How to Go from AI Skeptic to AI Champion in Your Organisation

Struggling to get your team onboard with AI? This article breaks down 5 clear steps to go from resistance to results with practical, proven strategies to build trust, show ROI, and scale confidently.

AI adoption is on the agenda for many organisations, but actual implementation often stalls. The reasons are familiar: unclear return on investment, internal resistance, and a sense that it’s just too complex to get right.

These concerns are valid. But they don’t have to stop progress.

With the right starting point, teams can move from hesitation to momentum - without a major budget or in-house AI team. The most effective approach we’ve seen is not to push technology, but to focus on solving real business problems, quickly and visibly.

We’ve seen this play out across dozens of teams. That’s why we developed ForgeWorks, a delivery framework to help clients adopt AI in practical, measurable ways. More on that later.

The Real Barriers to AI Adoption

AI adoption isn’t failing because of a lack of tools. It’s failing because teams are misaligned, internal trust is low, and delivery processes aren’t built for scale.

Surveys from both Writer.com and Fast Comnpany highlight the growing gap between intention and impact:

• 66% of executives say AI adoption has caused tention and divide inside their company
• 42% believe that AI is actively tearing the company apart
• 71% of C-suite leaders admit AI is being built in silos across their organisation
• 31% of employees (including 41% of Gen Z) are intentionally sabotaging AI efforts, from refusing to use AI tools to undermining their output

Despite these challenges, 90% of the workers and executives remain optimistic about their company’s approach to AI.

The companies seeing meaningful outcomes share a few things in common: they have clear organisation-wide strategies, they invest heavily to achieve higher ROI, they activate internal AI champion, and they work with vendors who help shape the long-term vision.

And it pays off.

94% of employees who become AI champions report seeing a career benefit.

In this article, we unpack what an AI champion can actually do to push the boundaries of better AI adoption - and how that momentum leads to results for the business and growth for the individual.

Five Steps to Build Trust, Deliver Value, and Scale Responsibly

1. Start with a Real Problem

It’s tempting to begin with the tech. But we’ve found the better starting point is to identify a problem that’s already slowing things down.

A company we worked with wanted to explore generative AI. We asked what problem they wanted to solve. Their operations team was spending 6 - 8 hours a week updating new client engagement contracts manually. Within two weeks, we deployed a templated generation model that automated 80% of that task. It wasn’t complex, but it saved them time and kept their clients happier.

Similarly, a client in the auditing industry flagged delays in their document review process. We used an existing model to pre-classify sections of incoming PDFs, speeding up QA checks and freeing up senior staff to focus on risk review.

The point is to anchor AI efforts in a pain point your team already recognises. That creates momentum without needing to sell the idea of AI at all.

Bring your team in early

When people see how a problem affects them and how a solution helps, they’re more likely to support the change. We always involve users in defining success and reviewing early results - before anything is scaled.

2. Choose a Low-Risk, High-Impact Use Case

You don’t need to launch a multi-year programme. Start with something narrow, testable, and visible.

One financial services team needed a faster way to assess financial statements submitted by SME clients. Rather than building a full credit risk engine, they started with a 2-week pilot using an AI model to extract and validate key financial fields (like revenue, profit, and net assets) from uploaded PDFs.

The pilot covered only a subset of clients, but it cut review time from 45 minutes to under 10 minutes per document and flagged inconsistencies analysts had previously missed.

Because it required no changes to existing systems and delivered immediate time savings, the team viewed it as a low-risk test, and the outcome gave them confidence to expand AI across onboarding and compliance.

Good use cases tend to fall into three categories:

• Automating time-consuming admin work
• Enhancing a current process (like segmentation or triage)
• Supporting better decisions with more timely data

Define success and boundaries upfront

In these projects, we spend just as much time defining what not to do. That keeps the scope manageable and gives the pilot the best chance to succeed.

3. Use Existing Tools to Move Quickly

You don’t have to build from scratch to get started. Platforms like the Highwind AI Marketplace offer plug-and-play AI components that work out of the box and can be customised later if needed.

In one HR use case, recruiters were already receiving CVs via email. Instead of changing the workflow, we automated it. CVs were parsed directly from forwarded emails, key details extracted, and candidates ranked against the job spec, with results synced into a chatbot-powered ATS interface.

The setup uses tools that can interact easily with AI Agents, and the flow was live in less than a week.

Within days, 150+ CVs were processed - cutting manual admin by over 60% and freeing up recruiters to focus on interviewing.

Start small and adjust

Most tools allow for quick feedback cycles. Run a two-week pilot. Look at usage, performance, and user feedback. Improve from there.

4. Measure and Share the Outcomes

Internal support grows when people can see what’s working.

We typically track:

• Time saved (manual vs automated)
• Accuracy gains (fewer mistakes or rework)
• Cost impact (either hard savings or time reallocated)
• User sentiment (via quick surveys or informal feedback)

A mid-sized insurance team deployed a lightweight chatbot to handle inbound product queries. From Day 1 of testing with their brokers, they began tracking:

• Coverage rate (what % of queries could be answered),
• Response helpfulness (measured via upvotes/downvotes),
• and a few other important technical metrics for product enhancement

The team reviewed these metrics regularly and used early wins to justify expanding coverage to additional products. Clear and transparent metrics also helped to build trust, maintain momentum, getting leadership buy-in for further automation.

Make it visible

We recommend internal demos, short videos, or even Slack updates. Sharing progress helps others see the value and puts internal teams in the spotlight—this is critical for wider adoption later.

5. Scale What Works, With the Right Support

Once you’ve proven value in one area, the question becomes how to scale without losing quality.

We’ve seen the best outcomes when companies invest in internal capability - not just by hiring more data scientists, but by building repeatable habits, pairing technical and business people, and documenting what works.

This is the approach behind ForgeWorks, our delivery framework for scaling AI inside organisations. It’s not just about deploying more tools, it’s about building the rhythm to do it again, better, and more efficiently each time.

We focus on:

• Creating internal visibility through an AI portfolio or marketplace
• Embedding feedback loops between users and delivery teams
• Developing internal champions through guided, hands-on delivery
• Working in short, iterative sprints to test and improve solutions quickly

Want to Build Internal AI Capability?

ForgeWorks is our programme for teams that want to move from pilot projects to sustainable AI adoption.

We embed with your team to co-deliver practical AI solutions, while building internal skills and laying the groundwork for long-term capability. You bring the problem. We help your team solve it on your infrastructure, with your people, and in your context.

If you’re exploring your first AI project or looking to scale what’s already working, we’d be happy to share what we’ve seen work.

👉 Reach out at info@melio.ai

Or visit melio.ai/forgeworks