The Opinion
by Ricardo Martinho, President of IBM Portugal
Small steps, big wins: selecting the right use cases to succeed with AI
When generative AI became accessible to everyone, organizations felt pressured to act. Leadership teams wanted to demonstrate they were already using AI and quickly launched pilot projects without actually having a clear plan for how to derive value from those initiatives. Some of these pilots were successful, but many stalled because they were not aligned with business priorities and lacked a way to measure performance.
We are now witnessing a shift towards a more thoughtful approach, focusing on AI use cases that solve real problems, achieve measurable results, and lay the foundations for long-term success.
Selecting the right initial use cases can determine whether AI delivers real value or is just another experiment. There are four key factors to consider.
1. Start with a use case that already generates business value
The most successful AI projects enhance processes that already generate business value. If leadership is already monitoring how long a process takes, how much it costs, or how accurate it is, then the impact of AI on that process will be easy to measure. This makes it much easier to demonstrate results, gain leadership support, and scale AI across the organization.
Many early AI pilots fail because they focus on new, untested ideas that seem promising but are not tied to clear business metrics. Without a benchmark for comparison, it’s hard to prove whether AI is truly making a significant difference. Organizations that start with well-established, value-added use cases where success can be clearly measured are able to demonstrate value more quickly and secure long-term investment.
2. Solve a problem that teams are already facing
AI is more likely to gain traction when it improves processes that employees already see as a challenge. If AI speeds up a slow process, eliminates repetitive tasks, reduces errors, or even redefines a process, teams will see the benefits immediately—making them more likely to trust the technology and advocate for its use in other areas.
For ideal adoption, AI should be integrated into existing workflows without adding complexity. If a solution forces teams to completely change the way they work, adoption may slow—even if the technology is effective. The best AI use cases improve efficiency while fitting into daily operations, making it easier for teams to adopt and scale them over time. That said, new agentic AI approaches can help simplify and redesign processes so that an organization is not just automating an ineffective process.
3. Make sure your organization has the right foundation
A critical barrier I consistently observe with clients is the lack of data readiness. Fragmented data ecosystems, quality issues, and outdated infrastructure severely limit AI’s potential to deliver transformative business outcomes.
AI works best when it has the right support. High-quality data, robust infrastructure, and clear processes ensure that AI delivers reliable and consistent results. If the data AI is working with is incomplete, outdated, or scattered across disconnected systems, even a promising use case may fail to deliver value.
Organizations that successfully scale AI take the time to organize their data first. By addressing data quality early on, organizations prepare for AI that delivers reliable results and creates a more direct path to broad adoption.
4. Have a plan for what comes next
A successful AI initiative is only the beginning of broader AI adoption. When AI measurably improves a process, it generates trust and progress. Teams see its value and start looking for ways to use it in other areas. Leaders, seeing clear results, are more willing to invest in AI at a larger scale.
To scale AI effectively, organizations need to think ahead, which includes identifying new opportunities for AI, ensuring the necessary infrastructure to support more AI initiatives, and creating processes to monitor and measure AI’s impact over time. Planning for growth ensures that AI can continue to add value as it expands across the business.
Conclusion: The difference between AI becoming a strategic advantage versus an expensive experiment lies in starting with the right use cases. By focusing on practical, high-value use cases already aligned with business priorities, you’re laying the groundwork for early, measurable wins. These small victories generate excitement and build momentum among teams, creating advocates who champion AI and help it scale across the organization. Start with a clear success metric, solve a real problem, build on strong data foundations, and your organization won’t just adopt AI—it will thrive with it.