The Operational AI Stack — Why Most Organisations Are Stuck at Assist
- Terry Chana
- 1 day ago
- 5 min read
Updated: 4 hours ago
When you’re responsible for digital experience, you’re constantly dealing with the challenge we now face: environments where work is no longer limited by a lack of tools or information, but slowed by how difficult both are to navigate.
The first meaningful shift for most organisations isn’t automation — it’s augmentation.
This reflects a broader shift in how work is operating, what we describe as the Operational AI Stack.
Introducing the Operational AI Stack
In most organisations, AI is already appearing, but in isolated ways. A copilot in one application. A chatbot in another. A standalone tool used by a few teams. There is value there, but it’s fragmented and difficult to scale across your organisation.
In reality, these capabilities are part of a broader development, which we describe as the Operational AI Stack for Intelligent Work.
This stack reflects how AI evolves within enterprises:
Assist — AI supports employees with knowledge, content, and decision-making
Automate — AI-guided workflows structure and coordinate operational processes
Agentic — autonomous AI agents execute and optimise outcomes across systems
Most organisations today are operating within the Assist layer. This layer delivers clear value, but it’s important to recognise what it is and what it isn’t.
AI assistants don’t fundamentally change how work is structured; they change how people interact within that complexity. While they don’t remove the underlying complexity, they reduce the effort required to operate within it.
The Challenge of Modern Knowledge Work
The struggle in most organisations is no longer a lack of information. It's the opposite. Employees are navigating more systems, more repositories, and more communication channels than ever — and a significant portion of their working day is spent not doing the work, but finding, assembling, and reconstructing the context needed to do it.
That friction compounds. It slows decisions, increases cognitive load, and quietly erodes engagement.
As we explored in the piece ‘How Connected Workspace Systems Improve Digital Experiences’, when systems don’t integrate effectively, employees are forced to bridge the gaps themselves. AI assistants can provide a real response to this problem.
The Assist Layer: A New Interface for Work
AI assistants introduce a different way for your teams to interact with your systems. Instead of navigating applications, employees can simply ask for what they need. A shift from navigation to conversation. It's subtle, but powerful.
While this doesn't remove complexity, it changes who has to manage it. Rather than opening multiple tools, searching across repositories, and manually assembling information, employees can ask a question and get a coherent answer.
The assistant becomes a unifying interface across fragmented systems, retrieving and presenting context in real time — so the human can focus on the decision, not the search.
How AI Assistants Augment Work
At their core, AI assistants reduce the effort required to access and use information. Their impact typically falls into three areas:
1. Simplifying Access to Context
AI assistants act as a conversational layer over organisational knowledge, often grounded through retrieval mechanisms that connect large language models to enterprise data sources. Instead of searching across systems, employees can simply ask a question and receive a coherent answer.
An employee preparing for a client meeting no longer has to search across documents, emails, and CRM systems. Asking a single question, they’ll receive a consolidated view of relevant information.
This dramatically reduces time spent searching and reconstructing context, giving your teams more time to apply their insights.
2. Accelerating Content Creation
A significant proportion of knowledge work involves producing content such as emails, reports, summaries, or updates. AI assistants reduce the effort required to get started and structure information.
Drafting a response or internal update shifts from a blank-page exercise to guided refinement, with AI generating a structured starting point that allows the user to focus the message.
This doesn’t remove human input; it amplifies it, reducing the time taken for the work while strengthening the end result. Instead of starting from scratch, your people are starting with momentum.
3. Supporting Faster Decision-Making
AI assistants provide contextual support that enables faster, more informed decisions.
A service agent handling an escalation doesn’t need to manually gather context. The assistant summarises the customer history, highlights relevant policies, and suggests possible actions.
The human remains accountable, but the effort required to reach a decision is significantly reduced. The decisions don’t change, but the path to them becomes shorter.
Reducing Friction, Not Replacing Work
It’s important to recognise where the value of AI assistants actually comes from. In most cases, productivity gains are not driven by replacing work; instead, they come from removing friction.
Across your organisation this shows up in small but constant ways: less time searching, less effort reconstructing context, fewer interruptions from switching between systems.
These small reductions compound, improving both productivity and the day-to-day experience of work. This reflects a broader reality: when organisations reduce friction in how work happens, both productivity and engagement improve.
We know that where adoption challenges exist, they often come not from capability but from the experience of working with systems that are difficult to navigate or poorly integrated. This is a problem we explored in our article From Software Spend to Business Value: The Case for Application Adoption.
The Limitations of Assist
While AI assistants help compensate for system complexity, they don’t eliminate it. If you’re already seeing value from AI assistants, this is where the limitation becomes clear; the experience improves, but the underlying structure of work remains unchanged.
Yes, they:
Support individuals
Improve interaction with systems
Reduce cognitive load
But they don’t:
Structure workflows
Coordinate processes
Ensure consistent execution
In many cases, assistants sit on top of existing complexity, relying on underlying data structures, identity layers, and system integrations that remain unchanged.
The challenge isn’t deploying assistants, it’s operationalising them across systems, data, and workflows.
Assist improves how people work within systems. It does not structure work.
Most organisations are making progress at the Assist layer while the underlying complexity of their workflows remains untouched. The gap between what AI appears to deliver and what it actually changes is wider than most realise.
That requires a different layer entirely.
What Comes Next: Structuring Intelligent Work
If the Assist layer changes how people interact with work, the next step is to change how that work itself is structured.
This is where the next layer of the Operational AI Stack comes into play:
Guided workflows that coordinate tasks across systems
Rule-based automation that structures operational processes
Real value emerges not just from the model itself, but from how it is grounded in enterprise data, integrated across systems, and governed within operational constraints.
In the next article, we’ll look at how organisations move from supporting individuals to structuring operations through AI-enabled workflows and why this shift is critical to scaling intelligent work.
AI assistants represent an important step forward.
They make work easier. They make knowledge more accessible.They reduce the friction of navigating complex systems and workflows.
But they are only the starting point.
The organisations that realise the greatest value from AI won’t stop at making work easier — they’ll rethink how work is structured, how decisions are supported, and how outcomes are delivered.
Defining how digital services and AI-enabled workflows operate, perform, and scale.
