As your client base grows, so does the volume of information tied to each relationship: meetings, emails, notes, tasks, financial details, personal context, and historical decisions.
That information is valuable, but only if you can retrieve it quickly and accurately.
Searching through CRM records, old emails, or past meeting notes is time-consuming and often incomplete. Important context gets missed, especially when preparing for meetings, responding to client emails, or answering ad-hoc questions during the day.
Vega’s Contact agent is designed to solve that problem by letting you ask direct questions, in plain language, and get grounded answers based on your actual client records.
The Contact agent
At the center of this capability is one of our AI agents: the Contact agent.
What is an AI agent in Vega?
An AI agent is a specialized component inside Vega that has:
domain-specific expertise
access to dedicated data sources
a well-defined role in the overall reasoning process
The Contact agent is responsible for retrieving and synthesizing information related to a specific client and prospect across your connected systems. This includes:
CRM contact and household records
Past meeting notes and transcripts
Email conversations (sent and received)
CRM notes and activities
Tasks and workflow context
When you ask a question about a contact (client, prospect, or other relationship), Vega activates the Contact agent to gather the relevant information before generating a response. This includes determining:
where to look for the right data
what information is relevant
how to interpret that information in the context of your workflows
How Vega uses RAG under the hood
This is also where retrieval-augmented generation (RAG) comes into play.
RAG is the process by which Vega:
retrieves the most relevant information from the AI agents (firm data, economic data, market data, and more)
injects that retrieved context directly into the AI’s reasoning process
then generates the final response using both your request and the retrieved information
Because the Contact agent performs its own targeted retrieval across past emails, meetings, CRM records, and connected financial planning and portfolio management systems before the final answer is composed, Vega’s responses are always built on current, specific, and context-aware data, not on generic training knowledge alone.
How "chat with your client data" works
You can chat with your client data in two places:
The Vega chat in the web app
The Vega Outlook plugin and the Vega Gmail assistant
The behavior is the same in all locations. Below are a few practical tips to get the most accurate and useful answers when chatting with your client data.
Referencing the right client or household
To get the most accurate results, include the full name of the contact or the household name in your query. This helps Vega route the request to the correct records, especially when names are common.
For example:
“Give me a recap of my recent interactions with John Smith.”
“What have we discussed this year with the Smith Household?”
Asking Vega to look at a specific object
You can also be explicit about what type of data you want Vega to focus on. Being specific helps Vega narrow the retrieval and return a more focused answer.
For example:
“What notes do I have on Jane Doe?”
“Summarize recent emails with the Johnson household.”
Asking Vega to retrieve a specific information
You can ask targeted questions about personal, financial, or contextual details that have already been discussed. Vega will look across meetings, emails, notes, tasks, and related records to surface the most relevant information.
For example:
“What did the Smiths say they were going to do about funding their daughter's college?”
“How much does Jane Doe have in her 401 (k)?”
“What has the Martinez household mentioned recently about travel or major life changes?”
Controlling the time range
You can constrain the time window Vega should look at. If you do not specify a time range, Vega will use a reasonable default based on the question.
For example:
“What did we discuss with the Wilsons in our last meeting?”
“Summarize interactions with Sarah Lee over the past 6 months.”
“What personal updates has the Martinez household shared this year?”
On the roadmap: cross-sectional search. This will allow you to query across multiple contacts at the same time to identify shared attributes or patterns, instead of looking at one client or household in isolation. For example, you’ll be able to ask questions like “Which clients have a birthday this week?”, “Which clients live in the Dallas area?”, or “Which households mentioned a potential move this year?”, even when that information is spread across notes, emails, meetings, and tasks.
By connecting all your client data into a single conversational interface, the Contact agent is most often used to:
Get up to speed quickly before a meeting
Refresh context before replying to a client email
Recall personal or financial details without searching the CRM
Confirm what has already been discussed or decided
Instead of piecing together information from multiple places, you can ask one question and get a consolidated answer.