Product release roundup: Scale globally, stay compliant, and access conversation data

As enterprises accelerate AI agent adoption, global growth presents a number of new challenges: managing agents across markets, languages, and channels without copy/paste chaos, meeting strict data residency requirements, universally accessing conversation data for analytics and compliance, and maintaining centralized visibility into configuration changes as teams scale. 

Our recent releases tackled these challenges head-on. From Agent Composition—our most significant release for large-scale, diverse CX footprints—to enhanced data sovereignty options and enterprise-grade conversation access, these updates help teams scale CX automation while maintaining control, compliance, and visibility.

Here's what's new.

Build once, deploy everywhere: Agent Composition

Managing AI agents across global markets shouldn't require maintaining dozens of duplicates. But that's exactly what happens when teams scale: each new market, language, or channel demands another version to deliver local or contextual nuance. A simple policy update becomes a manual slog across countless agent instances, with every edit introducing risk.

Agent Composition eliminates this duplication entirely. Build one composable, centralized agent, then dynamically adapt it across every region, language, and channel without rebuilding from scratch. When you need to update your refund policy or adjust store hours for a specific market, you make the change once. It flows automatically to every relevant deployment—whether that's your German phone line, Spanish web chat, or French mobile app.

This isn't just templating or cloning. Agent Composition is embedded in Parloa's core architecture, enabling agents to adapt contextually while maintaining a single source of truth. The result: global consistency without sacrificing local relevance.

What this unlocks for your team:

• Centralized control with local flexibility: Define your agent's core logic once, then customize greetings, compliance notices, or regional nuances through environment variables.

• Instant, synchronized updates: Change an instruction or policy in one place, and watch it propagate across your entire deployment network automatically.

• Unified brand experience: Customers get consistent service quality whether they're calling from Berlin or Barcelona, using voice or chat.

For enterprise teams managing high call volumes across multiple countries, this changes the economics of global automation. Less time synchronizing versions means faster launches, fewer errors, and lower operational overhead, while customers experience seamless service no matter how they reach you.

ElevenLabs voices with EU data processing

Voice quality matters, but so does compliance. With ElevenLabs EU Processing, enterprises can now adopt expressive, natural-sounding AI voices while meeting strict data residency requirements.

You choose where text-to-speech processing happens (EU or US) on a per-agent basis. For compliance-sensitive customers, this opens up new possibilities. A logistics company, for example, can now use AI agents to proactively call customers when deliveries need rescheduling—scaling outbound voice communications while keeping all processing within EU borders. No human agent required unless requested, and no compromise on voice quality or data sovereignty.

Parloa's Zero Retention Mode ensures no input or output data is stored or transferred, regardless of region. The result is GDPR-aligned voice automation that sounds natural and scales confidently.

Note: You will need to accept ElevenLabs as a subprocessor to enable EU processing.

Secure, real-time conversation access with Transcript API

Understanding what happens in AI agent conversations is critical for handovers, analytics, and compliance. The Transcript API provides secure, real-time access to complete conversation data in our Conversation Store.

Key capabilities include:

• Instant access: Retrieve full transcripts or summaries in seconds for seamless AI-to-human handovers. When a human agent needs to take over, they have complete context immediately.

• Granular retention controls: Define separate time-to-live periods for raw and anonymized data to stay aligned with GDPR requirements.

• Enterprise security: Tenant-level encryption, PII anonymization via PIIANO, and full audit traceability come standard.

This solves three critical problems: loss of context during handovers, delayed access to conversation data (replacing batch exports with real-time retrieval), and compliance risks through tenant-level retention controls.

The API supports retrieving full, anonymized, or summarized transcripts through a single endpoint, with AI-generated summaries available via the Conversation Insights tool. Performance is optimized for enterprise scale, with retrieval under one second even under load.

Complete visibility into configuration changes: Centralized Audit Logs

As AI agents scale across teams and environments, maintaining visibility into who changed what becomes critical for security, compliance, and troubleshooting. While audit logging existed previously, Centralized Audit Logs bring a more reliable, enterprise-ready experience to AMP with a unified, tenant-level record of every critical action in one place.

This tracks operationally and security-relevant actions: configuration updates, user invitations and role changes, tenant-level modifications, API client creation and updates, and resource creation or deletion. Each entry provides complete context, including the specific user or service account, the object affected, and detailed before-and-after states for configuration changes.

What this means for enterprise operations:

• Faster incident resolution: When an agent stops working as expected, correlate the issue directly to specific configuration changes instead of guessing what broke. No more manual investigation across scattered systems.

• Simplified compliance: Meet internal governance and regulatory requirements without compiling evidence manually. Audit Logs provide the structured, immutable record auditors require.

• Accountability across teams: In collaborative environments where multiple users manage live agents, clear attribution eliminates confusion about who changed what and reduces risk from unauthorized modifications.

Built for enterprise governance, Centralized Audit Logs give security teams and developers the operational transparency they need as AI deployments grow in complexity and scale.

Current customers can learn more in our documentation.

Looking ahead

Each of these releases reflects our commitment to building AI agents that scale intelligently across markets while giving teams the operational control and visibility they need. Agent Composition sets a new standard for global agent management, while ElevenLabs EU Processing and Transcript API ensure your voice and data strategies meet the highest standards for compliance and performance.

As we move into 2026, expect deeper capabilities around adaptability, governance, and insights, so your agents can handle more complexity while your teams maintain full command.