IT leaders will play a key role in ensuring their organisations leverage the latest surge in digital innovation driven by AI breakthroughs.
The opportunities are enormous: by 2028, 33% of enterprise software applications are expected to include agentic AI, enabling 15% of day-to-day work decisions to be made autonomously.[1]
Yet despite 84% of CIOs believing AI will be as transformative as the internet, only 11% have fully implemented the technology.[2] There is clearly much to do. The path to AI adoption is challenging and requires skilful navigation. CIOs should carefully consider the following three questions as they continue their AI journey.
1. Where should CIOs focus their AI optimisation efforts?
The need to maximise business value should always be the primary driver of AI adoption. Paul O’Sullivan, SVP Solution Engineering, UKI, Salesforce, says CIOs should identify use cases that are scalable, high value and closely aligned with their strategic objectives. Other key optimisation areas needed to enable AI include:
- Data optimisation. There is a direct correlation between data quality and the quality of AI outputs. When CIOs plan their AI strategy, they should also optimise their data strategy on at least four fronts: data quality, data accessibility, data governance and data integration.
- AI transparency. Clear AI audit trails give CIOs visibility of how AI models make decisions, generate outputs and interact with enterprise systems (ERP, CRM, OMS etc). Transparency also enables CIOs to monitor and maximise AI performance and accuracy.
- Human/digital collaboration. AI elevates and empowers employees, enabling them to innovate, improve customer service and enter new markets. New research from Slack reveals 80% of employees are already more productive thanks to AI.[3] The key to unlocking even more value is enabling seamless collaboration between humans and AI agents. Luxury travel firm Secret Escapes, for example, is using agents to automate routine tasks and simplify workflows, providing faster, more efficient customer support.
- Sustainability. CIOs must also be constantly mindful that AI, and the data centres used to power AI, are highly energy and water intensive. CIOs are advised to seek out responsible AI partners who have access to the sustainable resources needed to minimise carbon emission and wider environmental impact.
2. When should CIOs innovate?
The decision to either ‘build’ or ‘buy’ depends on factors such as customisation needs, resource and skills availability, time to value and strategic goals. Smart firms are doing both – building and buying – but they’re carefully tailoring their approach according to each individual use case. Secret Escapes went live with Agentforce in just two weeks, thanks to low-code tools.
O’Sullivan suggests CIOs should initially leverage out-of-the-box AI capabilities baked into their existing platforms. These solutions will provide initial marginal gains. Cost savings can then be invested in wider innovation.
Innovative models can offer a competitive edge, but they can take around two years to generate a return on investment, according to O’Sullivan. Meanwhile in some use cases the focus should be on leveraging the right AI infrastructure, rather than focusing on large language models.
O’Sullivan says: “The AI landscape is evolving fast. CIOs must consider if their use case will still generate return on investment in two years’ time, otherwise they’re wasting precious resource.”
3. How should CIOs protect their tech stack from fast-emerging AI risk?
AI is reliant on networked enterprise data; therefore, CIOs must implement robust cybersecurity measures such as encryption and access controls. Intellectual property is also at risk. This includes product details not covered by patents and copyrights, proprietary algorithms, unique data and AI training methods.
Criminals are rapidly developing new ways to expose and exploit AI weaknesses as well as using the technology themselves. AI prompt injection, for example, uses prompts to reveal IP, embed instructions and override moderation rules. In response, CIOs are well advised to establish AI governance frameworks integrating risk management, regulatory compliance and ethical considerations.
“The sheer speed of AI adoption increases the risk of systems vulnerability, so it’s critical CIOs take a methodical approach to securing their stack,” says O’Sullivan.
Conclusion
The pressure is on CIOs to navigate the AI landscape, optimising their operations, making the right buy-or-build decisions and protecting their tech stack from fast-emerging AI risk. By adopting agentic AI, CIOs can innovate, protect and optimise: simpler, faster and better.
_________________________________________________________________________________________