Ask a business question. Explore the results. Create agents to act on what matters.
EVA, the Enlighta Virtual Assistant, provides a natural-language interface for exploring vendor information and creating and managing AI agents.
Executives can ask questions about their third parties, investigate relevant information from Enlighta and external sources, and refine their questions through follow-up prompts. EVA can also help authorized users create new AI agents using natural-language descriptions, test their behaviour, and execute or schedule them through the supported workflows.
These capabilities bring two important activities together: gaining insight and creating agents to carry out defined tasks, with human review and accountability.
EVA helps executives explore vendor and third-party information without navigating multiple screens and dashboards. It can query relevant information in Enlighta and combine it with external web information where the applicable sources and tools are configured.
Start with a business question:
EVA queries the available vendor records and returns the matching results for review. In this demonstration, the query returns 88 vendors matching the specified criteria.

Continue the investigation with a follow-up question. Where configured, EVA can use relevant external web information to explore reported financial developments, management changes and other events associated with the selected vendors.

External findings should be checked against their original sources, publication dates, company identity and context before they are used in a vendor-risk decision. Internal records and externally sourced findings remain distinct: the first query matches Enlighta data, the follow-up reaches beyond it.
The investigation can continue in the same conversational context:
EVA can help surface relevant contract and SLA information available through the configured tools and data.
Available where the necessary fields and queries are supported in the configured EVA environment.
Demonstrated above, including the follow-up on adverse financial news over the previous six months. Results and source attribution are specific to the demonstrated session.
A follow-up can investigate external developments concerning vendors identified in an initial Enlighta query. Internal records and externally sourced findings are presented as distinct sources.
EVA is not limited to answering questions. Authorized users can describe a task in natural language and use EVA to create a new AI agent for a defined purpose.
Describe the task you want the agent to perform. EVA prepares a proposed configuration that the user can review before confirming the action. This provides a natural-language path to creating a new agent without requiring the user to build the entire configuration manually.

Human review remains part of the process. Users should check the proposed configuration and dry-run behaviour, confirm appropriate actions, and monitor the agent after it is activated. A proposed action takes effect only after the user reviews and confirms it.
These use cases illustrate the kinds of agents users can describe. Each states what the agent monitors, what information it uses, what it produces and how the result reaches a responsible user.
EVA supports configured tools that can query information and, where authorized and available, support read-write operational workflows. One example from the roadmap:
The availability, permissions, approval requirements and exact behaviour of each configured tool depend on the implementation. EVA does not execute arbitrary actions or modify records without appropriate configuration and authorization.
Creating an agent is only the starting point. The full set of agent-management capabilities includes:
These capabilities help users manage a portfolio of agents according to their operational needs. Each is available only in releases and configurations where Product has enabled it.
EVA helps users explore information and create or manage agent workflows, but its outputs must remain subject to appropriate review and oversight. Users should validate material findings, review agent configurations, assess dry-run behaviour where supported, and confirm actions in accordance with their organization’s permissions and governance policies.
See how EVA supports natural-language vendor insight and agent creation through a guided product walkthrough.
EVA is Enlighta's natural-language assistant. It helps users explore vendor and third-party information through questions and follow-ups, and, where authorized, create and manage AI agents from natural-language descriptions.
Where the appropriate sources and tools are configured, EVA can combine relevant information from Enlighta's application data with external web information. Coverage depends on the configured sources, tools and permissions of the implementation.
Yes. Authorized users can describe a task in natural language, review the proposed configuration that EVA prepares, and confirm, execute or schedule the agent through the supported workflow, subject to the enabled capabilities and permissions.
Depending on the available functionality, users can test agent behaviour, execute or schedule agents, query and monitor them, start or stop them, and review execution metrics.
Where configured and authorized, EVA supports operational workflows through configured tools, such as dispatching vendor emails using predefined or custom templates. The available actions depend on configuration, permissions and approval requirements.
No. EVA provides a conversational way to explore information and configure agents. Dashboards, records and configured workflows continue to support their respective functions.
A guided EVA walkthrough is available on request. Interested teams can contact Enlighta or visit the Demo Center for the demonstrations currently available.
Screenshots show the actual EVA interface from an approved demonstration environment. Agent and tool capabilities described here depend on the deployed product configuration, and no results or findings have been fabricated.