FEATURE: INVESTMENTS engineers and general users to build AI literacy and foster a data-first culture. At this stage, many realise their data infrastructure can’ t support their AI ambitions, prompting significant investment in cloud platforms, modern data warehouses and governance tools.
The third phase is about scaling. Companies build Centres of Excellence that act as central hubs for strategy, technical leadership and governance. These teams manage projects, set policies, standardise tools and ensure ethical and regulatory compliance. They’ re also building reusable AI components and platforms that let business units build on existing models.
What CIOs miss about resilience
Too often, CIOs view resilience in AI as a technical concern. It’ s a strategic one. A resilient AI programme aligns with core business needs and evolves through failure, disruption and change. This resilience separates enduring transformation from hype cycles.
Many companies fall into“ pilot purgatory” where AI initiatives never progress. Resilience ensures every project is rooted in a clear business case and guards against overinvestment in trendy tools. Most importantly, it bridges the gap between business and technical teams. AI built with both perspectives gains traction and impact.
How to build a system that supports both innovation and control
Jeremy Brown, CTO, GitGuardian
We’ re in the middle of an AI explosion and CIOs are caught between a rock and a hard place. Everyone wants to experiment with AI tools, but you’ re worried about security disasters waiting to happen. You can’ t stop this train, so you’ d better learn to steer it.
My approach: let a thousand flowers bloom but build the right garden walls first.
Your marketing team is already using ChatGPT, your finance team is automating, your engineers are coding with AI assistants. The question isn’ t whether people will use AI – it’ s whether they’ ll do it safely or go rogue with shadow IT. The key is foundational infrastructure.
First, use an LLM proxy to route all AI interactions through a central system with enterprise agreements – this ensures your data isn’ t training someone else’ s model.
Second, provide secure app hosting with sandbox and production-grade environments, so teams can experiment without bypassing security.
Add security tooling – SAST, DAST, NHI security, firewalls, role-based access – but keep it lightweight so it doesn’ t block innovation.
Non-technical users are building workflows using tools like Zapier, n8n and Make. com. These platforms let people plug systems together safely without code.
CIOs must create structures that protect the enterprise without killing creativity. Start with forming an AI governance committee including voices from across technical, legal, compliance, finance and frontline business. This group should prioritise projects and allocate resources.
Use a dual-track investment model. One track focuses on resilience: projects with defined business cases, ROI and strategic alignment. The other track supports open innovation: a portion of the AI budget goes to high-risk, experimental ideas. These projects are judged on learning potential, not just returns. If they fail, they must be reviewed and lessons shared.
The new CIO mandate
Today’ s CIO must lead like both a venture capitalist and risk officer. It’ s no longer about uptime or trends. It’ s about building durable AI foundations, aligning with strategy and enabling responsible experimentation.
Implement smart data access management. Let AI agents access only what they’ re meant to. When AI uses real company data securely, that’ s when the magic happens.
And here’ s what few talk about – you’ re about to see an explosion in service accounts, API keys and automated systems. You need tools to manage these non-human identities before they become a nightmare.
At GitGuardian, we use AWS Bedrock with data protection agreements, and an LLM proxy that maintains role-based access to our CRM, data lake and internal wiki.
The result? People are building agents that talk to agents. Our automation is sophisticated and secure.
Your job as CIO isn’ t to say no to AI – it’ s to make saying yes as safe as possible. Build the infrastructure, set the guardrails, then get out of the way.
Getting AI right doesn’ t mean playing it safe. It means playing it smart.
The companies that master this will innovate fast, safely and together. p www. intelligentcio. com INTELLIGENTCIO LATAM 27