Accountancy

Armstrong Watson - Building a Firm-Wide AI Adoption Programme

How Armstrong Watson moved from Copilot licences to firm-wide adoption: discovery, 84 hands-on sessions, targeted Premium learning and an AI adoption framework.

97%
improved their understanding of Copilot and AI
99%
intend to experiment with Copilot after training
95%
improved their understanding of Copilot and AI

"I felt fully engaged for the 2 hours and found it practical, supportive and informative. My attention span is not great, so this is quite the compliment!"

"I felt fully engaged for the 2 hours and found it practical, supportive and informative. My attention span is not great, so this is quite the compliment!"

Armstrong Watson is one of the largest independent chartered accountancy firms in the north of England, with around 900 staff across multiple offices. The firm had already invested in Microsoft 365 and Copilot licences and could see the potential of AI to improve productivity and access to knowledge across the firm. But confidence and experience varied widely. Some staff had experimented with Copilot on their own initiative while others had never opened it. There was no shared baseline of understanding, no consistent approach to using AI and no formal programme to support adoption.

Leadership, led by Chief Executive Paul Dickson, recognised that AI adoption was a behaviour and culture change as much as a technology rollout. The firm needed a structured programme that would build organisation-wide confidence, align with its existing AI policy and professional standards and create momentum for deeper, role-specific adoption later. It also needed evidence to inform decisions on Copilot Premium, rather than guessing where the licences would add most value.

Armstrong Watson's programme follows the four stages of our Copilot Adoption Programme. The diagnostic and foundational work was delivered once, across the whole firm. Adoption and embedding now continue as an ongoing cycle, team by team.

Stage 1: Diagnostic and alignment

We started with a two-hour discovery session with senior stakeholders, including Paul Dickson, IT leadership and learning and development. Together we mapped how Microsoft 365 and Copilot were being used across the firm, reviewed document management and communication practices and shaped the content around Armstrong Watson's workflows, AI policy and compliance requirements.

A pilot Copilot Essentials session with seven senior and technical users then tested the content through hands-on exercises in Copilot Chat, Excel, Word and prompt design. Their feedback refined the programme before it reached the wider firm, so everyone received a tested programme rather than a generic one.

Stage 2: Foundational training

The core rollout was 84 scheduled training sessions over six weeks, from 30 March to 8 May 2026, with catch-up sessions for anyone who missed their original slot. Sessions ran on Microsoft Teams in two-hour interactive formats, capped at around 10 participants so they stayed hands-on and conversational rather than passive.

Each session covered:

  • How AI and Microsoft Copilot fit into the Microsoft 365 suite
  • Armstrong Watson's internal AI policy, security and data handling expectations
  • Hands-on prompting across Copilot Chat, Word and Excel using realistic scenarios
  • Prompt design and best practice, including the Prompt Coach feature
  • Safe and responsible use, reinforced throughout

Every attendee received a custom training guide branded for Armstrong Watson, recordings of both parts of every session, daily prompt emails to build the habit and a direct link to the firm's AI policy. We managed the logistics too: a custom availability form, automatic session allocation, central attendance tracking and email support throughout the programme.

Stage 3: Adoption and targeted development

With a shared foundation in place, the focus moved to specific teams and roles. We delivered a Copilot Premium showcase to senior stakeholders, covering the practical differences between standard and Premium Copilot and demonstrating agents including Researcher, Analyst, Word, PowerPoint and Excel. This gave leadership what they needed to make informed decisions on Premium licence allocation.

Departmental discovery followed, looking at how individual teams work and where Copilot could genuinely help. For the Client Advisory team, we developed Copilot Premium learning built around realistic advisory activities, with supporting materials designed for those scenarios and professional judgement and human review built into the learning.

Stage 4: Embedding and scaling

To sustain momentum beyond training, we developed an AI Adoption Framework for the firm. It sets out:

  • A Senior Copilot and M365 Adoption Group to steer adoption across the firm
  • Departmental adoption groups, supported by department champions
  • Clear criteria and a decision process for allocating Copilot Premium licences
  • A forward plan for Copilot Premium, Microsoft 365 apps and Copilot agents training

Across 109 evaluation responses, 97% of staff said the training improved their understanding of Copilot and AI, 99% said they intended to experiment with Copilot and 95% felt able to apply what they learned to their day-to-day work. 97% rated the training as good and worthwhile and 99% rated the trainer as competent.

In practice, that meant a shared baseline of confidence across the firm. Staff who had been hesitant about AI had a clear, practical starting point. Those who had been experimenting on their own had a framework to work within. Leadership had the evidence to decide where Copilot Premium should go next.

Adoption has kept building since the rollout finished. Copilot Fridays and local office-led sessions are running, with staff sharing real use cases and encouraging a culture of experimentation. A Workflow Improvement Team has been set up, with department champions driving adoption within their own teams. A selective Copilot Premium rollout is underway, informed by training evaluation data and usage analytics, and departmental discovery continues to identify where Premium and automation can add most value. The firm has also approved the use of Anthropic (Claude) models alongside GPT within its Copilot environment.

Key benefits

  • Organisation-wide AI awareness and a shared baseline of Copilot confidence
  • A structured framework for long-term AI adoption
  • Evidence-led decisions on Copilot Premium licences
  • Advanced use cases identified team by team
  • A foundation for ongoing improvement and future automation

If you are planning structured Copilot adoption across a professional services firm, see our Copilot Adoption Programme, Copilot Essentials training and AI Implementation Partnership. We also deliver sector-specific Copilot training for accountancy firms.

Ready to move from Copilot access to Copilot adoption?

Book a free consultation to talk through where your organisation is with Copilot and what the right next step looks like. No pressure, no hard sell. Just an honest conversation.