By Dipti Narang, Director – AI & Digital Experience, To The New
Every boardroom conversation on AI eventually comes down to the same question: How do we scale it?
Over the past two years, enterprises have moved rapidly from experimenting with generative AI to putting it into production. Copilots are assisting employees, AI-powered systems are transforming customer service, and intelligent agents are beginning to automate tasks that once required human intervention. The technology is advancing quickly. The harder question now is no longer whether AI works, but whether it can work consistently across the enterprise.
And that’s where many organisations are beginning to hit an invisible barrier.
The next phase of AI adoption will not be constrained by the sophistication of the models organisations choose. It will be constrained by the digital foundations they have spent the last decade building.
That may sound counterintuitive. Enterprises have spent the last decade investing heavily in cloud, customer experience platforms, data modernisation, and digital transformation. Yet as AI moves from isolated use cases to enterprise-wide deployment, those investments are revealing a new challenge. Deploying an AI capability may be relatively straightforward; scaling intelligence across business functions securely, responsibly, and with measurable business value is considerably harder.
The reason lies beneath the surface. Most enterprise technology landscapes were designed to support applications. AI depends on something fundamentally different. It requires connected data, shared business context, interoperable systems, and governance that allows intelligence to move seamlessly across the organisation. In other words, AI isn’t simply another capability added to an existing technology stack. It is changing the underlying operating model required to make digital enterprises work.
AI is exposing the limits of the enterprise we built
Adobe’s 2025 AI and Digital Trends report identifies data quality, integration, and organisational readiness as some of the biggest obstacles to delivering AI-powered customer experiences. The message is becoming increasingly clear. Competitive advantage in the AI era will depend less on who deploys AI first and more on who builds an enterprise capable of putting intelligence to work at scale.
Much of today’s AI conversation still revolves around models, copilots, and agents. They are the visible layer of the transformation. The more consequential shift is happening underneath: AI is exposing how fragmented enterprise technology environments have become and how difficult it is for intelligence to operate across them.
The more important story is happening underneath. For years, organisations have improved customer experiences by adding digital capabilities. Marketing platforms enabled personalisation. CRM systems strengthened engagement. Commerce platforms simplified transactions. Customer data platforms unified customer profiles. Each investment solved a specific business problem and delivered value. But each also added another layer to an increasingly complex technology environment. That complexity was manageable when applications could operate largely within their own boundaries. AI changes the equation.
Enterprises are moving beyond experimentation into large-scale AI deployments. At the same time, AI agents are evolving from task-based assistants into systems capable of completing end-to-end business activities. That shift is forcing organisations to confront a reality that previous technology waves rarely exposed: disconnected systems, fragmented data, and inconsistent governance are no longer operational inefficiencies. They have become barriers to enterprise intelligence.
The enterprise no longer revolves around applications
For decades, enterprise applications have been at the centre of digital strategy. CRM managed customer relationships. ERP managed operations. Content management systems powered digital experiences. Analytics platforms generated insights. Enterprise architecture was largely about connecting these systems through integrations and APIs. That model worked because applications were designed to perform defined functions. AI introduces a different paradigm.
Applications are not becoming less important; their role is changing. They will increasingly serve as systems of record, while AI becomes an intelligence layer across the enterprise, one that understands context, connects information across systems and orchestrates decisions and actions in real time. Unlike traditional applications, this layer is not confined to a single business function. It can draw on customer data, operational knowledge, business rules and governance to deliver outcomes rather than simply execute transactions.
Consider an airline responding to a passenger whose flight has been cancelled. Resolving that seemingly simple interaction may require information from booking systems, loyalty platforms, weather feeds, crew schedules, payment systems and customer service records. The passenger does not care where that information resides. They expect one seamless response. This is where the intelligence layer becomes meaningful: AI can coordinate across those systems only when the underlying enterprise is connected enough to support that orchestration.
The same principle applies across industries. Whether it is a retailer connecting inventory and customer preferences, a healthcare provider connecting clinical and patient data, or a bank bringing together transactions, risk and regulatory controls, the value of AI increasingly depends on its ability to work across organisational boundaries.
This is also where the future of customer experience begins to change. For years, digital experiences were designed around applications. Increasingly, they will be orchestrated by intelligent agents that can move across applications on behalf of users. The customer will experience the outcome; the complexity of the enterprise architecture will remain invisible.
Architecture is becoming a boardroom conversation
Until now, enterprise architecture was largely viewed as an IT discipline, owned by architects, platform teams and technology leaders. AI is changing that. The quality of an organisation’s digital foundation now influences how quickly new products can be launched, how consistently customer experiences can be delivered, how effectively regulatory requirements can be met and how confidently AI can be scaled. Architecture is becoming a strategic business capability, not simply a technology concern.
This also calls for a shift in leadership thinking. The most successful AI initiatives will not be measured by the number of copilots deployed or agents launched. They will be measured by how seamlessly intelligence can move across the enterprise, connecting people, processes, data, and decisions to create measurable business outcomes.
The next evolution of customer experience will go beyond personalisation and automation. We are moving toward autonomous journeys, where AI can anticipate needs, coordinate actions across functions, and resolve increasingly complex interactions with minimal human intervention. But that future will depend on more than powerful models. It will depend on enterprises designed to support intelligence from the ground up.
Years from now, organisations will not be remembered for deploying the most advanced AI model. Those capabilities will evolve rapidly and, over time, become widely accessible. What will endure is something far more difficult to replicate: an enterprise that can continuously absorb new intelligence, apply it with context, operate with trust, and adapt at the pace of change. The next competitive advantage will not be built at the customer interface. It will be engineered into the enterprise behind it.