AI in Clue is designed for intelligence and investigation professionals, where compliance, accuracy, and evidential integrity are critical. Our AI is human-in-the-loop by design, ensuring our users remain in control of decisions. This protects evidential integrity while delivering significant gains in productivity.

Built for the realities of modern intelligence management

AI-assisted functionality in Clue

• Streamlines information triage and reduces manual effort to enable professionals to work more thoroughly and efficiently across large volumes of data.
• Support consistent and auditable compliance using automated redaction and alerts to help teams maintain evidential integrity and meet disclosure requirements with confidence.
• Enhances professional judgement and outcomes, highlighting key entities and relationships across records supports faster, more consistent, and more insightful outcomes.

Features

AI-assisted Watchlists

Create private lists of key entities (names, emails, phone numbers, and more) and be notified when they appear in new records or files.

Watchlists enable users to define key entities and receive automated alerts when matching references are detected in new records and files in Clue. Each Watchlist can be created to include one or more entity types allowing the matching engine to intelligently distinguish between homonyms  (identical terms with different meanings) as new records and files are scanned. For example, a mention of ‘Jordan’ as a location will not trigger a match for ‘Jordan Banks’ as a person, helping to reduce false positives and improve match accuracy.

These notifications help investigators and analysts quickly identify new intelligence related to known entities, without relying on manual searches. By narrowing focus to relevant matches and filtering out noise, users can spend more time assessing meaningful leads and responding to genuine connections. This approach to monitoring ensures that critical relationships and risks are surfaced early and reliably.

AI-assisted Triage

Quickly scan records and attachments to identify people, organisations and other key data, then search, review and create links to other records seamlessly.

Record Triage uses Microsoft Azure AI Named Entity Recognition (NER) to automatically extract and display found entities on key new registers within Clue. Extracted entities are presented as structured data tables and include confidence classifications and duplicate counts with links to Global Search for efficient entity triage. Users can also trigger scans manually for any existing records created before NER was introduced. The feature is initially available on triage specific Registers; Information, Incident, and File.

This feature helps transform large volumes of unstructured text into actionable insights, giving users a clearer starting point for triage and analysis. By automatically surfacing entitles, NER reduces time spent on manual identification and improves data consistency across records. Feedback from users will guide future enhancements and expansion to other record types such as Material, Statement, and Intelligence – ensuring the system evolves in line with real investigative needs.

Record Subscription Notifications

Receive notifications when a record of interest changes, so you stay updated without needing to check manually.

Record Subscription Notifications allow users to stay informed when records they care about are updated or linked to new information. These event-driven updates trigger near real-time email notifications, supporting collaborative workflows and ensuring that key stakeholders are kept up to date as cases progress.

By automatically alerting users to important changes, subscription notifications reduce the need for manual tracking and repeated searches. This ensures that teams remain aligned, helps prevent missed updates, and supports more transparent communication throughout an investigation. In fast-moving cases, timely awareness can be the difference between a quick response and a missed opportunity.

Technology powered by NER

Named Entity Recognition (NER) is the foundation of Clue’s first AI-assisted capabilities. As new information enters the platform, NER automatically identifies key entities such as names, organisations, and locations within unstructured text. Underpinning the AI-assisted features, this structured insight enables you to discover, link, and act on critical information with far less manual effort.

Our responsible approach

Human-in-the-loop by default
AI supports your expertise. You, the professional, decide how to act on outputs.

Compliance first
Clue’s AI adheres to UK GDPR/GDPR and is not considered to be high-risk under the EU AI Act.

Secure and isolated
Customer data is processed via Microsoft Azure AI Foundry NER in UK data centres. No data is used for model training, and strict segregation and encryption safeguards apply.

Transparent and controlled
All features are opt-in and configured with your team.

The future of AI with Clue

With AI, Clue is amplifying the productivity and insight you achieve through the application. Our roadmap includes:

• Generating summaries of lengthy documents and attachments
• Prioritising records based on risk indicators
• Natural Language Search through Chat
• Translation, Transcription, Sanitisation & Disclosure

These capabilities extend Clue’s strengths across intelligence development, investigation management, and impact reporting.

Pricing

Access to our AI-assisted features is included as part of your existing subscription – there are no additional fees for using these capabilities within fair usage limits. To ensure optimal performance and fairness for all customers, usage will be subject to limits that are currently being finalised and will be communicated shortly.

If your usage exceeds these limits, you will still be able to continue using the AI-assisted functionality; however, additional fees may apply for usage beyond the defined thresholds.

Terms & Conditions

1. Introducing Clue’s Named Entity Recognition (NER) service for AI-assisted Watchlists and Triage
1.1 Clue is committed to implementing AI responsibly. When developing AI capability, our overarching principle is to use AI to augment and assist human decision making, e.g. to present information or automate narrow procedural tasks for Clue users.

1.2 Taking this into account, Clue is introducing its AI-assisted technology to surface insights and information to Clue users as follows:
1.2.1 AI-assisted features will initially include Named Entity Recognition (NER) to enable watchlist and triage features. It will recognise a range of key entity types from text, thereby enhancing Clue’s search capabilities.

1.2.2 Clue’s NER service will search information coming into the Clue software application for matches against pre-defined entities in watchlists. Clue’s watchlist feature enables users to create and manage lists of key defined entities. When new information is received into Clue, an automatic search is performed on the new information for entities defined in each watchlist (on a “semantically likely” basis which enables additional precision) and notifications are sent to Clue users if a match is returned to enable follow up action.

1.2.3 Clue’s AI-assisted triage feature will automatically search for and identify key defined entity types in new records and files, and existing workflows allow users to search and establish links to existing records. Existing records can also be searched manually on-demand for the same purpose.

1.2.4 When using features that employ the NER service, the user must decide whether to act upon the outputs of the searches (in accordance with our principle of responsible AI above).

1.2.5 For more information see here.

2. Are AI-assisted features optional?
2.1 Yes. You can choose whether or not to receive AI-assisted features. You can request that it be enabled by accepting these terms. Please contact us so we can enable this sign-off process. AI-assisted features are disabled by default. Clue users themselves can choose not to use the matches generated by AI-assisted features within the Clue application. If you do not accept the terms set out below in section (6), AI-assisted features will not be enabled for use for your Clue instance. Once usage reaches a certain level, additional costs will apply. These will be determined by Clue once further usage data from Customers is available.

3. Which system will Clue use for its AI-assisted features?
3.1 Clue’s AI-assisted services will be provided through services orientated architecture using a Microsoft AI platform.

3.2 Specifically, it will use Microsoft Azure’s AI Foundry platform, and its Named Entity Recognition functionality (“Azure AI Foundry NER”).

3.3 Azure AI Foundry NER is a specialised feature available through Azure AI Language (which is part of the Azure AI Foundry platform). Azure AI Language provides Natural Language Processing (NLP) features for text mining and text analysis.

3.4 To learn more about Azure AI Foundry NER and the use of Azure AI Language, see these overviews:
https://learn.microsoft.com/en-us/azure/ai-services/language-service/named-entity-recognition/overview
https://learn.microsoft.com/en-us/azure/ai-services/language-service/overview

3.5 See also the transparency notices for Azure AI Foundry NER and Azure AI Language at https://learn.microsoft.com/en-us/azure/ai-foundry/responsible-ai/language-service/transparency-note-named-entity-recognition
and
https://learn.microsoft.com/en-us/azure/ai-foundry/responsible-ai/language-service/transparency-note

4. Do Clue’s AI-assisted features adhere to the UK GDPR/GDPR and other applicable legislation?
4.1 Yes. Clue’s AI-assisted features adhere to the UK GDPR/GDPR and applicable legislation. NER and watchlist services as implemented by Clue are not considered to be high risk AI activities under the EU AI Act. Both services surface information to users and do not pose significant risk to individual’s fundamental rights. For more information, including a template controller PIA -which Clue can provide for informational purposes – contact dpo@cluesoftware.com

5 Will customer data be used for retraining Microsoft’s models?
5.1 No. Customer data will not be used by Microsoft Azure to train or improve its shared models, nor does Microsoft share logs or content with model providers. However, Clue may itself keep usage feedback data (not underlying customer data in the customer database) provided by customers (e.g. a thumbs up or thumbs down usage feedback) for internal analysis for making improvements to AI-assisted features. This usage feedback data will not be used by Microsoft Azure or any third party.

Where will the data processed by AI-assisted features be stored?
5.2 Customer data will be submitted to the Azure AI Foundry NER to process text data. This data will be processed in a shared Microsoft Azure tenant set up by and administered by Clue for this purpose, using a UK Microsoft Azure data centre (or if specifically requested by our customers and agreed by us, another data centre location), via a secure API. No data will be shared between customers’ instances, unless customers require this (e.g. if a customer requests data sharing with another entity). Customers’ data will interact with AI-assisted features and Azure AI Foundry NER as a shared service and will be segregated virtually. Results will be provided to the customer’s instance via a secure API. Data is encrypted by Azure AI Language at rest. Clue has also prevented temporary storage of input data, by setting the Logging Opt Out query parameter with Azure AI Language. For more information, see this article https://learn.microsoft.com/en-us/azure/ai-foundry/responsible-ai/language-service/data-privacy For a limited period of time, for initial Clue AI users, existing Microsoft Azure single tenants will be used rather than a shared service, as determined by Clue.

6. Will additional/varied terms and conditions apply, for receiving AI-assisted features?
6.1 Yes. In line with emerging market practice, transparency and our responsible AI principles, in order to receive the benefit of AI-assisted features, Clue requires its customers to provide the following AI risk acknowledgment.

AI risk acknowledgement

6.2 AI-assisted features use AI technology (“AI Technology”). AI Technology is new, still developing and is not without complexity. Like any technology, AI Technology may not always function as intended and there may be times when AI Technology encounters errors or performance issues.

6.3 Any use of AI Technology involves inherent risks, such as bias risk and unexpected behaviour or outcomes. It is not a substitute for human decision making by the end user.

6.4 You therefore acknowledge and agree that your use of our AI Technology is intended as an assistive tool for your authorised users and is not a substitute for users own independent skill, judgment and expertise. You and your users will fully review all AI Technology outputs for completeness, accuracy and regulatory compliance, prior to use, distribution or exploitation, and will use reasonable skill and care when doing so. We shall have no liability to you or any third party claiming through you for any losses or damages arising from your failure to do so.

6.5 You acknowledge and agree that Customers’ data will ultimately interact with the AI Technology as a shared service and will be segregated virtually (a shared service).

6.6 You acknowledge and agree that the AI Technology is subject to Microsoft’s privacy policy at https://learn.microsoft.com/en-us/azure/ai-foundry/responsible-ai/language-service/data-privacy

6.7 You acknowledge and agree that Clue will limit consumption of AI-assisted features per Customer via usage metrics as determined by Clue from time to time, taking into account ongoing usage data by Customers. Clue will be entitled to seek additional charges if the Customer exceeds these metrics. Any additional charges will be agreed with the Customer before they are incurred. If the Customer does not pay these further charges, Clue may disable further use of AI-assisted features.

Getting started

If you’d like to start using the AI-assisted features, you’ll need to be on the latest version of Clue and accept the updated Terms & Conditions, to get started on this process, please contact your Customer Success Manager.