20 November 2025

For all the hype surrounding conversational AI, real enterprise adoption is still moving cautiously. The technology is advancing fast, but companies remain wary about handing sensitive customer interactions to autonomous systems. Leaders worry about unpredictable behaviour, opaque decision-making, and an AI damaging hard-earned customer relationships. In short: the biggest barrier to enterprise AI adoption continues to be trust.

This tension sits at the heart of GetVocal.ai, the Paris-based startup that has just closed its Series A funding round to scale what it calls “transparent, governed AI agents.” Led by CEO and cofounder Roy Moussa, the company argues that the future of AI in customer experience (CX) won’t be defined by the most autonomous agents, but by the most trustworthy and controllable ones.

“In today’s systems, nothing can guarantee that a probabilistic LLM will behave as expected,” Moussa tells MoveTheNeedle.news. “And when AI agents go bad, no insurance company will cover the fallout.”

For enterprises adopting AI, especially in regulated sectors, predictability is not optional. “Whether it’s a human or an AI speaking for your brand,” he adds, “you can only delegate tasks and scale operations when there is trust. And trust is the product of both control and transparency.”

This philosophy underpins GetVocal’s approach — and likely is the core reason investors are betting on the company.

---

From Black Boxes to “Glass Boxes”: A New Model for Transparent AI

One of the biggest concerns in deploying AI agents at scale is the “black-box problem.” Most large language model (LLM)–powered systems make decisions internally in ways neither users nor vendors can fully explain. For enterprises operating in finance, insurance, mobility, utilities or healthcare, that opacity introduces unacceptable operational and compliance risk.

GetVocal proposes an alternative: a “glass-box” AI architecture where all reasoning is transparent, rule-based, and fully governed by enterprise logic.

Instead of relying on the LLM to determine actions or interpret intent, GetVocal uses conversational graph technology built from the customer’s existing processes, protocols and compliance rules. The AI’s reasoning path is deterministic and visible, while the LLM is used exclusively for what it excels at — conversational generation, tone control and personalisation.

“It’s not about improving a black-box solution,” Moussa says. “Those decisions are still black-box decisions — or grey at best. We’ve built a system where the reasoning is completely transparent.”

This allows GetVocal to deliver two things enterprises rarely get from conversational AI platforms today:

1. Full decision traceability

Every step in a customer conversation follows the enterprise’s own logic, not the LLM’s improvisation.

2. Real-time human intervention

Managers can observe conversations live, validate decisions, halt a process, or seamlessly transfer the call to a human agent. Every action is logged and auditable — essential for industries with strict compliance requirements.

In the era of AI governance and trust & safety, this level of transparency positions GetVocal as one of the most enterprise-aligned players in the European AI ecosystem.

---

Human Oversight by Design — Not as an Afterthought

Most AI vendors talk about human oversight, but treat it as an add-on: dashboards, alerts, or short-term monitoring. GetVocal treats it as foundational.

“Human oversight is baked into the platform,” Moussa says. “Not just a feature added on.”

The platform allows enterprises to configure escalating workflows, human validation checkpoints, automatic handovers and flexible guardrails — all based on their own risk appetite and operational design.

The philosophy is simple:
AI should behave like a junior employee — escalating when unsure, learning from humans, and improving over time.

This aligns GetVocal with a growing industry shift from:

“humans assisted by AI” → “AI powered by humans”

Moussa notes that the boundary between human agents and AI agents differs widely across industries and geographies. Even within finance, a UK insurer and a US bank may have very different expectations of what AI should or shouldn’t do.

“With GetVocal, businesses define how deterministic, generative or creative they want the protocol to be,” he says. “They draw the borders of AI wherever they feel comfortable.”

In an era where customer expectations are rising and contact centres face labour shortages, flexible hybrid workforce design — with human-AI collaboration at its core — is becoming a competitive advantage.

---

What the Series A Funding Fuels

GetVocal’s Series A is not a “grow at all costs” raise. Moussa emphasises operational discipline.

“This funding ensures we can continue growing while shipping the right product for our customers,” he says.

The capital will be split evenly between:

• Product development, strengthening the governed AI engine and expanding enterprise features

• Commercial expansion, with new hires in sales, partnerships and customer success

This balanced strategy reflects a shift happening across the industry: enterprise buyers want long-term AI partners, not quick-deploy tools. With regulatory expectations rising in both Europe and the U.S., trust-first platforms like GetVocal are gaining traction.

---

Europe as the Perfect Testbed — With the U.S. in Sight

Europe’s AI market, shaped by GDPR, the AI Act and strong consumer-protection expectations, is becoming a proving ground for governed AI systems. GetVocal’s traction mirrors this environment.

The company sees major expansion opportunities across:

• DACH

• Benelux

• Nordics

• UK

• France

• Southern Europe, including Iberia.

These are markets with sophisticated CX operations and strict expectations around governance and transparency — exactly where GetVocal’s value proposition resonates.

But the company has global ambitions.

“Long term, we want to take the product to the U.S.,” Moussa says. “Having built the platform with the complexity of Europe in mind, we’re very confident we can tackle the U.S. market too.”

---

The Future of Customer Experience: A Hybrid Workforce That Actually Works

Moussa believes the contact centre as we know it won’t look the same in five years.

“We’re about to witness a major paradigm shift,” he says. “Call centres-as-a-service will not exist in the same way. Job roles are fast changing, and they urgently need a platform that can handle hybrid workforces.”

In GetVocal’s vision:

• AI agents handle predictable workflows reliably

• Human agents step in for complex or emotional tasks

• Enterprises maintain full governance, auditability and control

• AI becomes a trusted operational asset, not an unpredictable risk.

“GetVocal will be the next-generation CX platform,” Moussa states. “We’re accompanying enterprises on their transformation journeys, helping them gain efficiencies while keeping human oversight where it’s needed.”

So claims processing, onboarding, renewals, support, compliance workflows — almost any language-based function — could be rebuilt with transparent, fully governed conversational AI.

---

A Closing Thought: Trust Is the Real AI Innovation

In a year where dozens of startups are touting “AI agents,” GetVocal stands out by tackling the problem enterprises actually worry about: risk, transparency and trust.

“You can only delegate tasks and scale your operations when there is trust,” Moussa reminds us.

If the enterprise AI wave is truly going to scale — beyond pilots, beyond demos, beyond buzzwords — it will likely be built on systems like the one GetVocal is betting on: governed, explainable, controllable and designed to work alongside humans rather than replace them.