Parloa’s recent releases bring frictionless AI agent experiences

Customer conversations are evolving fast, and expectations for AI-powered interactions are rising just as quickly. And at Parloa, we’re focused on making AI agents not only functional, but natural to interact with. 

That’s why every recent release in our AI agent management platform reflects three priorities: more intuitive conversations, deeper operational transparency, and stronger enterprise technology for the teams behind them. As a result, CX leaders can deliver faster, more natural conversations while gaining the visibility and controls needed to run AI agents reliably at scale.

Here’s a roundup of the most impactful updates now live. And if you want to jump straight into how you can make the most of these features, check  out our documentation.

First impressions that adapt in real time

First impressions matter. With dynamic welcome messages, you can customize greetings at the start of every conversation. Agents can adapt the message right after a call or chat begins, drawing on past customer actions to set the right context from the first “hello.”

Better answers with knowledge skill improvements

We’ve upgraded our knowledge skill with a new embedding model that delivers significant benchmark improvements. This update enhances understanding of complex queries, technical terminology, and sophisticated content relationships.

Plus, advanced chunking strategies now break documents into segments while preserving semantic/markdown structure, ensuring more coherent and complete answers. 

Conversations that flow naturally with barge-in

Conversations with AI agents are now more fluid. With barge-in, callers can interrupt an agent mid-response and redirect the exchange. This makes support interactions faster, frictionless, and more natural.

Deeper conversation visibility with Data Hub

Data Hub brings secure, event-level interaction data directly into your analytics stack. Implemented via the open-source Delta Sharing protocol, it integrates cleanly with Tableau, Power BI, Looker, and BigQuery. Out of the box, Data Hub includes:

• Event-level data from every AI agent-customer conversation

• Built-in PII redaction

• Structured records optimized for high-volume queries (partitioned by tenant and event month)

• Automatic refreshes every 2 hours

This enables teams to correlate agent performance with CX outcomes, spot regressions quickly, and create dashboards using the BI tools you already rely on.

Smarter closeouts with improved hangup automation

Hangups are now more intelligent and flexible. There are two types of automation:

• Hangup events: Trigger backend processes such as ticket creation or case closure when a call disconnects.

• Hangup actions: Allow AI agents to proactively end calls when appropriate, preventing unnecessary time on the line.

The result is a smoother close to conversations and reduced manual workload.

Seamless connectivity with SIP Forward routing

We’ve expanded telephony integrations to support SIP Forward, allowing seamless routing of calls across a broader range of enterprise systems. This improves compatibility with legacy telephony setups and ensures workflows remain intact without additional engineering.

Personalization starts with contextual data

Conversations now start with more intelligence. AI agents can fetch and use relevant customer data at the beginning of the interaction, eliminating repetition, reducing friction, and creating more personalized experiences from the very first turn.

Handoffs that keep context with extended data

Handoffs between systems are smoother with structured data routing. AI agents can now store and forward key conversation outcomes to CRMs and support platforms. This ensures context carries over, reducing friction and enabling faster resolution when human agents step in.

Speedy fixes with agent debugging

Debugging agents is faster with a new user-friendly interface that provides clear, structured insights. This reduces troubleshooting time, simplifies onboarding for new developers, and helps teams resolve issues more efficiently.

Clear, consistent pronunciation

Voice interactions get a precision upgrade. With pronunciation control, you can fine-tune how names, brand terms, codes, or IDs are read aloud directly within the platform. Built-in lexicons and smart voice parameters reduce manual adjustments while ensuring voice consistency across experiences. The result is smoother customer interactions that sound natural, professional, and unmistakably on-brand.

Automated failover for peace of mind

Conversations shouldn’t grind to a halt just because a provider is having a rough day. That’s why we’ve introduced automated failover: a safety net that ensures your customer interactions keep running smoothly, even if one cloud provider stumbles.

Here’s what’s new:

• Two infrastructure options, one seamless experience: We now support both Azure and OpenAI for the same models.

• Built-in redundancy:  If performance dips on one side, Parloa automatically reroutes traffic to the other—no manual intervention needed.

• Secure and compliant: OpenAI is now an official sub-processor for select services, and everything remains compliant with the highest standards.

The result is higher availability, stronger resilience, and fewer headaches for your team.

The road ahead

Each of these releases strengthens our platform’s across usability, reliability, and analytics. Together, they help you deliver AI-powered conversations that build seamless, meaningful relationships with your customers. The next wave of improvements will go deeper into adaptability and enterprise-grade controls—so you can trust your agents to handle more, while your teams stay in command.

Full list of recent releases and documentation