Generative AI is everywhere. Whether it’s at industry conferences, in boardroom strategy sessions, or during casual peer conversations, CIOs can’t go a day without hearing it mentioned somewhere. The numbers are staggering. Gartner predicts global spending on generative AI will grow 76% to hit $644 billion in 2025¹. With that kind of investment, the pressure to deliver tangible results is higher than ever.

But here’s the reality check: not every GenAI project makes it past the starting line. In fact, Gartner estimates that at least 30% of these projects will be abandoned after the proof-of-concept stage, often due to unclear business value or poor risk management². I have seen first-hand, excited teams build flashy demos and then struggle to scale or demonstrate ROI.

So what should CIOs like us focus on to avoid these pitfalls? Here’s my take  based on experience and what I’m hearing from other leaders.

1. Embrace the Journey - Start Early, Learn Fast

AI isn’t a quick fix. It’s a journey. The sooner your team starts experimenting, the faster you’ll discover what GenAI can and can’t do for your business. Start small — maybe automate support tickets or streamline a workflow. As your team gains confidence, you can expand into more ambitious areas like predictive analytics or smarter automation. The key is to grow organizational maturity and adapt as the technology evolves.

2. Connect AI to Real Business Goals

It’s easy to get distracted by cool demos but every project must be tied back to clear business objectives, such as increasing revenue, improving efficiency or driving innovation. Set measurable goals, even if small, from the outset so you can track ROI and keep everyone focused on outcomes, not just novelty.

3. Be Prepared for Uncertainty

GenAI doesn’t always give predictable results. Unlike traditional IT projects, you might run into unexpected challenges or shifting timelines as data or training needs change, in turn leading to distrust between the project teams and executives. It’s important for both sides to understand this and communicate openly about progress and setbacks. Remind your stakeholders that AI is more of an exploration — patience and transparency go a long way in maintaining trust.

4. Data Quality Matters - A Lot

No matter how advanced your AI, it’s only as good as the data you feed it. Messy, incomplete or inconsistent data can stall even the most promising project. Many organizations underestimate the effort and collaboration needed to clean and organize legacy data. Investing in good data preparation and ongoing governance is essential.

5. Build Interconnected Systems

Data silos are a huge hurdle. When information is scattered, AI models can’t deliver reliable insights or sufficient end-to-end value as the required data is not linked or complete. Focus on creating interconnected systems using standardized APIs and interoperability frameworks. Simultaneously, keep your architecture consistent and your governance strong. Otherwise, you could end up with a chaotic mix of disconnected tools, inefficiencies and security headaches.

6. Prioritize Responsible AI Governance

Good governance is the backbone of successful AI deployment. That means having structured risk management frameworks, like the National institute of Standards and Technology’s (NIST’s) AI Risk Management Framework, to identify and mitigate risks throughout the AI lifecycle. Regular audits, diverse training data and transparency help prevent biases and misinformation. With more than 80% of enterprises expected to use GenAI APIs or deploy GenAI-enabled apps by 2026, getting governance right is important for the right AI culture and long-term success³.

7. Security and Privacy Are Non-Negotiable

GenAI introduces new security and privacy challenges. AI models often handle sensitive data and can be targeted by adversarial attacks or data poisoning. Robust cybersecurity — encryption, access controls, anomaly detection — is a must. Your systems should be built with privacy-by-design principles. Regular security audits and continuous monitoring are essential.

8. Stay Ahead of Regulations

Regulations are evolving quickly. The EU’s AI Act, for example, officially came into effect in August 2024 and is being rolled out in phases. Some rules, like bans on certain high-risk AI practices, started in February 2025, while the main requirements for high-risk systems kick in from August 2026⁴. Preparing for compliance means developing standards, conducting risk assessments and embedding transparency — especially for high-risk sectors like healthcare and finance.

9. Keep an Eye on Costs

Scaling GenAI can get expensive quickly. While pilot projects are manageable, costs can skyrocket with more data, retraining and broader deployments. I recommend designing architectures that balance performance and resource use, using dashboards to monitor costs, and optimizing models. Smart cost management ensures your AI investments deliver value without draining your budget.

10. Don’t Forget the Human Element

Technology alone is not enough. Often, the biggest challenges aren’t technical; they’re about unclear ownership, resistance to change, false expectations, or organizational silos. Boston Consulting Group’s 10-20-70 rule really resonates here. Only 10% of success depends on algorithms, 20% on data and technology, while a whopping 70% on people and processes⁵. What does that mean? Investing in training, upskilling and change management is essential for lasting results. 

Final Thoughts

AI is moving fast. The organizations that succeed will be those willing to start early, learn continuously and adapt quickly. Focus on targeted use cases, build internal expertise, digitize and unify your data systems, and expand gradually. This iterative approach will help you stay ahead in a rapidly changing landscape.

About Altius

Founded in 2019, by Big Four management consultancy professionals, Altius is a premier global advisory firm to Private Equity funds and their portfolio companies in the Mid-Market. From due diligence to value creation and from strategy through execution, Altius delivers sustained impact with velocity, precision and expertise at every stage of the M&A lifecycle. We are thought leaders in carve-outs, integration and value creation using levers such as Global Capability Centers, Global Business Services, Zero-Based Design and Digital Transformation. We support PE Deal and Operations teams from due diligence to exits, across asset-intensive and information-intensive industries in America, Europe, and Asia with a global footprint of practitioners.

Notes

      ¹ “Gartner Forecasts Worldwide GenAI Spending to Reach $644 Billion”, Gartner, March 31, 2025,
https://www.gartner.com/en/newsroom/press-releases/2025-03-31-gartner-forecasts-worldwide-genai-spending-to-reach-644-billion-in-2025#:~:text=CIOs%20Must%20Prepare%20for%20Rising,a%20forecast%20by%20Gartner%2C%20Inc.

      ² “Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept By End of 2025”, Gartner, July 29, 2024,
https://www.gartner.com/en/newsroom/press-releases/2024-07-29-gartner-predicts-30-percent-of-generative-ai-projects-will-be-abandoned-after-proof-of-concept-by-end-of-2025.

      ³ “Gartner Says More Than 80% of Enterprises Will Have Used Generative AI APIs or Deployed Generative AI-Enabled Applications by 2026”, Gartner, October 11, 2023,
https://www.gartner.com/en/newsroom/press-releases/2023-10-11-gartner-says-more-than-80-percent-of-enterprises-will-have-used-generative-ai-apis-or-deployed-generative-ai-enabled-applications-by-2026.

      ⁴ Elena Chinita, “The EU AI Act: What’s (Not) Going According to the Plan”, The Recursive, April 16, 2025,
https://therecursive.com/the-eu-ai-act-what-s-not-going-according-to-the-plan/.

      ⁵ “Artificial Intelligence at Scale”, Boston Consulting Group,  https://www.bcg.com/capabilities/artificial-intelligence.