# Practical Takeaways from the William Fry AI Summit 2025
Source: https://humanspark.ai/practical-takeaways-from-the-william-fry-ai-summit-2025/

Research shows generative AI adoption in Ireland more than doubled - from 49% to 91% - in just 12 months. Here's what smart businesses are doing next.

That kind of growth doesn't happen by chance. It signals a turning point: AI has moved from an emerging technology to an operational must-have.

![Close-up of a laptop and printed speaker agenda at a tech business conference, with attendees and stage in the background](https://humanspark.ai/wp-content/uploads/2025/05/IMG_6913-1024x768.jpeg)

At William Fry's 2025 AI Summit in Dublin, leaders from Microsoft and William Fry shared what's working in real-world business environments.

## AI is Now a Business Essential

**Irish organisations have gone from cautious to committed.**

In just a year, enterprise AI usage has exploded. Businesses that were once testing tools are now embedding them into daily operations.

Why this matters:

**Falling behind is now a real risk**
 **AI is driving measurable efficiency gains**
 **The tools are ready - the challenge is implementation**

## The Benefits Are Tangible

Firms adopting enterprise AI are seeing clear returns:

- Saving 2-4 hours per employee per week
- Shifting the workload from admin-heavy to high-value
- Improving consistency and cutting errors
- Speeding up work without sacrificing quality

Those time savings add up fast across large teams. And the work people do feels more meaningful when repetitive tasks are off their plates.

## What Works in Real Implementation

**A phased rollout beats big bang every time.**
 Here's the approach that's proving successful in large organisations:

### 1. Start Small and Targeted

- Pick clear use cases with measurable outcomes
- Run cross-functional pilot projects
- Track time savings and error rates
- Use early wins to build internal support

### 2. Scale With Intention

- Focus on areas with clear ROI
- Build internal champions across departments
- Create feedback loops for ongoing improvement
- Set up a centre of excellence to share knowledge

## High-Impact Use Cases

These are the areas where AI is already making a difference:

**Document drafting and review**
 Improves consistency, reduces editing time, and frees teams to focus on substance.

**Workflow automation**
 From email generation to onboarding follow-ups, AI is simplifying time-heavy processes.

**Knowledge access**
 AI-powered search cuts onboarding time and preserves institutional knowledge.

**This is important because** each of these use cases connects directly to business outcomes - time saved, errors reduced, and faster decision-making.

## Challenges You'll Need to Tackle

### 1. Closing the Skills Gap

Many teams aren't AI-ready, but you don't need experts everywhere.
 What works:

- Tiered AI training for different roles
- Cross-functional implementation teams
- Clear, practical learning - not just theory
- Support from external partners early on

(If you need help with [AI training](https://humanspark.ai/training/), reach out)

### 2. Managing Shadow AI

If you lock tools down too tightly, people will work around you.
 That opens up real data security risks.

Instead:

- Set permissive, structured AI policies
- Use enterprise-grade tools with proper controls
- Make security and compliance non-negotiable

### 3. Laying the Data Groundwork

AI needs clean, accessible data.
 That means:

- Running a data inventory
- Mapping sensitive data
- Tightening access controls
- Monitoring how data flows in your systems

**This groundwork isn't optional.** It's the backbone of safe, effective AI use.

## Legal & Compliance: What to Know Now

**Regulators are paying attention.**
 Organisations need to show:

- Documented AI training for staff
- Protocols to verify AI-generated content
- Transparent handling of personal data
- EU-based data processing where possible

**This is important because** compliance won't wait for you to catch up. Build it into your implementation plan from day one.

## What's Next: Trends to Watch

### Agentic AI

AI agents that handle multi-step tasks across systems are on the horizon. Think research, procurement, or legal analysis - automated end-to-end.

### Stronger Enterprise Commitments

Big providers are now offering:

- Local cloud infrastructure
- Resilience and uptime guarantees
- Clearer privacy and data safeguards
- Better support for open-source tooling

## Implementation Roadmap: A Strategic 90-Day Action Plan

Rolling out AI successfully takes more than just enthusiasm and tools - it needs a clear structure and steady pace. This 90-day plan is based on the framework I share in the "AI Adoption Roadmap" (part of the full AI Toolkit at [humanspark.ai/toolkit](https://humanspark.ai/toolkit)). It gives you a practical way to start small, build momentum, and scale with confidence.

### Days 1–30: Assessment & Preparation

- **Run a team survey** Capture where people stand on AI - what excites them, what worries them, and what they're already trying. It'll highlight quick wins and deeper gaps.
- **Pick your AI champions** Find curious, respected people across departments who can lead by example. They'll be key to encouraging adoption from the ground up.
- **Deliver hands-on training** Give your champions practical, no-jargon sessions on AI basics, ethical use, and useful applications. Focus on real work problems, not just theory.

### Days 31–60: Building Momentum

- **Set up an AI working group** Bring together IT, ops, legal, HR, and leadership. This group will set policy, track progress, and help avoid siloed efforts.
- **Create an AI sandbox** Let teams explore AI tools safely using synthetic or non-sensitive data. Keep it separate from live systems, with clear guidelines for testing.
- **Launch quick-win projects** Start 2–3 small, low-risk pilots with obvious benefits - like time saved or tasks automated. These early wins are crucial for buy-in.

### Days 61–90: Expanding Capability

- **Draft usage policies** Define how AI should (and shouldn't) be used. Balance control with flexibility. Make sure your policies support innovation, not block it.
- **Expand smartly** Roll out the best pilots to more teams - but don't skip the governance. Use what you've learned to improve the next phase.
- **Capture lessons learned** Start documenting insights, risks, and best practices. A shared knowledge base helps teams avoid reinventing the wheel.

### Beyond 90 Days: Strategic Transformation

- **Scale with purpose** Focus on use cases aligned with business goals. Don't chase every trend - build on what's working.
- **Strengthen your foundations** Invest in scalable data systems and clear governance. These are the guardrails that'll support long-term growth.
- **Embed it culturally** Make AI part of how you work and make decisions. It's not just a tool - it's a shift in how teams think and operate.

This roadmap supports responsible AI rollout that delivers real business value - while also improving how teams work.

**For more detailed guidance and templates, you can find the full toolkit at [humanspark.ai/toolkit](https://humanspark.ai/toolkit).**

## Success Factors: What Makes AI Stick

1. **Start with the data** Poor data equals poor AI performance.
2. **Engage your frontline** They know what's slowing them down.
3. **Measure real outcomes** Focus on time saved or errors reduced - not just "adoption."
4. **Get outside help early** Internal teams don't need to figure it all out alone.
5. **Build a community inside** Champions across departments drive sustained use.

## Pitfalls to Avoid

- Endless pilots with no follow-through
- Letting the tech lead instead of the business case
- Training that's too abstract or too light
- Treating governance as an afterthought
- Keeping AI siloed in IT or innovation teams

## Final Thought

AI is no longer about the future. It's about your next 90 days.
 The best implementations start by solving real business problems - not by chasing shiny tools.

Get the basics right. Involve the right people. Measure what matters.

**That's how you make AI deliver for your business.**

**What use case could AI help with in your team?**
 I'd love to hear where you're starting - or where you're stuck. [I posted about this on LinkedIn](https://www.linkedin.com/in/alastairmcdermott/) - please leave your thoughts on the post there, and [I'd love to connect there](https://www.linkedin.com/in/alastairmcdermott/).

*PS: Thanks to the William Fry team for a really practical look at how enterprise AI is evolving.*