Vesence is an agent built for professional work with documents, files, and business context.
That means Vesence is not limited to answering questions in a chat. It can understand an instruction, use the available context, decide what needs to be done, and carry out work inside the Vesence Workspace.
In simple terms:
Vesence is an agent that can act inside the Workspace, not just respond in the chat.
What “agent” means
An agent is an AI system that can move from instruction to action.
For Vesence, this means the user can describe a goal in normal language, and Vesence can use the available context to help complete the task. It can inspect files, prepare outputs, ask follow-up questions when needed, and save work product back into the Workspace.
How the Agent works with the Workspace
The Workspace gives Vesence a place to work. The Agent gives Vesence the ability to do the work.
The user adds files, connects context, gives instructions, and reviews outputs. The Agent uses that context to read source material, create outputs, organize files, and continue from the available task history.
Together, the chat, Workspace, and Agent make Vesence a file-based working system rather than only a text conversation.
What Vesence can do as an agent
Because Vesence can act in the Workspace, it can:
- Read files
- Create new files
- Edit supported files
- Organize files and folders
- Process structured data
- Compare documents or datasets
- Generate work product in different file formats
- Save outputs back into the Workspace
How Vesence handles larger tasks
For larger or more complex tasks, Vesence can break the work into focused parts and use sub-agents to handle those parts in parallel.
For example, a large contract review might involve different areas such as definitions, payment terms, liability, termination, intellectual property, confidentiality, and schedules. Vesence can assign those areas to focused sub-agents, combine the findings, resolve overlaps, and present one coordinated result to the user.
The user does not need to manage this process directly. The main Vesence agent coordinates the sub-agents and remains responsible for the final answer or output.
Custom agents
For recurring tasks or workflows, users can create custom agents that are tailored to a specific type of work.
- Give the agent specific instructions for how a task or workflow should be handled.
- Connect the agent to specific folders, matters, or source material in OneDrive, SharePoint, or iManage.
- Keep the agent private or share it with the organization.
- Choose whether the agent should be available on the web, in Word, in Excel, or in all supported places.
A custom agent can include task-specific instructions and workflow knowledge, so the user does not need to repeat the same setup each time. It can be kept private for personal use or shared with the user's organization when the same workflow should be available to others.
Custom agents are created on the web. Once created, they can be used on the web and inside supported Microsoft Word and Excel workflows. See: Custom Agents for more information.
Task-specific tools
When a task needs more than a simple text answer, Vesence can use task-specific tools inside the Workspace. Vesence is also a coding agent, meaning it can write and run code to create these tools, automate repetitive steps, and process documents or data when needed.
For example, if the user gives Vesence a math problem, Vesence does not need to guess the answer; it can build a calculator and solve it. Vesence may also create a structured tracker, compare many files, transform data, generate a checklist, or prepare a small workflow-specific output. These tools are created for the work at hand and should support the user's review rather than replace it.
Models, dictation, and web search
Alongside the core agent, Vesence offers a few options that change how you give instructions and what Vesence can draw on. These depend on your organization’s settings and permissions, so you may not see all of them.
Choosing a model
Vesence runs on advanced AI models. Using the model selector in the chat box, you can choose which model handles a conversation. The available models are the ones your organization has enabled, and the choice applies to that chat.
If you are unsure, the default model is a good general choice. You can switch to a different one for tasks that need more reasoning or a different balance of speed and depth.
Dictation
Where enabled, the chat box includes a microphone so you can speak your instruction instead of typing it. Vesence records what you say, turns it into text in the message box, and lets you review or edit it before sending.
This is useful for longer instructions, for dictating context on the move, or when speaking is faster than typing.
Web search
Where your organization turns it on, Vesence can search the web for current information as part of a task, instead of relying only on the files and context you provide.
Organizations stay in control of this: web search can be limited to approved websites, or kept off entirely. When it is not enabled, Vesence works only from the materials and connected sources available to it.
What users should remember
The main point is simple:
Vesence is an agent that can act in the Workspace.
To get the best results:
- Give Vesence the goal, not just a narrow command.
- Make the relevant files available in the Workspace.
- Tell Vesence which source is authoritative if there are several versions.
- Ask Vesence to flag assumptions, missing information, or judgment calls.