Quin is Vitally's AI assistant for understanding and acting across your customer base. It works with the customer context already in Vitally to help you ask questions, find what matters, and move from insight to action. Quin was previously called AI Copilot.

Quin is one assistant you can use in two places:

Quin in Objects lives inside individual Accounts, Organizations, Conversations, Meetings, Notes, and Custom Objects. Open Quin from the main navigation and that same assistant works across your entire book of business, so you can run analysis and take action without opening a single profile first.

Quin can access and understand all of your Vitally data, including dynamic properties like product events and custom fields. When you start a chat, it searches across every available property, decides what's relevant to your prompt, and builds the right query on its own. It applies filters and sorting and pulls only the relevant information into context.

In practice, Quin can do anything you can already do with Views, across every supported object. It can also combine data across objects, pulling from more than one at a time to answer a single question.

Open Quin from the main navigation when you want to:

From the main navigation, Quin can read and reason across the following objects:

Open Quin from the Quin button in the main left navigation.

From there you'll land on a full-page workspace with direct access to your recent global chats, AI Memories, and Saved Prompts. Chat names appear in the navigation by default. Hover over any chat to see its timestamp. Your chat happens in the main content area, so you have room to read responses, review citations, and keep working.

During the open beta, Quin has usage limits when you use it from the main navigation. Quin in Objects and AI Summaries don't count toward these limits and aren't limited at this time.

Limits are tracked per user. Each person on your team has their own allowance, and you can only see your own usage.

There are two limits:

A usage indicator sits below the chat input on the right side and shows how much of your usage limit you’ve used. The indicator doesn’t show how full the current chat’s context window is, so a chat can reach its context limit while the indicator reads well under 100%. Select the indicator to open the usage panel, which shows:

When you reach either limit, Quin shows a message that you've hit your limit and when it will reset. You can use Quin from the main navigation again once the limit resets. Limits can't be increased during the beta period. Quin in Objects stays available while you wait.

The best way to learn what's possible is to start chatting. Ask Quin what it can do, how it might solve a specific problem, or why a particular prompt didn't return what you expected. It can explain its own reasoning and limits.

A few prompts to get you going:

For more ideas, see the Quin Prompt Library.

Because Quin knows your Vitally user ID by default, prompts about your own work, like "Who are my top accounts?", resolve without extra lookups, which keeps responses faster.

Each answer ends with a suggested next step and one or two alternatives, so you can choose where to take the conversation next.

While Quin works through a question, its reasoning streams into the working section next to the tool calls it runs. The section shows the last few lines by default. Click it to expand the full reasoning. This makes it easier to see why Quin took a particular action and to spot where a run went off track.

You can point Quin at an existing table view or hub instead of describing its filters in chat. Paste the view's URL into your prompt, and Quin reads the view's name, filters, sort order, and tab, then searches the same set of records you see in Vitally. Browser URLs, vitally:// links, and bare paths all work. This works with saved views, hub table URLs, and default lists. You can also ask Quin to describe the view itself, for example "what filters does this view use?" or "what object is this list built on?"

Quin only reads views you have permission to open. If a view is missing or you can't access it, Quin tells you so and doesn't guess at the filters.

Relative date filters such as "this quarter," "last month," or "today" carry over from the view. They are calculated with your fiscal calendar, the same way they work in Explorer.

Example: "Look at the accounts in [table view URL] and flag any renewing in the next 60 days with no meeting logged this quarter."

You can create Notes and Conversations directly from Quin in the main navigation, without navigating to a profile first. The flow matches Quin on Accounts and Organizations, and a template picker lets you search for and apply a Conversation Template by name.

This opens up workflows like:

Example: "Find accounts renewing in the next 30 days and draft a check-in email for the highest-risk one using the renewal template."

Quin can suggest changes to existing Tasks, Notes, and Custom Object records, and can assign Key Roles on Accounts and Organizations. Ask in plain language and each proposed change shows up on a review card, so you can check everything before it's applied. One prompt can update several records at once.

Quin can change custom traits, including custom traits on Custom Object records, plus these built-in fields:

Quin can assign Customer Success Manager, Account Executive, or any custom Key Role. The review card shows the teammate Quin picked, and you can choose someone else before accepting. Accepting works the same as assigning the role yourself, so assignment emails and Playbooks run as usual.

Example prompts:

Quin can search the web. Ask it for recent news or other public information about an account, a company, or an industry. A web search returns snippets from relevant pages, and Quin summarizes what it finds. When a snippet isn't enough, Quin can read the full contents of a page. It only opens URLs you include in your prompt or links that came back in its own web search results. Some websites block AI agents, so Quin can't read every page. Web search also works in Quin in Objects on Account, Organization, and other record pages.

Example: "Find recent news about Acme and summarize anything that could affect their renewal."

Example: "Read [press release URL] and tell me what it means for Acme's expansion plans."

Web search and web fetch are both on by default. Admins can turn either one off for the whole workspace from Settings > Security, in the Quin section. Click Update Settings to save the change.

With Web Fetch off, Quin can still search the web and summarize the snippets it gets back. It just can’t open the pages themselves. With both off, Quin works only with data already in Vitally.

Quin in Objects is available on Accounts, Organizations, Conversations, Meetings, Notes, and Custom Objects. It helps you surface insights, ask the right questions at the right time, and act faster on key customer interactions. Quin in Objects includes AI Summaries, AI Actions, and Ask AI.

Choose from a set of automated actions on an Account or Org 360, Conversation, or Meeting to speed up follow-up. How you open Quin depends on the record you're in:

Example: "Make Dana the CSM for this account."

The core of Quin, whether you open it from the main navigation or inside a record, is asking questions about your customers and your engagements with them to dig into risk, opportunity, sentiment, and more.

When answering a question, Quin can search your objects and property values to find the specific records and values it needs. This works in Quin from the main navigation and in Quin on Accounts and Organizations. You can also tell Quin where to look, for example "do a dedicated search of the Product Requests custom object." When Quin searches custom object content, it searches only the custom object type your question is about, so unrelated custom objects don't crowd out the records you asked about.

Quin can also load your workspace's configuration objects, so it can filter on and assign real configured values instead of free text. This covers Project Categories, Project Statuses, Task Statuses, Task Categories, Task Tags, Note Categories, Note Tags, Account Key Roles, and Organization Key Roles. You can ask things like "which projects are in the Onboarding category," "find notes categorized as Escalation," or "create a task for the renewal and tag it Renewal Risk."

When your question is about Vitally itself, such as how a feature works or where to find a setting, Quin shows a button that opens the in-app support chat with your question already filled in. Send it there to get an answer from Viv, Vitally's support assistant, or to reach the support team.

The Quin input supports basic rich text formatting, which is also kept in your Saved Prompts. You can apply bold, italics, underline, strikethrough, and inline code, and structure your prompt with headers, bulleted lists, numbered lists, and links. Formatting helps Quin parse multi-part requests and keeps reusable Saved Prompts easy to read.

Each chat has a limited context window. As you ask questions, Quin pulls the data it needs into that window (account traits and properties, related records, citations, and its own responses), and the window fills up over the course of a chat. Quin has a dedicated tool for calculating totals, averages, and other aggregate metrics across many records, the same way Vitally's metric widgets do, so those questions don't require pulling every record into context. Questions that pull in a lot of raw record data fill the window faster.

When a chat approaches its limit, Quin warns you that you're nearing the context limit. When the limit is reached, Quin displays "This chat has reached its context limit. Please start a new chat." and can't take more questions in that chat. Click New Chat to reset the context and continue. When you use Quin from the main navigation, you can check how full the context window is by selecting the usage indicator below the chat input and looking at Context in the usage panel. The percentage on the indicator itself tracks your usage limits.

If Quin can't generate an answer, it explains what happened in plain language instead of showing a technical error, and tells you what to do next.

These messages stay in the transcript, so they're still there if you reload the chat or return to it later.

When Quin suggests several objects in a single run, the transcript groups them under shared headers instead of showing a separate status row for every suggestion. Suggestions of the same object type merge into one header, such as Suggested 19 Tasks. Suggestions that span multiple object types for the same Account bundle together under that Account, such as Suggested an Escalation, a Jira issue, and a Deal, along with a chip showing the Account name. Click the chip to open that Account in a new tab.

While Quin works, its background steps collapse into a single live section that shows its most recent activity. When the run finishes, that section settles into a compact summary such as Completed 6 steps. The retrieval step shows which content Quin searched, including the custom object's name when it searches custom object content. This live section shows what Quin is doing. It doesn't show a pass or fail status for individual steps. If a step doesn't work out and Quin can't work around it, Quin explains that in its response.

Your Quin chat history shows the live status of every chat. Each chat displays a visual indicator, and statuses update in real time across Quin and Quin in Objects. Opening a chat clears its unread state automatically, and background polling keeps statuses accurate even if your connection drops.

Each chat is titled automatically from a summary of its first response, so your history stays easy to scan even when your first message was long or pasted content. Titles update as chats complete, and a chat you rename yourself is never overwritten.

To rename or delete a chat, hover over its name in the history sidebar, click the ... menu, and select Rename Chat or Delete Chat. Changes show up instantly across every open Quin view.

Quin includes AI Memories, which are instructions you save once and apply globally across your account. Use AI Memories to store specifics about your business, your preferences, your brand voice, and anything else you want Quin to keep in mind. Quin takes those instructions into account every time you prompt it.

Because a global memory applies to every session across your workspace, keep memories short and broadly useful. Avoid loading them up with definitions for specific traits, fields, or custom objects. That detail belongs on the field descriptions themselves (or in a Saved Prompt), where Quin can pull it in only when it's relevant. Overloaded global memories can pull Quin's attention toward details that don't apply to the question at hand.

Saved Prompts let you keep your favorite prompts handy to reuse later. Create one for any workflow you repeat often, or whenever a Quin session goes especially well and you want to recreate it. Saved Prompts are individual to your account and aren't shared across your team.

How is Quin in the main navigation different from Quin in Objects?

Quin in Objects is scoped to a single record: one Account, Org, Conversation, Meeting, Note, or Custom Object. Quin in the main navigation works across your whole book of business and can combine data from several objects at once to answer a single question. During the beta, Quin in the main navigation has per-user usage limits and Quin in Objects does not.

Who has access to Quin?

All Vitally users have access to Quin on Account and Org 360s and Conversations. Users with the Gong integration or the Vitally AI Meeting Recorder enabled for their account also have access to Quin in Meetings.

Can admins see their team's Quin usage?

Not yet. Each user can see only their own Quin usage.

What kinds of questions can Quin answer across my book of business?

Anything you could express as a View, plus questions that span multiple objects. Good examples include "Which accounts have a health drop in the last 7 days and no logged activity?", "Which accounts renewing in the next 90 days have a health score below 60?", and "What's my average health score across my book of business?"

Why didn't my prompt return what I expected?

Ask Quin directly. It can explain why a prompt didn't work and suggest how to rephrase it. Vague or heavily multi-part prompts tend to do better when broken into clearer steps.

What data is analyzed in Quin in Objects?

Quin in Objects can be used on Accounts, Organizations, Conversations, Meetings, Notes, and Custom Objects. On an Account or Organization, Quin has access to all of its traits and properties. Quin can also search your objects and property values to retrieve details it doesn't already have loaded, so it can answer more specific questions about records tied to that Account or Organization.

Why did my chat reach its context limit?

Each Quin chat has a limited context window. Think of it as a bucket that fills as Quin looks up records, does analysis, and generates answers. The number of questions you've asked matters less than how much data each answer needs. Questions that pull in a lot of raw account data fill the window faster. Aggregate questions like totals, averages, and counts across many records use a dedicated tool instead of pulling each record into context, so they don't fill the window the same way. When the chat is close to full, Quin warns you. When it's full, Quin shows "This chat has reached its context limit. Please start a new chat." The limit is per chat, so starting a New Chat resets the context.

What data is analyzed in AI Summaries?

AI Summaries use Conversations, Notes, Tasks, Projects, Docs, Custom Objects, NPS responses, Custom Survey responses, and Meeting transcripts. Up to 500 Conversations are reviewed per summary.

If we only use Gong, what context is used for Ask AI?

If you use Gong and not AI+, the context used for Ask AI and Summaries is limited to 90 days.

Can Quin answer questions about how Vitally works?

Quin is scoped to your customer data, not to Vitally's product documentation. When it recognizes a question about the Vitally product, it shows a button that opens the in-app support chat with your question already filled in, so Viv and the support team can answer it from the Help Center.

What kind of Conversation Template should I use for AI follow-ups?

The Conversation Template forms the foundation of the follow-up email Quin creates, so we recommend you format these templates like a prompt rather than a traditional email template. Prompting the AI to pull out key details from the meeting, summarize them, and suggest action items keeps your brand voice consistent while personalizing every follow-up.

Here is an example of a good AI follow-up Conversation Template:

Does Vitally use my data to train AI models?

No. Data processed through Vitally AI is not used to train AI models. It's used only to generate the output you requested, such as a summary, task, or note.

Is customer data retained or reused by underlying model providers?

No. Vitally uses AWS Bedrock as its AI infrastructure provider. Model providers (for example Anthropic or Cohere) do not have access to your prompts, data, or completions, and your data isn't made available to them for training or reuse.

What specific data is retained when using Vitally AI?

Inputs, outputs, and prompts follow Vitally's existing retention and deletion protocols, the same as all other customer data in the platform.

How long is AI-processed data retained?

AI-processed data follows Vitally's standard retention policies. AI features add no additional or extended storage.

Are AI-generated outputs (like summaries or tasks) stored?

Yes. AI-generated outputs such as a summary, task, or note are stored in Vitally if you choose to save them, just like any other object created in the platform. They follow the same retention and deletion policies.

Who are Vitally's AI subprocessors?

Vitally's AI features run on AWS Bedrock. AWS is the only subprocessor involved in handling AI-related requests.

Can you confirm that AWS Bedrock model providers cannot access my data?

Correct. AWS Bedrock uses a "Model Deployment Account" structure, so model providers don't have access to customer prompts, completions, or logs. They also can't use your data for training or inference outside your request.

What models does Vitally use?

Vitally uses Amazon Bedrock, which gives us access to large language models from trusted providers like Anthropic, Meta, Cohere, and others. We use different models for different parts of our AI feature set. Your information is never used to train these models, and it never leaves our secure environment.

What happens if an AI model is unavailable?

If AWS Bedrock reports that the model handling your request is unavailable, Vitally automatically retries the request with the next configured model. Quin, AI Actions, and AI Summaries keep working through model-specific or regional provider outages, so you don't have to wait through repeated retries against a model that's down.

Is Vitally AI under the scope of our existing DPA and Privacy Terms?

Yes. Quin falls under your existing DPA and privacy terms.

Does using Vitally AI require a separate customer opt-in under GDPR or other frameworks?

No. AI processing is part of Vitally's standard product functionality under your existing DPA and privacy terms. No additional opt-in is required.