AI Lab: Digital Dictation for Global Law Firm

Kerv Transform AI Labs is a hands-on innovation programme where we design and deliver tailored AI and automation solutions that solve real business challenges. Whether it’s streamlining internal processes, enhancing user experiences, or surfacing smarter insights, the AI Labs bring your ideas to life. We work collaboratively to prototype, refine, and deploy solutions that fit seamlessly into your existing ways of working.

A global law firm, with a hybrid workforce and rising caseloads, needed to modernise a core part of its daily workflow: legal dictation. The goal was to move away from legacy systems and toward a smarter, AI-powered platform that aligned with its Microsoft-first strategy.

The firm’s previous dictation platform had become a blocker to productivity. It operated as a standalone fat client, lacked integration with case management systems, and came with rigid licensing costs and infrastructure demands. New user onboarding was manual and slow, and the system offered no room for innovation or AI enhancements.

The firm wanted a future-ready alternative that would:

• Integrate seamlessly into their day-to-day legal workflows
• Leverage cloud-native services to reduce operational overhead
• Enhance transcription accuracy and formatting using AI
• Improve security, compliance, and monitoring
• Reduce ongoing infrastructure and licensing costs

Through the Kerv Transform AI Labs programme, we designed and delivered a bespoke digital dictation platform built directly into Microsoft Teams. This wasn’t just a modernisation of an old system, it was a reimagining of how dictation should work in a legal environment.

Key features of the solution included:

Native Teams integration:
The dictation tool was built directly within Teams, delivering a familiar, frictionless experience across desktop and mobile. This ensured fast adoption with minimal training.

Whisper AI for legal-specific transcription:
A key component of the solution is Whisper AI, an advanced speech recognition model that delivers industry-leading transcription accuracy. It detects and formats punctuation intelligently, understands legal-specific terminology, and adapts to different speaking styles.

Secure, scalable architecture:
Built on a cloud-native, PaaS-based Azure foundation, the solution delivers on performance and security. It provides resilient compute, reliable storage, and visibility through integrated monitoring, all with reduced management overhead and alignment to the firm’s Microsoft environment.

Automated onboarding:
Using Entra ID and role-based attributes, user access is fully automated and centrally managed, eliminating repetitive admin work for the IT team.

The move to this cloud-native, AI-enhanced platform had a transformative effect on the firm’s legal operations:

• Dictation is now embedded within the day-to-day workflow, no switching tools, no uploading files. Everything happens within Teams.
• Transcriptions are faster and more accurate, with Whisper AI ensuring dictations are correctly punctuated, structured, and legally sound before they even reach support staff.
• Manual onboarding has been eliminated, saving valuable time and reducing errors.
• Matter integration is seamless, with dictations correctly linked to the appropriate case file every time.
• New features and enhancements are deployed rapidly, based on evolving business needs. No more waiting on slow vendor roadmaps.
• Significant cost savings have been achieved through the removal of infrastructure and rigid licensing models.
• And most importantly, lawyers and support staff are saving time and working more efficiently, with tools that are built around how they actually work.

Kerv’s unique blend of Microsoft expertise and legal sector insight meant we could move fast, co-creating a solution that delivered measurable impact from day one. Our AI Lab approach brought together architects, legal ops, and end users to design a platform that solved real problems, reduced friction, and opened the door to continuous innovation.