Great chat. What’s next?

Agent Goals: Deploy with purpose. Return with insights. 

A good conversation can clarify a question, solve a problem, change someone’s mind, or uncover something you did not know before. 

But in business, the best conversations usually do more. They lead somewhere. 

A prospect decides to take the next step. A learner demonstrates that they understand. A customer shares the information needed to solve an issue. Someone raises a concern you did not know existed. 

That is the idea behind Agent Goals, a new set of capabilities for Kaltura Agentic Avatars that gives organizations more control over what an avatar should accomplish during its conversations. 

Agent Goals introduces two complementary capabilities: 

Capture Data lets an Agentic Avatar present a form during an interaction and collect structured information directly from the user.  

Conversational Insights lets teams define the signals they want to identify from what users say during conversations. 

Together, they turn an Agentic Avatar from something that simply responds into something that can help move an interaction forward and leave the organization with useful information once the conversation is over. 

 

Where do conversations with AI lead? 

Kaltura’s Agentic Avatars already give people a more natural way to interact with an organization’s knowledge. Instead of searching through documentation, sitting through a presentation, or trying to find the right person to ask, someone can simply start a conversation with an avatar trained on organizational knowledge and customized to represent the brand. 

But useful interactions often require more than just an answer. 

A sales conversation may reach the point where a prospect wants to be contacted. A learner may need to demonstrate that they understood the material. A customer support interaction may need a few specific details before the issue can move forward.  

That is where giving the agent goals becomes useful.  

 

Why make an Agentic Avatar curious?  

Sometimes the most useful thing an avatar can do is answer a user’s question. For businesses, the ability to ask the user a question is even more valuable.  

With Capture Data, an Agentic Avatar can present a form directly within the interaction and collect structured information from the user. Instead of sending someone away to another page or system, the next step can remain part of the conversational experience. 

A prospect who has spent five minutes exploring a product can provide the information needed for a follow-up. A learner who has just worked through a topic can complete a short assessment. A customer can provide the account or support details needed to continue.   

The agent is helping the conversation go somewhere. 

 

What can the conversation tell you?  

Explicitly asking for information is only one side of the story.  

A conversation with an Agentic Avatar can also reveal useful context without requiring the user to enter it into a form, simply from actively listening. The questions someone asks, the topics they explore, what they already understand, and the problems they describe can all provide useful signals.  

For prospects, learners, and customers, those signals can reveal readiness, knowledge gaps, urgency, context, or pain points.  

With Conversational Insights, teams define which of those signals matter to them, and Kaltura can surface relevant insights from the interaction. That makes the information useful at two levels.  

At the individual level, teams can understand more about a specific person and respond with greater context. 

At scale, they can use the Conversational Insights dashboard to look across hundreds or thousands of conversations to spot recurring needs, objections, knowledge gaps, and patterns. 

 

Data Capture plus Conversational Insights – a powerful combo 

The combination of the two capabilities is where Agent Goals becomes a gamechanger. 

The structured data collected by Data Capture tells you who the user is. The conversational signals help explain what matters to them.  

Now the organization does not just know that someone showed strong interest in a particular use case. It may know which prospect did, what they care about, and what they are likely to need next. 

That combination can feed very practical downstream workflows. 

A sales team could enrich a lead or account record in its CRM, prioritize high-intent prospects, and give the account owner context before follow-up. Marketing teams could segment or nurture users differently based not only on profile data, but also on the interests and needs surfaced during the interaction. 

For Learning and Enablement, identified learner data combined with conversational signals could inform workflows in an LMS. Someone struggling with a particular topic could be assigned additional material, while recurring gaps across many learners could help teams decide where a course or curriculum needs improvement. 

Customer-facing teams could enrich support or customer-success workflows. 

Across many users, the same information can also reveal broader patterns that inform content, messaging, training, customer experience, and other business priorities. 

 

How it works? 

The new Agent Goals capabilities are configured as goals in the Playbook for an Agentic Avatar.  

For Capture Data, you create a Collect user data goal and define the form the user will see. Give it a title and description, then add the fields you want to collect.  

For Conversational Insights, teams simply describe the indicators they want the agent to identify from its interactions.  

At launch, each Agentic Avatar can have one Capture Data goal and up to four Conversational Insights goals. 

 

Review what the agent learned 

Once an Agentic Avatar is live, the analytics dashboard provides an overview of activity, engagement, and signals that deserve a closer look. 

Session Insights lets teams drill into individual conversations. Each session includes the conversation thread, an automatically generated summary, interaction details, and a score tied to the configured goals. Teams can search, filter, review specific sessions, and export results for further analysis or use in other workflows. 

Across many sessions, those same insights can reveal recurring objections, knowledge gaps, buying signals, or friction points. 

The progression is simple: define what you want the agent to accomplish, collect the information users choose to provide, learn from what the conversation reveals, and use both to inform what happens next. 

 

A good conversation leaves you with something to act on 

Making AI better at conversation was an important step. But a good conversation becomes much more valuable when it leaves you with something you can act on. 

That might be information someone explicitly chose to provide through a form. It might be an insight the conversation revealed about their needs, interests, or readiness. And when you can connect those signals to a known individual, while also looking for patterns across many interactions, the conversation starts to become useful well beyond the moment itself. 

Because the best conversations do not simply end. 

They give you somewhere to go next. 

 

Frequently asked questions 

What is Capture Data? 

Capture Data lets an Agentic Avatar present a form during an interaction and collect structured information directly from the user. Teams can define the fields they want to capture, such as contact details, company information, feedback, qualification data, or other information relevant to the experience. 

What are Conversational Insights? 

Conversational Insights let teams define the signals they want to understand from the conversations their Agentic Avatars are having. These insights are derived from what emerges during the interaction, helping teams understand factors such as user needs, interests, pain points, familiarity, or other indicators relevant to the experience. 

What is the difference between Capture Data and Conversational Insights? 

Capture Data collects information the user explicitly provides through a form. Conversational Insights help identify information revealed through the conversation itself. 

Used together, they can provide a richer picture of both who the user is and what their interaction tells you about their needs or intent. 

Can Conversational Insights be tied to a specific user? 

When the person behind the conversation is known, insights from that interaction can be understood in the context of that individual. The user may identify themselves through Capture Data, or they may already be known through an existing experience such as sign-in, registration, or proactive outreach. 

This makes it possible to combine conversational signals with known user information and build a more useful picture of what that individual may need next. 

Can I also look at insights across many conversations? 

Yes. Conversational Insights are useful at both the individual and aggregate level. Teams can examine specific interactions while also looking across larger numbers of conversations to identify recurring questions, needs, friction points, or other patterns. 

How many goals can I configure? 

At launch, each Agentic Avatar can have one Capture Data goal and up to four Conversational Insights goals. 

Where can I see the information that has been collected? 

Teams can review results in Session insights, where they can see individual sessions, conversation summaries, interaction history, and scores. Sessions can be searched and filtered by date or status, and results can also be exported for further analysis or use in other workflows. 

What can teams do with these insights? 

The value depends on the goal of the Agentic Avatar. A revenue team might use conversational signals to better understand a prospect’s needs or readiness. A Learning and Enablement team might identify where an individual needs additional support or where the same knowledge gap appears repeatedly across learners. Customer-facing teams can use the same approach to understand individual issues while also identifying recurring friction across many interactions. 

The common thread is that the conversation becomes more than a moment of engagement. It becomes something the organization can learn from and act on.