Dealfront AI Features

1. Purpose and Scope of this Document

At Dealfront we offer you AI Features that make it even easier to find what you are looking for. Whether you are looking for companies, contacts or other insights, by using the power of popular Large Language Models (LLM) our AI Features understand your intent and deliver the results you were hoping for.

We’ve built these features with security and compliance at their core, so that your data stays protected and you stay compliant. Our AI Features do not expose confidential or personal data to externally hosted LLM. The LLM only sees the information you include in your prompt. No information from your account is shared with the LLM provider.

This document gives you an overview of how our AI Features work and how they are designed to meet data-protection and AI-regulatory standards. If you’d prefer not to use them, you can easily turn the AI Features off at any time.

2. Overview of AI Features

2.1 AI List Enrichment

Our AI List Enrichment tool helps you add valuable insights to the companies that matter most to you. Simply select a list of companies and ask the AI to generate a new data point on a specific topic you are interested in. The AI enriches your list of companies with the desired information.

For example, if you are interested in environmentally friendly companies, select the companies most relevant to you and ask the AI: “How environmentally friendly are the companies?”. The result could look as follows:

Company A, Score: 9 / 10, Explanation: Company A operates on 100 % renewable electricity, publishes audited sustainability reports, and has achieved a verified 40 % reduction in CO₂ emissions since 2020.

Company B, Score: 2 / 10, Explanation: Company B has no publicly available sustainability targets, relies heavily on coal-based energy, and has been fined for environmental non-compliance.

The feature works with data from our global Dealfront database, including publicly available sources such as company websites and signals. It does not access or use your customer data. The AI List Enrichment tool will access the global Dealfront database to retrieve information on each selected company and forward it to the LLM to generate the results

2.2 AI Target Filter Generation

The AI Target Filter Generation makes it easier than ever to discover the companies or contacts most relevant to you. Simply describe what you’re looking for in natural language. The tool will take on retrieving the relevant information from the database by having the LLM transpose your natural language request into a powerful database query.

For example, if you ask the AI to show all German companies with at least 100 employees that have Data Science teams, the AI automatically converts your request into matching filters such as country (DE), employee count (100+), and job titles (Data Science, AI, Data Engineering). The tool applies these filters to our database and delivers the results instantly. Additionally, the tool will provide feedback to help you understand if your prompt needs refinement to retrieve better results.

To make this work, none of your customer data is exposed to the LLM. Only your prompt will be forwarded to the LLM together with tool specific instructions. The query provided by the LLM will be applied to the Dealfront database without any further involvement of AI.

2.3 AI Signal Alerts

AI Signal Alerts help you stay up to date with the companies that matter most to you. Set up customized alerts based on specific topics or events for a selected list of companies and let the tool inform you about updates on a regular basis (e.g. daily, weekly, or monthly).

For example, you could create an alert for companies in your list announcing the construction of new office buildings. The output might include entries such as “Siemens AG is building a new office in Munich” or “Volkswagen AG plans a new office complex in Wolfsburg to be completed in 2028.”

To generate these insights, the AI Signal Alerts searches the Dealfront global database for relevant signals related to your selected companies. The identified information, along with your prompt, is then processed by the LLM to compile the results. Aside from the user prompt, no user or customer data is sent to the LLM.

2.4 AI Company Insights

With AI Company Insights, you can quickly generate tailored insights about a specific company. Simply select a company of interest and ask a question in natural language. The tool will retrieve information on the company from the Dealfront global database, collect information from the internet and send this information to the LLM. The LLM will analyze the information and provide the answer to your question.

2.5 AI Contact Recommendations

AI Contact Recommendations helps you identify the most relevant contacts for your needs. Start with a selected list of contacts from the Dealfront database and describe in natural language which profiles you want to find. Your prompt will be forwarded to the LLM which interprets your request and returns the necessary filters. The filter will be applied to your list of contacts to return the matching contacts.

For example, you might ask to find all members of Data Science teams based in Germany or Austria. Contact Recommendations will translate this prompt into filters such as location and job title, apply them to your pre-selected contact list, and deliver the relevant results.

Only your prompt together with system instructions is sent to the LLM to define the filters. No other user or customer data is shared. The generated filters are then applied to the chosen subset of the Dealfront database to identify matching contacts.

2.6 AI Web Visitor Summary

AI Web Visitor Summary gives you instant, AI-powered summaries of how companies engage with your website. Simply enter a prompt, for example, “Show me a summary of this company’s engagement” and you’ll receive a clear, conversational overview of their activity.

Your summary might read something like:
“Company XYZ visited your website 50 times in the past 90 days, mainly from Berlin, Germany. Six unique visitors viewed 77 pages in total, spending the most time on your pricing pages — a strong indicator of buying intent.”

Behind the scenes, a LLM analyzes your tracked website events (such as page visits, sources, time spent, and visitor IDs) to generate insights tailored to your question.

All personal identifiers are pseudonymized before data is sent to the AI model, and then safely restored internally once the response is received. This ensures that no identifiable visitor data ever leaves the Dealfront environment.

The result: an at-a-glance, privacy-safe view of each company’s engagement, directly in the Dealfront App, helping you spot your most interested prospects faster.

3. Compliance

Our AI Features do not expose customers to compliance risks. In particular, the only user or customer data that is processed are the respective user prompts.

3.1 Data Protection

Protecting your personal and confidential data is a core element of our AI Feature design and a key part of complying with European and global digital regulations. We know that generative AI brings new challenges for privacy because LLMs may, in rare cases, reproduce information contained in their training data. To avoid this risk, we built our AI Features with data protection in mind and have taken the necessary legal safeguards to protect your data.

To eliminate the risk of a model reproducing your confidential information, your customer data is not forwarded to the LLM. The only individualized data the LLM will see, is your prompt. Otherwise, we only provide the LLM with data from our global database which is collected from publicly available sources on the internet.

This separation of data particularly applies to the tracking data related to website visitors. While we offer website tracking services to identify companies that may be interested in your products, we keep this data strictly separate from the AI Features and do not feed external LLMs with this data.

As an additional safeguard, we only use LLM providers that contractually commit not to use customer prompts or outputs for model training. Thus, even if personal or confidential data is included in a user prompt, it will not be stored or reused by the external LLM provider for model training. This ensures that your data remains private and protected at all times.

Prompts and AI-generated outputs are only retained for as long as necessary to provide the requested functionality and to ensure service quality. Once not needed anymore, prompts and outputs are deleted from our systems and the systems of the LLM providers.

To comply with regulation on data transfers to third party recipients, we have concluded data processing agreements with our LLM providers. Where external LLM providers process data outside the EU/EEA, we rely on appropriate safeguards such as the European Commission’s adequacy decisions or Standard Contractual Clauses (SCCs) to ensure an equivalent level of protection.

Finally, our AI Features are designed purely as assistive tools. They do not make automated decisions or create user profiles that produce legal or similarly significant effects.

3.2 AI Act

We design and operate our AI Features in full alignment with the principles and requirements of the EU AI Act. As part of our compliance approach, we regularly assess our AI systems to ensure that they meet applicable obligations and reflect the spirit of responsible and trustworthy AI.

Our AI Features do not qualify as high-risk AI systems under the EU AI Act. They are assistive tools that support users in retrieving information. They are not intended for use cases that are classified as high-risk under the EU AI Act.

Even though the strict requirements for high-risk systems do not apply, we take the AI Act’s key principles of transparency, accuracy, fairness and non-discrimination serious. To uphold these values, we continuously monitor and test our AI Features to identify potential risks or compliance issues early. That said, AI may occasionally generate incomplete or inaccurate results. Our AI Features are designed to assist your work, not to replace human judgment. We therefore recommend verifying AI-generated information for accuracy and suitability before using it in your specific context.

4. Turning off AI Features

If you feel uncomfortable with the AI Features, the administrator of your Dealfront account can always disable the AI Feature by setting the disable_ai_features flag. This will disable all AI Features described above and hide any onboarding messages and other information that could be perceived by users as a nudge to use AI Features.

5. FAQ

5.1 Which large language models are used?

We use Large Language Models (LLMs) to make our services more intuitive and efficient. Depending on the specific feature, the LLM helps interpret what you are looking for, generate search filters, or summarize relevant information.

The general process works as follows: You describe your request in natural language (the prompt). This prompt, along with limited contextual information from our global Dealfront database and predefined instructions, is sent to the LLM. The model generates a response, which our system then uses to produce the final result, sometimes involving additional data processing within our own systems. With the exception of the prompt you enter, your customer data is never forwarded to the LLM.

5.3 Will my data be used to train AI models?

No. We have contractually agreed with all LLM providers that neither your prompts nor the output generated are used for model training. In addition, our AI Features do not access or process your customer data.

The only individual data ever seen by the AI model is the text you include in your prompt. All other information required for the AI Feature to function is handled within our systems under strict data protection and security controls. We apply robust internal safeguards to prevent unauthorized access or use and to ensure full compliance with applicable data protection laws.

5.4 How does Dealfront ensure data protection compliance for its AI Features?

We ensure data protection compliance through a combination of contractual, technical, and organizational safeguards. All LLM providers we work with act as our data processors under dedicated Data Processing Agreements (DPAs). The LLM providers have contractually committed not to use any prompts or outputs for future model training. This prevents the risk of personal or confidential information being memorized and reproduced by the model.

In addition, our AI Features never access your internal or customer-specific data. They operate solely on data from our global Dealfront database, which is available to all customers. The only individual data element ever seen by the model is the text you include in your prompt.

5.5 How can I protect confidential information when using the AI Features?

Our AI Features are designed so that your confidential information is not exposed to the LLM. The only personal data ever seen by the model is what you actively include in your prompt.

If you prefer not to share certain information, simply avoid entering it in the prompt. However, even if a prompt contains personal or confidential data, this does not create a compliance issue. All LLM providers act as data processors and are bound by Data Processing Agreements as required by applicable law. They have contractually committed not to use prompts or outputs for model training or any other purpose.

In short: Your data remains protected at all times and you stay in control.

5.6 Can I rely on the AI Output?

We design our AI Features to be as accurate and reliable as possible. Their performance benefits from the quality of our underlying databases, and we continuously monitor and test the features to maintain consistency and precision.

That said, AI models can occasionally produce incomplete or inaccurate information and may sometimes generate statements that do not fully reflect the facts. For this reason, we recommend reviewing AI-generated results for accuracy and suitability before relying on them in your specific use case.

5.7 Can I turn the AI Features off?

Yes. You have full control over whether you use the AI Features. Please refer to section 4 for instructions. Turning them off does not affect your access to the rest of the Dealfront platform.

5.8 Does Dealfront use AI for automated decision-making or profiling?

No. Our AI Features are designed purely as assistive tools. They help you find and interpret information more efficiently but do not make automated decisions or create profiles that produce legal or similarly significant effects. All decisions and interpretations remain fully under your control.

5.9 Where is my data processed, and does it leave the EU/EEA?

Your data is primarily processed within the EU. However, as we use popular LLM providers, data sent to the LLM may leave the EU. The effect is limited, though, because your prompt is the only data individualized by you that is transferred to the LLM. We do not process your customer data with LLMs.

5.10 How does Dealfront ensure compliance with the EU AI Act?

Our AI Features are fully aligned with the principles of the EU AI Act. They are assistive tools helping users to identify information and do not qualify as high-risk systems. We ensure transparency and accuracy and continuously test and document our AI Features.

5.11 What happens to my prompt and AI-generated output after I use an AI Feature?

Prompts and AI-generated outputs are only stored as long as necessary to deliver the requested result or perform the requested service and maintain service quality. Afterwards, they are automatically deleted or anonymized. None of this information is used for model training or shared with third parties.

5.12 Can I report an issue or concern with an AI-generated result?