5 expert lessons on driving value with AI

What does it take to turn AI hype into real business value? At SoftwareOne Summit 2025, experts from AWS, Google Cloud, Microsoft, ServiceNow and SoftwareOne gave clear answers.

Here are five of the most important—and actionable—lessons we learned on the day.

1. Leverage the innovation engine you already own

Forget cost cutting as a necessary evil, today’s CIOs are discovering that IT portfolio optimization is an engine that drives innovation. Imagine launching your next AI initiative not with a plea for more budget, but with funds you’ve liberated from hidden inefficiencies, using disciplines like ITAM and FinOps.

This approach can deliver value with surprising speed, so long as an organization has a comprehensive picture of its spending. SoftwareOne’s Trent Allgood shared the story of a financial services client that for the first time extended their search for savings to FinOps.

They were seeing those savings immediately, and they were able to use those funds to start innovation projects.

– Trent Allgood, SoftwareOne

AI itself is changing where inefficiencies and risks lie and as Trent noted, “we now have not only shadow IT, shadow SaaS, now we have shadow AI”.

As a result, IT cost management disciplines are adapting at speed. ITAM and FinOps are converging to manage spend and risk holistically and AI is helping CIOs keep track of their estate. Today’s solutions are breaking down silos, according to Simon King from ServiceNow by “embedding AI capability to be agent to agent across workflows”.

With AI posing new challenges and opportunities, it’s never been more important to put measures in place to control cost and risk – and unearth fuel for innovation.

2. Democratize AI now, don’t wait for perfect data

Another theme from the summit is that the power of AI is multiplied when it’s accessible to everyone. The latest workplace solutions, like Google’s Gemini Enterprise or M365 Copilot, are intuitive and customizable so both super users and frontline staff can benefit from AI-driven insights and automation.

Furthermore, line-of-business experts outside IT are the ones most likely to find valuable and unique applications for AI.

If you think about deploying AI, think at scale. So put it in the hands of everyone in the organization, because that’s going to unlock some of the most exciting use cases.

– Daniël Rood, Google Cloud

Democratizing AI tools will not only boost productivity, it will also reduce risk. The use of unauthorized AI tools at work is booming, increasing the likelihood of sensitive data being leaked. Merely banning these apps is ineffective, as illustrated by SoftwareOne’s Kevin Bernstein in his story about a CIO secretly using a personal AI assistant at dinner.

The very definition of shadow AI…the thing that they are trying to block, they themselves are using.

– Kevin Bernstein, SoftwareOne

The lesson is that rules are easily ignored. To tackle risk, organizations must accelerate adoption of safe and compliant tools that match or exceed the capability of what’s freely available online.

And concerns about data maturity is no longer a reason for IT leaders to hold back. As Daniel explained, today’s models like Gemini are designed to work effectively with imperfect data, spanning images, text, code, and voice.

Don’t wait six months or longer before you deploy these models, because the outputs will be of high quality from day one.

– Daniël Rood, Google Cloud

The message is clear: start your AI journey now and let the technology drive value right away.

3. Have a clear vision and build strong foundations

Whether buying or building AI solutions, the same rules apply: while you shouldn’t wait for perfection, you should aim to build on solid ground.

AWS’s Simon Treacy issued a reminder that all successful AI initiatives start with a clear business vision, “If you don’t fully understand the problem you’re trying to solve, the tooling behind it... is almost immaterial.”

IT decision makers should therefore place greater emphasis on refining the idea behind the project and ensuring the business impact is clear, rather than fretting about technical details like model selection.

So in some ways, going after the right ideas, working backwards from your customers is more important, maybe than the exact model you're choosing today, which could potentially change 6 or 12 months down the line.

– Simon Treacy, AWS

The most successful organizations start with clear business outcomes, develop fit-for-purpose data, and invest in a flexible, secure platform. Starting out with AI envisioning workshops with expert partners like SoftwareOne and adopting a structured approach to delivery can help provide the foundations for scalable, sustainable AI.

Having the right principles in place from the start is key, and that extends to governance and cost control. Simon highlighted tools like AWS Bedrock’s guardrails, ensuring “you’re protecting your organization, your brand, and your reputation” from data security and compliance risks.

FinOps is also critical to ensuring initiatives achieve ROI, and regular reviews, tagging, and collaboration with legal and finance teams will mean innovation doesn’t come at the expense of compliance or runaway costs.

4. Security must move at machine speed

Another lesson is that while artificial intelligence is reshaping the threat landscape, making attacks more sophisticated and frequent, it is also empowering defenders. The integration of AI into security tooling—such as Microsoft’s XDR and Security Copilot—enables organizations to analyse vast volumes of signals and respond autonomously to incidents.

Attackers are using AI, but defenders can as well. Just responding at a human speed is no longer enough… it’s not possible without having AI in the loop to help respond at machine speed.

– Ben Hart, Microsoft

And while the threat landscape is exploding, organizations are urged to get the basics right. Zero Trust is now the cornerstone of any robust security strategy, especially in hybrid and multi-cloud environments.

Verifying explicitly, enforcing the principle of least privilege, and assuming breach at all times is absolutely key to ensuring that you’ve got a safe environment.

– Ben Hart, Microsoft

With AI’s growing role, data privacy and ethical considerations are more critical than ever, especially as some risks are hard to remediate.

Once personal information gets into an AI tool, it is nearly impossible to take it out

– Martin Roskelly, SoftwareOne.

Clear AI policies, user awareness, and robust controls are therefore crucial to prevent sensitive data from being exposed or misused. The future will demand regular auditing of AI activity and a culture where security is everyone’s responsibility.

5. AI is, in fact, a human story
— empower, educate, engage

Forget the hype about algorithms and automation — our final take away is that real transformation starts with people. As stated by Kevin, industry analysts now agree that “70 to 80% of this story is about people, not about the technology itself” and AI is no longer just an IT initiative; it’s a company-wide movement.

Collaboration between IT, HR, marketing, and every business unit is essential. The organizations that thrive will be those that bring people together in more meaningful ways, using AI to enhance—not replace—the human element.

Indeed, the best technology is worthless without enthusiastic users. SoftwareOne’s Alex Galbraith summed it up: “It doesn’t really add a lot of value if we build our amazing apartment block... and nobody wants to live in it.”

Real business value comes from solving problems that matter, engaging users early, and investing in change management. To succeed, put people at the heart of your AI transformation strategy.