Skip to main content

Service Agreement

Last updated 20 September 2026. Reviewed every six months alongside our AI Transparency page.

The terms on which TrakIntel provides its platform to customers located in, or established in, the EEA, the UK or Switzerland. It should be read together with our Privacy Policy and our AI Transparency page.

Parties

This Service Agreement is entered into between TrakIntel.eu (“TrakIntel”, “we”, “us”) and the customer identified on the applicable order form, or, absent an order form, the entity or individual registering for or using the Service (“Customer”, “you”).

This Agreement applies to customers located in, or organisations established in, the EEA, the UK or Switzerland accessing TrakIntel's platform via its EU product. It governs your use of the Service and, together with our Privacy Policy, our AI Transparency page, and any signed order form or Data Processing Agreement, forms the entire agreement between you and TrakIntel.eu.

1. The Service

TrakIntel provides an AI-native technology intelligence and innovation platform (the “Service”), comprising:

  • IRIS, continuous monitoring and scoring of external technology and market change
  • ARIA, technical evidence-building, including patent and research landscapes and white space analysis
  • Venture Clienting, governed sourcing, proof-of-concept management and scale-decision workflows
  • Agent Studio, comprising six AI agents: Signal Agent, Mr Z, IP Agent, Scouting Agent, FinSight Agent and VC Agent

together with associated APIs, dashboards, reports and support. TrakIntel may update, add to or deprecate features of the Service from time to time. We will provide reasonable notice of any change that materially reduces core functionality you are actively subscribed to.

2. Accounts and access

  • You must provide accurate registration information and keep your account credentials secure. You are responsible for all activity under your account.
  • We may suspend or terminate access for breach of this Agreement, non-payment or security reasons, subject to reasonable notice and an opportunity to cure where practicable, except where immediate suspension is necessary for security, legal or safety reasons.
  • You must be authorised to bind your organisation if registering on its behalf.

3. Fees, term and termination

  • Fees, billing cycle and payment terms are as set out in your order form or the pricing published on our website.
  • Unless otherwise agreed, fees are non-refundable except where required by applicable law or in the case of a billing error.
  • Either party may terminate this Agreement for the other's uncured material breach, on 30 days' written notice specifying the breach.
  • Upon termination, your right to access the Service ends. We will handle your Customer Content as described in section 7, our Privacy Policy and any applicable Data Processing Agreement.

As an EEA, UK or Swiss customer, where a statutory right of withdrawal from a distance contract applies to you, that right is not superseded by this Agreement's cancellation terms.

4. Service levels

TrakIntel does not currently commit to a formal Service Level Agreement for uptime, response times or issue resolution. If your order form specifies service level commitments, those terms control for your subscription. Absent such terms, the Service is provided on a reasonable-efforts basis.

5. AI systems, and how they work in this Service

This section reflects the current, factual state of our AI systems and supersedes any inconsistent description elsewhere in our marketing materials. It should be read together with our AI Transparency page, which we keep more frequently updated and which controls in case of conflict on AI-specific facts.

5.1 What the agents do. Signal Agent, Mr Z, IP Agent, Scouting Agent, FinSight Agent and VC Agent each generate a specific type of output — scored signals, position briefs, landscape reports, shortlists, financial models and investment memos respectively — from company, market, patent, research and your own uploaded business data. Some agents operate in a conversational interface; where this is the case, the Service clearly discloses that you are interacting with an AI system, consistent with Article 50(1) of Regulation (EU) 2024/1689. TrakIntel does not operate a public chatbot on its marketing website.

5.2 Delivery without human review. Agent outputs are delivered to you directly, without a TrakIntel staff review step in between. You are responsible for reviewing outputs before relying on them, consistent with 5.5 below.

5.3 Data used, and data in outputs. No personal contact details such as emails are used as direct agent inputs, and none appear in generated outputs. Some outputs, such as a scouting shortlist or an evidence brief, may name identifiable individuals such as founders or executives, as relevant business context. TrakIntel does not treat this as profiling of the named individual's personal characteristics, but you must still handle such outputs in accordance with applicable data protection law if you process them further. You must not upload special category data, as defined in Article 9 GDPR, to the Service.

5.4 Training on your data. If you upload documents or data to the Service, we may use that content to fine-tune a model instance dedicated exclusively to your account. That customer-specific model instance is never shared with, or used to produce outputs for, any other customer. Your data is never used to train or improve TrakIntel's shared, generic foundation models used across other customers.

5.5 Accuracy, limitations and no professional advice. Outputs are generated using machine learning models, including large language models, and may be incomplete, outdated or contain inaccuracies, including content not supported by the underlying sources. We evaluate output quality using internal methods, including BERT-score-based evaluation, but have not commissioned independent third-party accuracy or bias audits at this stage. Outputs are not validated as financial, investment, legal or tax advice and must not be treated as such. Content generated by our agents is labelled as AI-generated in the product interface, and each output records the model and agent version that produced it.

5.6 High-stakes uses. You must not rely solely on Service outputs for decisions producing legal or similarly significant effects on an identifiable individual, including decisions regarding creditworthiness or lending, insurance eligibility or pricing, employment, or access to essential services. If your organisation intends to use outputs for such a purpose, you must notify us in advance at founder@trakintel.ai, as this may require a separate agreement or additional safeguards.

5.6 is an advisory restriction stated in this Agreement. It is not enforced by a separate technical control.

5.7 No formal AI dispute process. TrakIntel does not currently operate a formal process or service-level commitment for contesting or appealing a specific AI-generated output. Questions about a specific output can be directed to founder@trakintel.ai as a general enquiry; this is not a guaranteed review or appeals procedure.

5.8 Our AI Act role and classification. TrakIntel acts as a provider of the AI systems described above under the EU AI Act. Based on an internal self-assessment, not independently verified by external counsel or a regulator, against the Annex III high-risk use cases, we classify these systems as minimal to limited risk. We will notify customers of any change to this classification via our AI Transparency page.

5.9 Models, and where they run. Both foundation models used by the Service are open-source, open-weight models that TrakIntel self-hosts and fine-tunes itself, run entirely on our own infrastructure: one based on OpenAI's open-weight model family, and one based on Google's open-weight model family. Neither is called through a third-party API, so no data, personal or otherwise, is transmitted to OpenAI or Google to generate these outputs. All outputs naming identifiable individuals are generated exclusively by the self-hosted model based on OpenAI's open-weight family. Our primary infrastructure, including both self-hosted models, is hosted with Hetzner in Germany. We also use Microsoft Azure in its India region to generate vector embeddings; per our internal assessment this does not process personal data, though we treat the question of embeddings derived from documents containing names as an open item for legal review. We use Google Analytics for website usage tracking, processed on Google's EU-based infrastructure per our current understanding; because it sets cookies and processes IP addresses, which the GDPR treats as personal data, we obtain explicit consent before setting these cookies.

6. Intellectual property and Customer Content

  • TrakIntel and its licensors retain all rights in the Service, including the underlying AI models, taxonomies and non-customer-specific data.
  • You retain all rights in the data and documents you upload (“Customer Content”). You grant TrakIntel a licence to process Customer Content solely to provide the Service to you, including the customer-specific model fine-tuning described in 5.4.
  • Outputs generated specifically for your account may be used internally by you under your subscription terms. You may not resell, redistribute or sublicense outputs to third parties outside your organisation without our written consent, and must attribute TrakIntel as the source where outputs are shared internally in reports or presentations.

7. Data protection

Processing of personal data under this Agreement is governed by our Privacy Policy and, for Customer Content that constitutes personal data processed on your behalf, by a separate Data Processing Agreement available on request. In case of conflict between this Agreement and that Data Processing Agreement on data protection matters, the Data Processing Agreement controls.

8. Confidentiality

Each party will protect the other's confidential information, including Customer Content, non-public business and technical information, and this Agreement's commercial terms, with reasonable care, use it only to perform this Agreement, and not disclose it to third parties except as permitted by law, court order, or with the disclosing party's consent. These obligations survive termination of this Agreement for 5 years, except for trade secrets, which remain protected for as long as they qualify as such under applicable law.

9. Warranties and disclaimers

The Service is provided as is. Except as expressly stated in this Agreement, TrakIntel disclaims all warranties, express or implied, to the maximum extent permitted by law, including warranties of merchantability, fitness for a particular purpose, and non-infringement. Nothing in this section excludes or limits liability that cannot be excluded or limited under applicable EU or Member State law.

10. Limitation of liability

To the maximum extent permitted by law, neither party is liable for indirect, incidental, special or consequential damages, or loss of profits, revenue or data. Each party's total aggregate liability under this Agreement is capped at the fees paid or payable by the Customer in the twelve months preceding the claim, except for liability that cannot be limited under applicable law, such as death or personal injury caused by negligence or fraud; breach of confidentiality; and a party's indemnification obligations, if any are agreed in an order form.

11. General

  • Assignment. Neither party may assign this Agreement without the other's consent, except that TrakIntel may assign it to an affiliate, including TrakIntel.ai, or in connection with a merger, acquisition or asset sale, provided the assignee assumes our obligations, including data protection obligations.
  • Force majeure. Neither party is liable for delay or failure to perform due to events beyond its reasonable control, including failures of third-party model or infrastructure providers.
  • Severability. If a provision is unenforceable, the remainder of this Agreement remains in effect, and the unenforceable provision is replaced with one that most closely reflects its intent.
  • Review cadence. We review this Agreement, alongside our AI Transparency page, at least every six months, and sooner if our AI systems, sub-processors or risk classification change materially.
  • Order of precedence. In case of conflict: a signed order form or enterprise agreement, then this Agreement, then the AI Transparency page for AI-specific facts only, then the Privacy Policy, then general website content.

EU consumers may also access the European Commission's Online Dispute Resolution platform at https://ec.europa.eu/consumers/odr.

These four documents describe the same facts and are kept in step with each other.