CIOs are under increasing pressure to embrace AI-driven innovation without compromising security, stability or cost-efficiency. Industry experts share practical strategies for balancing transformative technology investments with enterprise resilience in today’s data-driven landscape.
Alexandre Bastos Borges, CIO, Pernambucanas
There’s an inflection point in sight: data and AI drive new business models every day while the negative headlines fall on those who move forward without a solid foundation – ransomware attacks, cloud outages, GDPR violations.
How, then, does a CIO balance technological boldness with operational robustness? Three movements complement each other:
- Start with the strategic “why”:
Each investment needs to be linked to a measurable business objective – revenue growth, efficiency or risk reduction. Map corporate OKRs and associate them with clear indicators (avoided churn, gross margin, average recovery time). AI initiatives that don’t demonstrate impact on these KPIs fall into the category of low-cost experiments rather than the executive priority queue. - Strengthen the core for resilience:
Before moving on to sophisticated designs, consolidate the foundation: architecture with multiple availability zones, immutable backups, end-to-end observability. Consider these layers as invisible capital – they rarely shine on the committee but they raise the innovation ceiling. An MLOps pipeline only generates ROI if it runs on reliable data and fail-safe infrastructure. - Manage a disciplined bimodal portfolio by distributing the budget across three buckets:
a. Run – maintain essential operations: 50-60%
b. Grow – scale what already works: 20-30%
c. Transform – explore AI/data bets: 15-20%
Reevaluate this allocation every quarter. If a Transform project proves value, it migrates to Grow. If something in Run is automated and cheaper, it frees up resources for new bets. Transparency in this “bucket dance” minimises friction between innovation and support teams.
Metrics that matter:
- Time-to-value of AI experiments (90≤day proof)
- MTTR of the data platform (goal: < 30 min)
- Savings reinvestment ratio (how much of the efficiency gains goes back to Transform)
- Adoption rate of internal AI solutions (end-user engagement)
Conclusion
Prioritising is, in practice, choosing where not to invest now. Without robust foundations, AI initiatives become expensive and risky bets. With stable operations, each new project converts into tangible value for the business. The CIO who masters the essentials creates the space to innovate with confidence – delivering customer benefits while keeping the company prepared for the unexpected.
Dr Roman Atachiants, Head of AI and Data Architecture, inDrive
Balancing innovation with resilience
Today, we face a critical challenge: how to drive innovation while building resilient, trustworthy systems. At inDrive, we’ve learned that the most effective technology investments are those that push boundaries while reinforcing the foundations of trust, security and human connection.
- Start with purpose
The best AI investments solve real problems that users and businesses face. Rather than automating for efficiency’s sake, companies should focus on technology that creates genuine value and empowers people. At inDrive, we use AI for driver identity verification and fraud detection not because it’s trendy but because safety is fundamental to our mission.
As a first step when considering new technology investments, CIOs should ask themselves two essential questions: does this investment align with our core values, and does it enhance human capabilities or simply replace them? Technology should never be a substitute for human judgement and empathy. Instead, businesses should focus on using it as a force multiplier that empowers their teams and helps them better serve their customers.
- Build platforms that scale
Companies should invest in unified data and AI platforms with strong governance from day one. This helps ensure compliance and enables your systems to adapt and grow as you scale without compromising security or putting user trust at risk. At inDrive, we’re scaling our data and AI platform to enable real-time analytics and efficient MLOps while maintaining strict privacy controls (e.g. EU’s GDPR, AI Act), regardless of local regulations.
Resilience comes from having robust foundations. So, establishing clear AI governance, internal codes of conduct and cross-functional teams that can experiment safely while scaling what works is essential. This approach allows companies to innovate at pace while maintaining safeguards to protect what matters most.
- Automate while keeping a human touch
Not every process needs AI, and not every AI solution needs to be built in-house. At inDrive, we build when differentiation matters, like with our consumer experience and logistics algorithms, and buy when it accelerates time-to-value. This ensures resources are focused where they can create the most impact.
Some interactions require empathy and human judgement that AI cannot replicate, so companies should be intentional about where not to use AI. For example, at inDrive, our support agents handle complex customer enquiries that require emotional intelligence while AI chatbots handle more routine questions. This hybrid approach improves both operational efficiency and customer satisfaction.
Additionally, the most successful AI systems should improve, rather than replace, human decision-making. For example, our price recommendation system uses AI to suggest fair ride rates based on real-time data, but the app allows users to negotiate and ultimately choose the price they offer. This preserves human agency while leveraging the benefits of AI.
- Foster an AI-first culture
Technology investments must be paired with cultural transformation. Businesses should create environments where teams can experiment safely, learn from failures and scale what works. At inDrive, we promote AI literacy through training programmes and champion networks. We also invest in local talent and run company-wide mentorship schemes. The most resilient organisations develop internal capabilities while staying connected to broader industry trends and best practices.
While changing the culture, companies should define success beyond traditional metrics. Yes, measure speed and operational efficiency, but also track trust indicators: user satisfaction, reduced friction and improved outcomes. For example, at inDrive, we measure success by reduced cancellations, faster support resolution, higher customer satisfaction scores and increased conversion rates – metrics that reflect both innovation and user value.
- The path forward
There is a lot of hype around AI – some of it justified, some not. CIOs should prioritise investments that create a virtuous cycle: technology that enables innovation while building trust, systems that scale efficiently but remain human-led, and platforms that drive growth while reinforcing organisational values. The future belongs to organisations that can move fast without breaking things, and harness AI’s potential while keeping human connection and fairness at the centre.
Igor Benincá, Head of AI, Indicium
As AI races from pilot projects to core business infrastructure, CIOs face a dangerous paradox. They are expected to lead Digital Transformation at breakneck speed while also ensuring enterprise systems remain secure, stable and aligned to long-term goals. Many will fail not because they lack ambition, but because they invest in AI the wrong way.
To balance innovation with resilience, CIOs need a smarter investment strategy that matches their organisation’s stage of AI maturity. They must also build structures that allow experimentation without derailing enterprise stability.
Three phases of AI investment most companies must navigate
Across industries, companies are investing in AI across three distinct phases. Each phase reflects a different level of maturity and success depends on investing in the right priorities at the right time.
In the first phase, companies lay the foundation. Strategy takes centre stage. CIOs are spending on consulting, internal workshops and leadership alignment to define a vision for AI tied to business objectives. They’re also identifying internal talent, conducting skills gap analyses and developing training or hiring plans. Modest investments in pilot platforms or proof-of-concept tools are common – not to build production systems but to explore possibilities and identify early roadblocks.
In the second phase, the focus shifts to discovering value and enabling adoption. CIOs are funding collaborative efforts to prioritise AI use cases through structured workshops and team integration. Organisations are rolling out tailored training for both 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
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.
Getting AI right doesn’t mean playing it safe. It means playing it smart.
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.
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.
The companies that master this will innovate fast, safely and together.
