise mutually agreed.

• After any termination of the contract, the customer data or knowledge will be returned to the customer and deleted on the findIQ side, whereby findIQ’s artificial intelligence for processing the data remains with findIQ.

The current list of our GDPR-compliant subcontractors and the respective processing purpose can be found in the DPA (order processing contract).

• First Choice: findIQ operates, maintains and updates the software as a Software-as-a-Service (SaaS) cloud service, whereby our software is hosted on Hetzner servers in the EU.

• Second Choice: The software is deployed on the customer's private cloud, which only indirectly enables findIQ to provide control and support and can therefore slow it down.

• Third Choice: The software is deployed on-premise on the customer's servers, which places operation and maintenance in the hands of the customer. This means that the customer is also responsible for fully ensuring IT security and must provide dedicated resources for this purpose. If there are also updates to the findIQ software, installing them in the customer's environment involves considerable additional work, and FINDIQ-side access in the event of support is also more complex.

  Recommended system requirements:

  • 4-core CPU

  • 16 GB RAM

  • 120 GB Disk space

  • Ability to run Docker containers

  Note: For optional feature, our software connects to GDPR-compliant subcontractors, which are listed in our DPA. These features can be deactivated on request or have to be configured by the custoemr in case of a private cloud or on-premise deployment.

• The implementation of maintenance, updates, backups and up-to-date security measures are permanent expenses that the customer saves and can have handled by findIQ.

• In addition to IT security at the highest level, findIQ always brings best practice experience from its customer portfolio. To date, the majority of recommendations and a high level of satisfaction have centred on the provision of the software as a SaaS cloud service, particularly with regard to the rapid provision of the software and quick responses in support cases.

  • We offer an uptime rating of better than 99.5% and provide an incident response team to ensure it stays that way.

• Fully automated backups of the entire system are carried out every 2 hours, which is above the industry average in terms of the best possible data backup.

• The backup system is also automatically checked for functionality.

• The backups are also stored in the Microsoft Azure Cloud to provide greater redundancy - with full end-to-end encryption. The data center is located in Germany.

• We have a strict speration between our internal IT and our production enviroment (SaaS) where the application data of our customers are stored. Access on the production enviroment is strongly restricted.

• We are following a zero-trust strategy with our internal IT, which is supplemented by additional security measures such as full-device encyrptions and intrusion detection systems.

• We use modern security measures for our production enviroment (SaaS) such as:

  • A Brute Force Detection system

  • A modern IAM (Identity and Access Management) system

  • Firewalls and virus scanners

  • Strict access and update management for the sofwtare and servers.

  • Automated security tests

  • Encryption at rest and of in-/outgoing data transfers (TLS/HTTPS encryption)

• We offer secure authentication procedures such as SSO (through integration with e.g. Microsoft Entra ID) or 2FA.

• We offer our customers a fine-grained role and group-based user management system to keep full controll on access and restrictions regarding their knowledge data.

• We regularly have external IT security audits (carried out by LocateRisk GmbH) and are also available for individual IT audits of the respective customer at any time before the start of a new partnership.

• We have a GDPR compliant DPA & TOM

• We are currently undergoing a ISO 27001 certification.

• Our hosting provider Hetzer has an ISO 27001 certification and additional external audits by TÜV Rheinland

• Service-relevant operations (such as past diagnoses) are visibly tracked for the customer in a service history.

• System-side operations are logged by findIQ (e.g. error messages, access logs and system perfomance metrics)

Only work-related information is documented in the logbook – i.e. who opened or processed a case, when this happened and what steps were taken.

• Name/initials of the person who opened or processed a case

• Time of creation and changes

• Status of the case (e.g. in diagnosis, in repair, resolved)

• Reference to the machine/system

• Result of the fault diagnosis

• Specific fault that occurred

• Work carried out and steps taken to rectify the fault

The data is used exclusively for traceability in the service process, better collaboration and continuous improvement of the knowledge base.

The logbook entries are visible to all authorized users who, within the scope of their role, have access to the respective system or machine. They are not limited to the person who opened the case. This ensures that relevant information is available for joint processing and that all parties involved share the same level of information. Access is clearly regulated through a role and permission concept.

• Personal information, such as individual user accounts, can be edited or deleted by the user themselves.

• After any termination of the contract, the customer data or knowledge will be returned to the customer and deleted on the findIQ side, whereby findIQ artificial intelligence for processing the data remains with findIQ.

findIQ uses three roles, plus a group structure that controls access to machines and templates.

1. Operator has read-only access. Through the Service Guide, an Operator can run the guided troubleshooting and find the matching solutions. During a real service case, they can also submit user suggestions (new symptoms or causes). An Operator cannot edit the knowledge base itself and has no access to the Knowledge Hub.

2. Editor can additionally maintain knowledge. Editors have access to the Knowledge Hub, where they create and edit templates, chains, heatmaps, and routines. Editors also review incoming user suggestions and either add them to the knowledge base, adjust them, or reject them.

3. Administrator has all the rights of an Editor and, on top of that, manages the organization. Only Administrators can invite new members, assign them roles and groups, and manage the organization's settings. Administrators also have access to the Insights Panel.

The group structure:

User groups control who can access which machines and templates, for example separated by department or customer. You create user groups in the settings. Using the key icon on a given machine or template, you then select the matching group and assign the access rights. The machine list can be filtered by user group in the navigation.

Security-relevant reports from third parties can be sent confidentially to security@findiq.de at any time. Incoming reports are assessed by our Incident Response Team and handled promptly.

• It is important for us not to burden our machine service employees with yet another isolated solution, which is why we can not only integrate, but WANT to.

• The FINDIQ architecture is therefore designed to integrate into existing system landscapes, e.g. field service management systems, ticket systems, maintenance management systems and service portals.

• We have standard interfaces (OpenAPI-compliant HTTP interface/REST API) for this purpose.

• In the event of integration, a single sign-on function (SSO) can be used to provide fast and uniform access for the user (e.g. with Microsoft Entra ID or any other OpenID-Connect compliant IAM system).

• Customizing in the form of individualising the user interface with company colours and logo is possible without any problems and is clearly intended.

• Changing the layout, i.e. shapes and arrangements, is possible but more time-consuming and will be charged accordingly.

• By default, the software the software requires the internet to utilise the full range of features including most of the artificial intelligence functions.

• In the event that internet access is not possible on site for technical or strategic reasons, we are currently providing a limited offline mode which includes the most important informations and functionalities to service a machine on site. If requested, we can add you to the closed beta.

• Hardware and device requirements for the offline version:

  • Browsers: Safari or Chrome

  • 3 GB disk space

  • 6 GB RAM

  • Android Version 15 or higher

  • iOS only latest version

• findIQ uses a novel and self-developed AI process.

• The AI process is the sole property of FINDIQ GmbH and is operated independently by FINDIQ GmbH on hosted servers.

• For optional knowledge extraction Microsoft Azure AI models are used, all features using Microsoft Azure are optional and can be disabled. This data is not used for training these models.

• We use AI to "translate" the knowledge base (error matrix) into a dynamic, user-guided error diagnosis that guides the user step-by-step from error symptoms to the cause of the error as quickly as possible.

• A further AI process has been implemented to optimise the knowledge base in a self-learning manner (feedback system) with frequent use of the software:

  • On the one hand, the step-by-step queries change automatically in their sequence and the diagnoses become faster and better over time.

  • On the other hand, updates and additions to the knowledge base can be made repeatly in the course of use "in the field", which are processed via the feedback system.

  • Based on both types of feedback, the AI generates suggestions on how to optimise the knowledge base, which can be accepted or rejected by an admin.

• The specially developed AI of the FINDIQ software is a combination of several self-learning statistical procedures that emulate the natural reasoning and learning processes of humans.

• The selected AI processes have been developed and validated in real industrial conditions as part of a research project from 2020 to 2022 between several Fraunhofer Institutes and industrial partners.

• The AI methods FINDIQs used differ fundamentally from the generative AI approaches, also known as Large Language Models (LLM), such as ChatGPT (see below), which have become popular with the public since 2023 and are also increasingly being evaluated by industry.

• Our core AI procedures are NOT a Large Language Model (LLM), also discussed in public as ChatGPT, but a model-based approach.

• For optional features, findIQ is using some LLM based processes to extract knowledge out of documents. Those features can be turned off.

• The choice for the type of our core AI models was made for various reasons, at the end of an extensive validation phase (see below):

  • findIQ AI enables deterministic (not affected by chance) fault diagnosis and self-optimisation of expert knowledge in an environment where information about machines and processes is often imperfect and heterogeneous (unlike, for example, the mass of freely available information on everyday topics or B2C areas on the Internet).

  • findIQ AI does not add false information or allow expert knowledge to be reinterpreted in a dubious way, which is important in the context of e.g. troubleshooting technically complex systems (machines and plants) in order to avoid aggravating machine or production errors and ultimately also to avoid personal injury.

  • The findIQ AI and how it learns is permanently under the control of the customer, specifically a "system supervisor". They are provided with their own functional area in which they are shown the AI-based changes to the knowledge base transparently and can intervene and control them. In contrast to the reinforcement learning of LLMs, the changes are not randomised.

  • findIQ AI consumes significantly less computing power than generative AI methods and is a more sustainable alternative in times when the gloable energy consumption of data centres is coming under increasing criticism.