FEATURE: INVESTMENTS
Magnus Tagtstrom, Corporate VP of AI
Transformations and GM Europe, Iterate. ai tolerance in relation to the initiative. It is also a CIO’ s responsibility to collaborate with the business to understand the costs associated with different levels of resilience and determine the optimal cost-benefit mix for the project.
Embrace a Platform Mindset
CIOs should consider platform-based investments such as SaaS solutions, cloud-native ecosystems, unified data layers and modular AI toolchains, which can evolve with the organization while providing consistency and reduced risk through ready-made integrations.
Implementing a well-defined platform provides:
• Reduced vendor sprawl and integration headaches
• Better cost management through reuse and scale
• Easier adoption of new capabilities
• Simplified training and recruiting
Embed AI Into Operations
The integration of AI into current business processes can yield significant productivity gains if implemented correctly, but CIOs must also ensure the security of their organization’ s data.
This means:
• Protect sensitive data with access controls through automation
• Establish governance for AI development and implementation incorporating defined measures
• Continuously monitor AI through automation to enforce those measures
Invest in People and Processes resilient foundation enables confident experimentation, while well-governed innovation drives efficiency and strengthens risk posture. CIOs who strike the right balance between investing in both durable infrastructure and forward-looking capabilities will unlock the transformative potential of AI in a secure and sustainable manner.
Magnus Tagtstrom, Corporate VP of AI Transformations and GM Europe, Iterate. ai
Just about every CIO I speak with is under intense pressure to start tying AI strategy to business results. Some have reacted by jumping on the biggest, mosthyped AI models and found that pouring money into giant, general-purpose LLMs led to budgets quickly spiralling out of control, systems slowing down and challenging new security concerns piling up. But as with tech advances that have come before AI, the best CIO strategy is one that fits the use case at hand. Some of the smartest investments right now focus on specialized, right-sized AI models that deliver what a CIO’ s business actually needs without unnecessary complexity or cost.
In no uncertain terms, a much higher percentage of CIOs need to prioritize small language models( SLMs) to balance innovation with resilience. Large models( requiring huge GPU clusters) are expensive upfront and even more so in the long run, while often introducing lag that can make real-time applications impractical. SLMs offer a clear alternative by still achieving the performance CIOs are looking for but at 10 % of the cost( or less) of massive models. That cost saving alone can make the difference between an AI pilot project that fizzles and a scaled deployment that truly transforms a workflow or application.
CIOs should prioritize investments in training, crossfunctional collaboration and operating models that foster agility, as people are at the core of a successful AI implementation. For example:
• Provide just-in-time training that focuses on the skills required for employees to implement AI in their roles effectively
• Facilitate working sessions with cross-functional teams to collaborate on implementing and enhancing their processes using AI
Innovation and Resilience Can Reinforce Each Other
Innovation and resilience are often viewed as competing priorities, but when approached strategically, they become powerful complements. A
Talking to one CIO recently, I heard how their team moved from a general-purpose LLM that was draining US $ 30,000 a month to a specialized model trained on their own support documents. The switch brought costs down to US $ 2,000 a month while improving the accuracy of answers for customers. Those freed-up funds were then used to modernize other key systems that had been waiting for budget approval.
Stories like this are a reminder that targeted AI investments can unlock innovation across the entire organization( in areas that might not even overlap with AI). With the rise of Agentic AI and different models starting to talk to each other, the cost containment that is possible with SLMs becomes even more important.
Investing in specialized models also strengthens resilience. Smaller models can be retrained quickly
36 INTELLIGENTCIO NORTH AMERICA www. intelligentcio. com