SciTransfer
BISON · Project

AI-Powered Speech Analytics That Turns Contact Center Calls Into Business Intelligence

digitalPilotedTRL 7

Imagine every phone call your contact center handles is a goldmine of information — but right now, someone has to listen to each one manually. BISON built a system that automatically listens to calls in 9 languages, figures out what's being said, who's speaking, and what it means for your business numbers. Think of it like having a super-fast analyst who never sleeps, turning thousands of daily conversations into actionable reports on agent performance, customer satisfaction, and sales outcomes.

By the numbers
35,000
contact centers in Europe targeted by this technology
3.2 million
jobs generated by European contact centers
~1%
of Europe's active population employed in contact centers
9
languages supported by speech mining technologies
8
consortium partners with complementary skills
2
working prototypes delivered (smallBison and bigBison)
The business problem

What needed solving

European contact centers produce massive volumes of multilingual call data every day, but most of it goes unanalyzed. Supervisors can only manually review a tiny fraction of calls, meaning critical insights about customer behavior, agent performance, and business outcomes are lost. With 35,000 contact centers and 3.2 million jobs across Europe, the gap between data generated and data actually used represents an enormous missed opportunity.

The solution

What was built

Two working prototypes: smallBison — an initial system with speech recognition, speaker verification, language identification, and voice activity detection in 9 languages, deployed for real user feedback; and bigBison — the final open-environment platform with full capabilities, noise-robust optimized speech technologies, business outcome mining, and complete integration with contact center infrastructure.

Audience

Who needs this

Large BPO companies running multilingual contact centers across EuropeTelecom operators with high inbound call volumes needing automated quality monitoringBanks and insurance companies with regulatory call recording and monitoring requirementsContact center technology vendors looking to add speech analytics to their product stackEnterprise customer service departments wanting to extract KPIs from call data automatically
Business applications

Who can put this to work

Customer Service Outsourcing
enterprise
Target: BPO companies running multilingual contact centers

If you are a business process outsourcer handling calls across European markets — BISON developed a speech mining system covering 9 languages that automatically extracts Key Performance Indicators from live calls. Instead of supervisors manually sampling a fraction of conversations, the system processes all calls and flags issues in real time. Two large contact center operators in the consortium (EBOS and Atento) were involved as end users and deployed prototypes in real scenarios.

Telecommunications
enterprise
Target: Telecom providers with large inbound call volumes

If you are a telecom company dealing with millions of customer calls and struggling to extract meaning from them — BISON built speech recognition, speaker verification, and keyword spotting tools that work across 9 languages with noise-robust performance. The system translates raw call data into business outcome metrics, helping you understand why customers call and what drives resolution. Telefónica I+D and Telefónica Móviles were direct partners in the project.

Financial Services
mid-size
Target: Banks and insurers with compliance-heavy call recording requirements

If you are a financial institution required to record and monitor customer calls for regulatory compliance — BISON developed an open environment that integrates with existing contact center infrastructure and includes language identification and voice activity detection. The University of Bologna specifically addressed legal and regulatory aspects of processing call data. The bigBison prototype supports stand-alone or third-party integration configurations.

Frequently asked

Quick answers

What would it cost to implement this speech analytics system?

Pricing details are not available in the project data. The coordinator Phonexia SRO is a Czech SME that develops and markets speech technology products, so commercial licensing or SaaS pricing would need to be discussed directly with them. The system was designed as an open environment supporting both stand-alone and integrated deployment.

Can this handle our call volume at industrial scale?

The bigBison prototype was designed for full integration with contact center hardware and software infrastructure and was tested with real operational data. The project targeted the European contact center market of 35,000 centers generating 3.2 million jobs, indicating the system was built with large-scale operations in mind. Noise-robust and optimized versions of all speech technologies were delivered.

Who owns the IP and how is it licensed?

The project consortium includes 6 industry partners and 2 universities across 5 countries. The coordinator Phonexia SRO already places speech technology products on the market, suggesting they hold commercial rights to core technologies. Specific licensing terms would need to be negotiated with the consortium members.

What languages does the system support?

The system supports speech mining technologies in 9 languages, covering the operational needs of end-user partners in the consortium. Technologies include speech recognition, speaker verification, language identification, and voice activity detection across all supported languages. The system was also designed for rapid and cost-effective adaptation to new languages.

How does this integrate with our existing contact center setup?

The bigBison prototype was specifically built as an open environment that works in stand-alone configuration or integrates with third-party infrastructure. MyForce, a skilled contact center hardware and software integrator, was a consortium partner responsible for integration work. This means the system was designed from the start to fit into existing operations.

Is this compliant with data protection regulations?

The University of Bologna was included in the consortium specifically for its experience in regulatory and legal aspects of processing contact center data. The project recognized that call data involves significant legal issues and built compliance considerations into the system design. Specific GDPR alignment would need to be verified given the project ended in 2017.

How long would deployment take?

The project produced two prototypes: smallBison after the first year for initial deployment and feedback, and bigBison at project end with full capabilities. Based on available project data, the system was designed for relatively rapid deployment since end-user partners EBOS and Atento were ready to deploy prototypes immediately in real scenarios.

Consortium

Who built it

This is a commercially-oriented consortium with 75% industry participation — 6 out of 8 partners are companies, including 3 SMEs. The coordinator Phonexia (Czech Republic) already sells speech technology products, giving the project a clear path to market. Two large contact center operators (EBOS and Atento) served as end users who defined requirements and tested prototypes in real operations. Telefónica contributed both R&D expertise and business outcome mining capability. MyForce handled integration with existing CC infrastructure. The two universities (Brno University of Technology for speech tech, University of Bologna for legal/regulatory) provided targeted academic support rather than driving the project. Spread across 5 countries (Belgium, Czech Republic, Spain, Italy, Luxembourg), this consortium was built to commercialize, not just research.

How to reach the team

Phonexia SRO (Czech Republic) — an SME specializing in speech technologies with existing commercial products

Next steps

Talk to the team behind this work.

Want to explore how BISON's multilingual speech analytics could work for your contact center operations? SciTransfer can connect you with the right consortium partner. Contact us for a one-page technology brief.