Putting AI Agents to work across eDiscovery use cases

Ask the question. Select the data. Let Aviator Agents turn search into action.

Imagine having an AI Assistant you can converse with using a more natural way to begin and get powerful responses. This allows the inquiry to develop with the matter rather than forcing the legal team to define every possible search or step at the beginning of an investigation or legal action. 

Contextual understanding is particularly important in eDiscovery because legal teams are rarely searching for documents in isolation. They are trying to assess risk and determine what the evidence means for a matter and reconstruct events. An agentic workflow can also surface contradictions, such as when a witness account conflicts with a calendar entry or contemporaneous communication, and route those inconsistencies for focused human review.

In the second part of our two-part blog series, we will look at how and when it is most appropriate to use OpenText™ eDiscovery Aviator™ Agents to support eDiscovery AI workflows. 

How OpenText eDiscovery Aviator Agents works

Aviator Agents can extend legal tech solutions beyond conversational prompting to intelligent search and intelligent automation within the same matter workflow. Users can describe a goal in natural language, and an agent can plan and execute the multi-step search, review, analysis and summarization tasks required to complete it. Each step is logged and traceable, allowing users to inspect the process, validate the output, and re-execute individual steps when needed. 

Ask the question

Legal professionals may begin with an open-ended question or choose an agent aligned with a defined workflow. This lets teams explore without Boolean expertise or dependency on specialists to build every query.

Select the data

Users choose the matter, document population, or data set the agent should examine. This keeps the analysis aligned with the legal inquiry and gives the team control over the scope.

Explore and refine

Aviator Agents analyze the selected information, extract key data for analysis by the LLM, and report relevant findings. Users can ask follow-up questions, adjust criteria, and pursue new lines of inquiry as the matter develops or new facts emerge.  Aviator Agents even go as far as prompting you to ensure you find what you are looking for or refine your focus.

Connect the evidence

Aviator Agents can help identify relationships among documents, people, events, communications and themes, providing context around individual findings.

Turn findings into auditable action

Depending on the objective, Aviator Agents can summarize, classify, organize, and generate outputs that support investigations, document review, reporting, and legal action.

From legal objective to actionable result

Legal teams do not simply need more documents or more AI-generated content. They need to understand the evidence, assess risk, meet obligations, and decide what to do next.

A look into Aviator Agents

Through a centralized dashboard, Aviator Agents provides access to pre-configured agentic templates for immediate use and enables teams to build fully custom agents for matter-specific workflows. Production-ready agents support broad research, search and fact-development needs, along with beta templates providing starting points for more specialized workflows.

General-purpose and research agents

Some matters do not fit a predefined workflow. Legal teams may need to investigate an emerging issue, test a theory or ask a question that could not have been anticipated in advance. Teams may want to define their own instructions in plain language and direct the agent to perform a connected series of searches, analyses and reporting tasks. The OpenText Custom Agent allows teams to create their own agent to support workflows not covered by any specific pre-configured templates. Similarly, teams can use the OpenText Research Assistant Agent to investigate a user-defined topic and synthesize those findings into an evidence-backed research report. These general purpose agents give teams the flexibility to explore information for matter-specific workflows where pre-configured templates are not applicable or useful.

Specific legal workflows for key eDiscovery use cases 

Early Case Assessment & Investigations

During early case assessment or investigations activities, teams use eDiscovery software to understand the facts quickly to make informed decisions about exposure, scope, strategy, and potential resolution. Agentic eDiscovery workflows are particularly valuable when a team needs to understand what happened, who was involved, and how the relevant events unfolded. Aviator Agents can help investigators understand who said what, when and why; identify relationships among people and events; and reveal patterns that keyword searching may miss.

Identify key players

Aviator Agents can help examine the relevant data, identify important people and communications, surface significant events, and explore emerging themes. The OpenText Key Custodian Identification Agent can analyze communications, roles, and relationships to identify key players or individuals in a data set to improve matter analysis. The OpenText Sentiment Discovery Agent creates a report identifying documents matching a target sentiment, specific attitudes, or even reactions that create greater context for the information being analyzed.

Reconstruct events

Applied against specific document sets, the OpenText Timeline Reconstruction Agent helps legal teams create a factual timeline of events, provides document references to statements, and offers next step options. Users can refine the inquiry as the factual picture and case story develops to better understand the evidence most likely to influence strategy.

Data Privacy and Breach Response  

Personal data protection is critical because global privacy laws create legal, financial, and reputational risk. The OpenText Person Discovery Agent can identify documents containing a person’s direct PII or personal data by focusing on known identifiers.  This helps teams improve the completeness and accuracy of person-centric requirements and data subject response efforts.  

The OpenText DSAR Report Agent helps teams identify in-scope personal data and extract third-party PII. The findings can then be organized into an auditable report to support subject rights obligations and regulatory requirements.

The OpenText Data Breach Investigation Agent helps investigators and legal teams analyze potential personal data breach exposure from an identified set of compromised data. It accelerates breach investigations by identifying affected individuals and documents containing PII. Teams can continue refining the findings as the investigation develops, supporting notification analysis, regulatory response, and remediation planning.

eDiscovery Document Review

Search and review quality

Aviator Agents can support the quality and reliability of document review by applying responsiveness criteria across a selected document population, organizing potentially responsive material and flagging uncertain or inconsistent classifications for human review. Using the OpenText Query Support Agent legal teams can develop, test, and refine search queries using metadata analysis and iterative validation to find documents associated with a specific topic or issue – improving search accuracy (precision and recall) while reducing the manual trial and error required to identify the most relevant documents. Users can use these eDiscovery AI tools to evaluate how a query performs, identify potential gaps or false positives, and refine the search to improve precision or recall.

Evaluate document classifications

The OpenText Classification Quality Control Agent helps by reviewing the accuracy of document-review classifications. It improves classification accuracy by automating quality-control review, providing clear reasoning for classification decisions, and identifying opportunities to refine the criteria used for first-pass review. This allows teams to concentrate human quality control on uncertain, inconsistent, or higher-risk results rather than reviewing every classification with equal intensity.

Protect privileged and sensitive information

Privilege analysis is highly contextual. It requires legal teams to evaluate participants, roles, content and the purpose of a communication. Some of the highest risk eDiscovery activities involve identifying information that should not be disclosed. The OpenText Privilege Discovery Agent supports the identification of potentially privileged communications. Using the OpenText Sensitive Content Search Agent, reviewers can compile a list of documents containing sensitive information. Against an identified document set, the OpenText Privilege Log Agent determines privilege type and reason as well as creates a preliminary privilege log for review. Together, these workflows can reduce lower value manual work while helping teams focus their attention and judgment on difficult privilege and disclosure decisions.

From legal objective to actionable result

Moving beyond isolated prompts and disconnected technical steps, Aviator Agents helps set the direction, select the relevant data, and use guided workflows to move more efficiently from question to evidence and from evidence to action. It is powerful agentic eDiscovery in action.

All you need to do is:

Select the data. Ask the question. Let Aviator Agents turn search into action.

For a preview of OpenText Aviator Agents, you can watch the ACEDS webinar where we provide a hands-on walkthrough of three workflows attorneys and paralegals are running today for meaningful operational benefit.

Contact us to discuss whether Aviator Agents may support your eDiscovery activities.