What is Lyra Agent?
Lyra Agent is a CX-trained AI agent built by Chattermill that runs customer feedback analysis. It analyzes signals across the entire customer journey, spots problems and opportunities, and grounds every insight in quantified data.
It works like a human analyst and an extension of your team - query it, analyze with it, or delegate to it. It was trained to be safe, accurate, and secure to help you understand your customers at scale.
How can I talk to Lyra Agent?
Chattermill users can use prompts to communicate with Lyra Agent. The best approach is to speak to Lyra Agent like you would with a coworker - simply ask a question in plain English, then follow up, set a task, or give instructions. The prompts can range from simple questions to complex requests:
Simple questions:
"What are customers saying about our delivery experience?”
“What was our NPS last quarter?”
"What issues are customers reporting with the mobile app?"
Complex requests:
"Give me the full picture on checkout — the sentiment, the top issues, and what customers are asking us to fix."
“Our CSAT dropped last month — dig into why, show me the top three issues behind it, and pull a few customer quotes for each.”
“Why is our NPS higher for iOS customers than Android? Break down what's behind the gap.”
Tips for getting started
Start simple: It helps to begin with straightforward questions, so you can get a feel for how Lyra Agent responds.
Be specific: The more specific your request, the better Lyra Agent can assist you.
Iterate: Feel free to ask follow-up questions to refine the analysis.
Explore: Once you're comfortable, try going beyond basic questions — ask Lyra Agent to build a chart, a table, or a full report.
Use separate chats: Where possible, it helps to start a new chat when moving to a different topic rather than continuing in the same thread. This keeps context focused and responses more accurate — a good habit with any AI agent, not just Lyra.
What can I do with Lyra Agent?
Ask any question, get analyst-quality answers
Ask Lyra Agent anything about your customer feedback, and get a precise, evidence-backed answer. Our agent was trained for customer experience intelligence, so every answer is grounded in verified data.
Get proactive insights
Working constantly in the background, Lyra Agent scans your feedback, analyzes it, and automates the work behind the scenes – delivering proactive insights, while you sleep. Learn more about Proactive Insights here.
You can use Lyra Agent to:
Find top issues and wins: See the most frequently mentioned issues, and the opportunities worth doubling down on.
Spot growing issues: See what's trending in customer feedback — which issues are spiking this week.
Identify bugs and defects: See what customers complain about, and where the biggest friction points, bugs, and errors are.
Explain metrics: Understand how key metrics such as NPS, CSAT, and customer sentiment have changed and why, with comparisons across time.
Understand churn reasons: Identify why customers leave — for example, why they abandon checkout.
Identify feature requests: See what customers want built next, and surface opportunities for the product roadmap.
How can you access Lyra Agent?
Lyra Agent is currently in Closed Beta, and we're actively gathering feedback from early users to shape the feature ahead of a wider rollout.
Customers who'd like early access should reach out to [email protected] or their dedicated XLG Programme Lead to request it.
You can also sign up for the waitlist, and we'll get you set up to run it against your own customer feedback data.
For early access users: you can access Lyra Agent in the Chattermill app. There is nothing to set up — available as soon as you log in. You can ask a question, get an answer, follow up, set a task, or give instructions.
Please note you can also connect Chattermill to your own AI agent, using the Chattermill MCP.
Understanding Lyra Agents Capabilities
What analysis tools does Lyra Agent have access to?
Lyra Agent has access to all of Chattermill's advanced analysis tools, and puts them to work the moment a question is asked. It processes all available customer feedback data instantly, applying the relevant filters and segments needed to answer that specific question — no manual report building or segment creation required.
From there, it retrieves metrics, themes, feedback summaries, observations, customer snippets and quotes. It is also capable of building and adding charts and tables directly into the response.
How can I share the results with my team?
Every conversation is saved as its own chat, so you can always return to it later. Any charts generated within a chat can be downloaded and shared via the Reports page.
We're also working on making it possible to save results directly as a report or dashboard — giving you a shareable insight artifact with just a click.
Does Lyra Agent have memory?
Yes. Lyra Agent remembers useful context across sessions, at two levels: organization-level memory, shared across your whole team, and user-level memory, specific to you.
It builds this memory automatically as you use it, but you can also tell it what to remember directly — for example, explaining what a specific segment or metric means to your team, so it applies that context in every future conversation.
Are there any usage limits?
For the first 90 days, all Chattermill customers get unlimited usage of Lyra Agent, so you're free to explore it as much as you'd like.
Does it integrate with Slack or Microsoft Teams?
We’re currently working on bringing Lyra Agent to Slack and Microsoft Teams, and we’re looking forward to making this integration available soon.
Security
How can I make sure my data is secure?
Security is built into how Lyra Agent works. Analysis stays where your data lives — nothing is exported into an external tool or agent.
The same user roles and permissions that govern Chattermill today apply to every Lyra Agent interaction, and Space carry through automatically, so no one sees data outside their existing Chattermill access.
To see how Chattermill handles AI security, privacy, and compliance, click here.