TRENDING
Why composability matters
towards developing the ability to reassemble systems, processes and teams rapidly as business priorities evolve.
Composable IT breaks technology estates into capabilities that can be reused, replaced and combined without rebuilding entire systems. For AI, this matters because agents require controlled access to data, applications and business processes.
For CIOs, composability can therefore become an execution model as much as an architecture. It enables organisations to introduce AI incrementally, isolate risk and avoid waiting for multi-year legacy replacement programmes before delivering new capabilities.
Open, composable architectures can also provide the managed APIs through which AI agents interact with enterprise components, giving organisations greater control over how automated systems access data and execute actions.
The breakdown of phased transformation
Previous waves of cloud and Digital Transformation were expected to prepare enterprises for this environment. In many cases, however, modernisation stopped before delivering the simplification CIOs had anticipated.
Core platforms were migrated, wrapped or connected to new environments rather than retired. New systems were subsequently layered over existing infrastructure, moving complexity instead of eliminating it.
That compromise becomes more visible with AI. Models and agents depend on reliable integration, accessible data and consistent orchestration across multiple systems. Fragmented technology stacks therefore make AI projects slower, more expensive and more difficult to scale.
For CIOs, this creates an uncomfortable contradiction. Leadership teams increasingly expect AI to generate productivity gains and cost savings, yet the underlying IT estate can make achieving those economics significantly harder.
The problem is not necessarily that previous modernisation programmes failed. Many delivered important improvements. The difficulty is that they often removed insufficient friction to support the speed, interoperability and flexibility now required.
The operating model constraint
Technology represents only part of the challenge. AI is also exposing weaknesses in enterprise operating models.
Unlike a traditional technology deployment, AI rarely arrives as one self-contained system. It spreads across customer interactions, network operations, engineering, software development and internal workflows. Multiple models and agents may operate simultaneously, with actions in one area influencing outcomes elsewhere.
As deployment expands, unresolved organisational and architectural problems surface through higher costs, delivery delays and difficulties moving successful pilots into production.
Earlier transformation programmes gave CIOs greater freedom to sequence competing priorities. Organisations could introduce new capabilities first, pursue optimisation later and gradually retire legacy technology.
AI compresses those timelines. CIOs are expected to maintain existing services, modernise legacy platforms, extract savings and deploy AI at the same time.
Traditional optimisation programmes built around incremental efficiency improvements can struggle with this level of simultaneous demand. Equally, multi-year transformation roadmaps risk appearing increasingly detached from pressure to demonstrate measurable AI value now.
Execution becomes the capability
The response requires a change in emphasis. Rather than positioning AI as another programme alongside cloud, automation and modernisation, CIOs can treat execution itself as a strategic capability.
AI can be applied internally to automate engineering activities, streamline operations and identify inefficiencies hidden across complex technology estates. The resulting productivity improvements and savings can then be reinvested in further modernisation and AI deployment.
That creates the possibility of a reinforcing cycle in which AI helps finance its own expansion, an increasingly important consideration in a telecom industry where IT capital expenditure allocations remain constrained.
Modernisation consequently becomes less focused on reaching an ideal architectural
54
INTELLIGENT CIO EUROPE www. intelligentcio. com