Katie Obi, Chief People Officer at OneAdvanced on why disconnected systems are holding back AI and what HR must do next.
Artificial intelligence has rapidly moved from future ambition to present day imperative. Organisations across industries are investing heavily in AI tools, expecting transformative gains in productivity, innovation, and decision-making. Yet, despite this enthusiasm, many are finding that AI adoption is failing to deliver on its promise.
Is the issue with the technology itself, or is it with how organisations are implementing it?
From a people and operating model perspective, fragmented workflows and disconnected systems are quietly undermining progress. Rather than simplifying work, AI is, in many cases, adding complexity, driving tool-switching, duplicative tasks, and increased cognitive load for employees. If we want AI to truly unlock productivity, we need to rethink how work gets done.
The hidden cost of disconnected workflows
In many organisations today, employees operate across a multitude of systems, HR platforms, CRM tools, project management software, messaging apps, and more. Each system holds valuable data, but rarely are they seamlessly connected.
AI thrives on context. Fragmented environments lead to missing context, which in turn leads to poorer results.
Instead of a single flow of work, employees are forced to reconcile conflicting outputs from different tools or redo work due to incorrect conclusions delivered. AI-generated insights may be useful in isolation, but without context or integration, they can create additional work rather than reducing it.
Hands on usage is the best way to build competence.”
In addition, even when AI tools are connected to multiple systems, e.g. using MCP, the way of working with these tools is fundamentally different. Requests are initiated, and the requester waits for the request to complete or for more information to be requested – often switching to another task in the interim.
This leads to four critical challenges:
- Context-switching fatigue: Constantly moving between platforms and activities breaks concentration and reduces efficiency.
- Duplicative effort: The same work is repeated or needs to be redone due to a lack of integration, data and context.
- Cognitive overload: Employees must interpret and validate fragmented AI outputs, increasing mental strain. People are also reporting cognitive overload based on the higher pace of work and the increased need to deploy high functioning critical thinking skills.
- Recipient overwhelm: Generative AI can make it easier to generate work. The outputs can often be overly generic and lacking substance, and tend to be much longer than a person would have produced. This makes the output easier to create for the person generating the content; however, it makes it less easy and slower to consume for the recipients of the content.
The result is a paradox: the more tools organisations introduce in the name of productivity, the harder it becomes for employees to work effectively.
Where organisations are going wrong in AI rollout
From my perspective, the root cause lies in how organisations deploy AI. Too often, it is treated as a technology deployment rather than a transformation of work.
There are three common missteps:
1. Leading with tools, not outcomes
Many organisations start by deploying AI solutions without fully understanding the workflows they are meant to improve. This creates a proliferation of point solutions that fail to connect meaningfully into the employee experience. Organisations and individuals need to really understand what they are trying to achieve through using AI tooling, and then structure the use of such tools accordingly. Organisations also should consider whether they wish to use AI to generate or complete work on a person’s behalf, or be an enhancement to a person’s thinking (by being a thought partner, helping to spot patterns, helping to spot risks etc.).
2. Not properly considering the operating model
AI changes how work is structured, how decisions are made, and how teams collaborate. Roles are starting to merge together. Yet organisations often attempt to overlay AI onto existing ways of working without redesigning roles, processes, or governance. Equally, organisations should be cautious of making decisions about their operating model that will become unsustainable in the future. For example, eliminating early careers programmes will likely cause a talent gap and wage inflation in 2-5 years for experienced talent. Fully replacing large volumes of roles with AI may provide short-term profits, however if AI consumption becomes more expensive than employee salary costs, that operating model becomes financially untenable.
3. Underestimating the human factor
AI adoption requires behavioural change. Employees need clarity on when to use AI, how to interpret its outputs, and when human judgment is essential. Without this, adoption is inconsistent and impact limited. Organisations should define what is it that they want only a human to do – this could range from decision making to crafting the narrative or perspective that summarises insights.
The case for more connected environments
To unlock the true value of AI, organisations must move beyond fragmented systems and towards more connected environments, where workflows, data, and intelligence come together seamlessly.
In practice, this means creating more unity across the organisation that brings together data from multiple systems and applies AI in context, rather than in isolation. We are guiding our customers in this direction, using our Intelligent System of Work, IQ, to connect workflows and surface insights where work is actually happening.
What matters here is not the system itself, but the principle: AI is most effective when it is embedded within the flow of work, not sitting alongside it.
In a connected environment:
- Data becomes more consistent and accessible
- AI outputs are contextual and actionable
- Workflows are streamlined rather than fragmented
The experience for employees shifts from managing systems to improving outcomes, with AI quietly providing better insights and supporting faster execution.
HR’s critical role in AI adoption
As organisations make this shift, HR has a pivotal role to play, not just in adopting AI, but in shaping how it is implemented across the business.
Designing work for the AI era
HR must take the lead in redefining roles and workflows to fully leverage AI. This includes identifying where tasks can be augmented, where processes can be simplified, how teams should operate differently, and how skills need to evolve.
Building capability and confidence
AI will only deliver value if employees are equipped to use it effectively. HR should focus on building AI literacy, providing clear guidance, and creating space for experimentation. Hands on usage is the best way to build competence. Educating not just on the technology is also important – people should be educated on the ethical and responsible elements of AI to properly understand risks and where to apply their human judgment. Finally, it’s also important to educate people on how to efficiently use AI tools, e.g. how to optimise token consumption so that costs are sustainable for the organisation.
Establishing governance and guardrails
As AI becomes embedded in decision-making, organisations need clear frameworks around its use. HR plays a key role in ensuring ethical, consistent, and transparent practices. I would caution organisations from including any type of consumption metrics as a substitute for performance – consumption does not mean effective usage or impact.
Strengthening HR itself
AI also presents an opportunity to elevate the HR function. By automating administrative work and enhancing access to insight, HR teams can spend more time on strategic priorities—organisational design, workforce planning, and employee experience.
When, how, and why to use AI
A key barrier to adoption is a lack of clarity around the problem the organisation is trying to solve using AI , with many organisations failing to define when, how, and why it should be used. To unlock value, businesses need to be explicit that AI is best applied to repetitive, data-heavy, or time-intensive tasks; that it should act as a complement to human judgment rather than a replacement, keeping the human in the loop; and that its purpose is to improve outcomes, not simply to introduce new tools. Leaders involved in implementation and design play a critical role in bringing these principles to life, ensuring AI is grounded in real workflows and employee needs so that it is both practical to use and capable of delivering meaningful impact.
Moving from experimentation to impact
Many organisations remain in the experimentation phase, testing tools, piloting use cases, and exploring capabilities. The challenge now is scaling beyond pilots to achieve meaningful impact.
This requires a shift:
- From fragmented solutions to connected ecosystems
- From deploying technology to redesigning work
- From short-term wins to long-term transformation
- From content generation to meaningful insight
Connected environments, whether enabled through systems like IQ or broader integration strategies are central to this shift. They provide the foundation for AI to operate not as a collection of tools, but as a cohesive capability embedded across the organisation.
A people-first approach
Ultimately, the success of AI will not be determined by the sophistication of the technology, but by how well it integrates into the human experience of work.
Disconnected systems and fragmented workflows are not just operational inefficiencies they directly impact adoption, engagement, and performance.
In the people function, we have a responsibility to address this. By championing more connected ways of working, aligning operating models with technology, and keeping employee experience at the centre, we can ensure AI delivers on its promise.
The opportunity is to create an environment where work flows more seamlessly, decisions are better informed, and employees are empowered to focus on what matters most.
That is where real productivity gains will be realised.
