Ascendion CTO & Global Head of Solutions Vaibhav Vora says CIOs risk falling behind by remaining stuck in AI experimentation and must embrace agentic AI as a core driver of productivity, innovation and Digital Transformation.
In the late 19th century, the transition from gas to electricity began to reshape how societies, workplaces and cities were organised. But its introduction was not smooth sailing and incurred much initial scepticism from business owners that the costs and complexity of implementation would outweigh the benefits.
Retrofitting factories and offices required extensive rewiring and new generating equipment, often forcing organisations to overhaul infrastructure that was already functioning perfectly well.
As a result, early adoption was frequently cautious and piecemeal. In fact, one newspaper in 1880 noted of local factories’ attempts at converting to electricity that, “The arrangements thus far are quite experimental.”
This infrastructural overhaul is likely familiar to CIOs, who are undertaking a similar shift with the development of AI. The technology is no longer just improving existing processes but, through autonomous agentic systems, has the potential to transform how organisations operate at a fundamental level.
But too many CIOs are dragging their heels in the experimentation phase and adopting the ‘Ostrich Strategy’.
They are ignoring the challenges and opportunities AI systems present and instead blindly clinging to old processes and assumptions.
Within a few years, the organisations that bury their heads in the sand to the transformations now underway could find themselves struggling to compete. By contrast, CIOs who act early and seize on the potential of agentic AI to power business operations could gain a powerful advantage in speed, productivity and innovation.
The agentic AI revolution
An excuse many CIOs use for evading the challenge posed by AI is that the technology is not ready — it is not mature, scalable or reliable enough to use. But this is no longer the case. Tools are improving rapidly and some of the world’s largest organisations are deploying them at scale – Gartner predicts enterprises will have an average of 150,000 agents in deployment by 2028.
At the same time, nearly US$3 trillion is projected to be invested into AI infrastructure in the next three years. This means that AI is becoming faster and more efficient and is increasingly embedded in enterprise platforms rather than experimental tools.
New agentic systems mark a step up in capability. They do not just produce new data in response to questions or queries, they can act, decide and coordinate independently. They are capable of orchestrating data and processes at multiple levels of an organisation with unparalleled speed and efficiency, reflected in this year’s HFS Horizons: Agentic Services research that uncovered productivity as the top priority for AI amongst 74% of enterprises.
There is a lot of AI hype and comparison but the change underway is genuinely as pivotal as the introduction of electricity into factories in the 19th century, which did not just make elements of the manufacturing process more efficient but powered the entire enterprise and reshaped how production was organised.
Leadership mistakes
Too often, CIOs are deploying and testing generative tools as part of fragmented roll-outs that offer measurable but comparatively limited gains instead of embedding agentic AI systems that run automatically across an end-to-end process.
Consider the difference. A helpdesk AI chatbot saves minutes; an agentic system embedded in core operations impacts profit and loss.
AI adoption must be a board-level priority that requires a fundamental reassessment of technical systems at every level of an organisation.
If not, workers often find themselves producing ‘AI workslop’ – that is, low-value or erroneous outputs typically generated by misconfigured or disconnected tools. This is because they are still navigating legacy systems with little clear guidance on how AI should be used safely or effectively.
Some of the key challenges for CIOs in deploying agentic AI are therefore about trust and coordination. How can agentic AI systems be put to work at scale without damaging trust, increasing risk or creating chaos while still delivering measurable productivity gains?
Many CIOs are encountering what has been called an AI ‘Trust Valley’. The AI Trust Valley occurs when it is unclear who is accountable for the AI technology, how systems can be monitored and how they can be audited.
CIOs hesitate and become stuck in a loop of endless experimentation with new tools. Breaking this is about ensuring that accountability, transparency and trust are engineered into AI systems from the start.
Abandoning the ostrich strategy requires not just the adoption of new technology but a change in leadership mindset.
Lessons for CIO leaders
There are several steps CIO leaders need to take to begin harnessing the power of AI systems in their enterprises and face head-on the challenges and opportunities they pose.
The change must start with how they view and approach the technology. CIOs should be comfortable using, discussing and deploying AI. They must model the new approach to agentic systems that successful integration will depend on, signalling adoption of the technology is a strategic priority.
They must then map every corporate function, including talent, engineering, marketing, finance, HR, IT and operations, to identify where agents and AI systems can improve productivity by orchestrating workflows and reducing friction across the organisation.
I like to call this Engineering to the Power AI (EngineeringAI), a discipline where human engineers and agentic systems operate as a single layer rather than as parallel workstreams. This method functions on three interlocking pillars: platform, process and people.
Platform
This is the technical foundation. AI has the capacity to significantly boost the productivity of workers through handling routine work while humans focus on higher-value tasks. But too often AI is implemented by experimental labs or as isolated experiments.
It is only by embedding it in core processes that measurable returns are achievable and AI’s benefits become scalable.
While up until recently most businesses sat firmly in the experimentation phase, the next chapter is where value is actualised – according to McKinsey, 23% of organisations have begun scaling their AI projects.
This involves integrating AI into the legacy systems many employees work with, including existing software, internal data, workflows and compliance systems. When this is achieved, AI helps unlock hidden value in legacy systems, enabling them to operate faster and more efficiently.
Process
How do we keep that foundation under control? Pre-built, plug-and-play agentic workflows with humans-in-the-loop oversight engineered in from the start, not bolted on afterwards.
Processes must be designed to ensure that operating discipline remains intact. CIOs must ensure that AI-enabled workflows operate smoothly, that risk is managed properly and that quality is maintained.
To achieve this, humans must be given oversight over key processes to maintain trust and accountability while allowing agentic systems to automate routine decision-making and operational coordination. Strong governance frameworks are as important to the successful deployment of AI as technological infrastructure.
People
A structured talent-transformation playbook that evolves engineering roles into agentic role families with clear growth pathways – the future-proof workforce.
The successful deployment of the technology depends on the new relationship being formed between humans and AI. It is what we call Silicon plus Carbon, or Si+C.
It comprises the combination of silicon-based AI systems offering unparalleled capacity to process and coordinate processes and carbon-based humans, providing the imagination, creative insight and ethical judgement that AI cannot replicate.
This is a relationship which is being invested in now for future success; the vast majority of senior executives (88%) are increasing their budgets for agentic AI.
CIOs must encourage workers to begin collaborating with agents now so that human expertise and machine capability reinforce each other rather than compete.
These three layers must work together to leverage the full potential of agentic AI systems because, ultimately, they are inter-reliant.
Human-AI partnerships will stall if the technology is not optimised. The technology will remain underutilised and mistrusted unless embraced and monitored by humans. AI will not achieve its full potential unless fully integrated into legacy systems.
But for those willing to take the plunge, a Si+C relationship is already paying off. It is encouraging to see knowledge workers and coders, amongst others, lead the charge to boost productivity and unlock new learning opportunities.
The CIO rulebook for deploying AI into businesses is still being written. Ditching the ostrich strategy requires courage, imagination and the willingness to initiate deep-level change.
But the organisations that initiate the process now will set the pace for everyone else to follow in an economy increasingly defined by how effectively organisations combine silicon and carbon to get work done.
