"AI moves at the pace of trust, and trust needs predictable outcomes. Large organizations can't run on rules the model will probably follow. They need rules that always hold."- Sam Hassani, CTO, Unily
There's no shortage of AI ambition in the enterprise.
New generative AI tools, AI agents, AI models, and platforms promise to transform how organizations operate. Business leaders want to move faster, employees want technology that makes their lives easier, and boards increasingly expect evidence that substantial investments in artificial intelligence are paying off.
For CIOs, the pressure is mounting.
You're expected to accelerate innovation while managing costs, protect the organization without slowing it down, and integrate emerging technologies into information technology environments that have often evolved over decades.
Scan any list of the latest technology trends and you'll find everything from quantum computing, edge computing, blockchain, digital twins, augmented reality, virtual reality, autonomous vehicles, and the internet of things (IoT) competing for attention. Those technological advancements are flashy and exciting. But for most CIOs, they aren't the problem keeping them up at night.
The challenge isn't a lack of technology. It's making everything work together.
At Unily's Unite 26 conference, CEO Lokdeep Singh described this growing complexity as the "Fog of AI": a world in which technological possibilities are multiplying faster than organizations can establish a clear path to value.
It's a challenge reflected in wider industry research.
Deloitte's Tech Trends 2026 reports that just 11% of organizations have agentic AI systems in production, despite 38% piloting them. Microsoft's 2026 Work Trend Index argues that organizations must rethink how work is structured as agents take on more execution and finds that organizational factors such as culture, manager support, and talent practices account for twice the reported AI impact of individual effort alone. And Gartner's Top Strategic Technology Trends for 2027 identify AI context platforms, AI agent identity, and disposable applications as important foundations for the next generation of enterprise technology.
These show different perspectives but a common message: the next phase of AI transformation will be less about acquiring capabilities and more about making them work together.
Our take? In 2027, the AI advantage won't come from the model. It will come from what your AI knows about your organization and your people.
For CIOs, we believe that creates an opportunity to rethink the digital workplace, not as another application in the enterprise technology stack but as the environment where organizational knowledge, technology, and people come together to get work done.
Here are eight shifts we expect to define that opportunity in 2027.
1. The Next Competitive Advantage Will Come From Simplifying Complexity, Not Adding Capability
The question for CIOs is changing from "What else can our technology do?" to "Why is getting work done still so complicated?"
Enterprise technology has become extraordinarily capable.
After years of digital transformation, and with software as a service (SaaS), cloud computing, and multi-cloud estates now the norm, most organizations have applications for almost every conceivable business requirement. Collaboration, communication, HR, finance, knowledge management, and service delivery are all supported by increasingly sophisticated systems.
Yet employees still struggle to complete relatively straightforward tasks.
At Unite 26, Norman Nowak, VP, Executive Management Technology & Innovation at Deutsche Telekom, shared a familiar example. Employees used to search the intranet for information that actually lived in a separate HR portal or another employee services system. Deutsche Telekom's answer was MyT, a unified employee experience layer that gives everyone one clear starting point across mobile and desktop.
It's a useful example of a broader architectural shift.
Unily and Gallup's 2026 research found that 54% of U.S. workers constantly or often switch between digital tools to complete their main tasks. Among employees whose information is spread across many locations, that figure rises to 78%.
The State of the Digital Workplace Experience
AI could help solve this problem. But deployed carelessly, it will make it worse.
Agent sprawl is the new app sprawl. CIOs spent a decade fighting application sprawl, and the same pattern is now repeating with agents. Every vendor is shipping one. Without a governed front door, employees won't get simplicity. They'll get a new layer of fragmentation.
"CIOs spent a decade fighting app sprawl. Agent sprawl is the same problem, only faster. Every vendor is shipping an agent, and without one governed front door, employees just inherit a new layer of fragmentation." - Sam Hassani, CTO, Unily
Complexity also isn't a simple function of size. A 5,000-person organization with acquisitions, regulators, and frontline staff can be far more complex than a tidy 50,000-person one. What matters is complexity relative to capacity: how much technology, change, and risk your teams are being asked to absorb with the people and budget they actually have.
In 2027, we expect more CIOs to prioritize the experience of the technology estate rather than the capabilities of individual applications. And with budgets under pressure, that will mean rationalization over new spend. The winning move won't be another tool. It will be a layer that makes existing Microsoft, HR, IT, and line-of-business investments, across hybrid cloud and on-premises systems alike, more relevant and accountable for every employee.
2. Context, Not the Model, Will Become the New AI Bottleneck
AI models are now a commodity. What will set one organization's AI apart from another's is the context it works from.
For the past few years, the CIO question has been "Which model?" In 2027, it will be "What is our AI actually grounded in, and can we trust it?" Whether it's generative AI, machine learning, or natural language processing, and whether it runs on cloud computing platforms or in your own data centers, AI is only as good, and as safe, as what it draws on.
"Models are now a commodity. In 2027, the AI advantage won't come from the model you pick. It will come from what your AI knows about your organization and your people, and whether you can trust it." - Sam Hassani, CTO, Unily
Every enterprise contains an extraordinary amount of information.
Policies, procedures, expertise, decisions, documentation, and institutional knowledge are distributed across applications, repositories, departments, and conversations.
Much of it is difficult to find. Some is duplicated or outdated. And some of the most valuable knowledge exists only in employees' heads. Unily and Gallup's research shows how deep the trust gap already runs: only 26% of employees who use digital information say their organization's information is very accurate, and employees are about twice as likely to trust people (60%) as digital resources (29%) when they need information to do their jobs.
As intelligent systems become more capable of answering questions and initiating actions, they need reliable organizational context to operate effectively.
A general-purpose model may understand how a procurement process usually works. But does it know how procurement works in your organization? Which policies apply? What approvals are required? And which information is an employee authorized to access?
Without that context, even sophisticated AI produces answers that are plausible but unhelpful.
This is where the digital workplace will take on a more strategic role: As a system of record for enterprise knowledge. This isn’t simply a place to store information but a governed foundation that makes clear which knowledge is authoritative, who owns it, and who can access it. This is the role Unily is built to play, bringing structure, ownership, permissions, and publishing governance to the knowledge employees and AI rely on. The distinction matters: finding an answer isn't the same as knowing it comes from a source the organization trusts.
The implication for CIOs is significant. Knowledge management can no longer be treated as the responsibility of content teams or individual business functions. It's becoming part of the architecture that supports intelligent enterprise operations.
At Unite 26, dentsu explained why. A year after launch, its intranet had drifted: editors had gone inactive, content was no longer trusted, and it wasn't fit to act as a knowledge base for AI. As dentsu put it, "AI can't scale on fragmented, outdated content." So the team ran an audit that removed hundreds of pages and editors, put a named owner and last-updated date on every page, and made Unily's content lifecycle review a monthly KPI. Only then did it build an AI Hub as "a single destination for everything AI." The hub now draws 70 times the views of the average site.
ASML faced a similar problem at greater scale. After hiring more than 11,000 people in two years, its intranet had become overwhelmed with outdated content from 1,300 publishers. ASML's response was to make the intranet a trusted "lighthouse": a radical cleanup, strict ownership limits, gold, silver, and bronze content tiers, and quarterly reviews by a permanent team. Pages fell from tens of thousands to around 3,000, and employee satisfaction with the intranet rose from 2.87 to 3.97 out of 5.
These initiatives illustrate why AI readiness often begins with work that isn't particularly glamorous: understanding what information exists, whether it's accurate, and who is responsible for maintaining it.
From permission-aware to person-aware AI
But context isn't only about content. Respecting permissions is table stakes. The next step is AI that understands who someone is: their role, location, skills, working pattern, and environment. That's how employees will get answers that apply to them, not just answers they're allowed to see. A nurse on a night shift, an engineer on a rig, and a finance analyst at head office should get different, relevant answers to the same question.
"Respecting permissions is table stakes. A nurse on a night shift, an engineer on a rig, and a finance analyst at head office can ask the same question and need three different answers. AI that doesn't know who it's talking to isn't really helping anyone." - Sam Hassani, CTO, Unily
The goal isn't simply better answers. It's helping people understand what matters, why it matters, and what they should do next.
3. The Digital Workplace Will Move From Providing Information to Enabling Action
Finding the right answer is useful. Being able to do something with it is where the real value begins.
Consider a common employee request, such as someone who wants to understand their organization's parental leave policy.
Today, they might search the intranet, locate a document, interpret the guidance, open an HR application, and work through several steps to submit a request.
An AI assistant or chatbot could improve the first part of that experience by providing a clear, relevant answer. But the employee still has work to do.
The bigger opportunity is to connect understanding with action.
What if the employee could ask a question, receive the appropriate guidance, and initiate the relevant process within the same experience? Or a manager could request help preparing for a performance conversation and move directly from guidance to the appropriate workflow?
This represents a meaningful shift in how enterprise technology is designed. Robotic process automation (RPA) and hyper-automation taught organizations to automate the steps inside a process. The next wave of agentic AI will go further, arranging information, services, and actions around what the employee is trying to achieve.
For decades, employees have largely been expected to navigate organizational structures through applications. In an intelligent digital workplace, the organization will increasingly arrange the appropriate information, services, and actions around the person instead.
At Unite 26, Microsoft research leader Alexia Cambon introduced an interesting distinction between information work and intelligence work.
She argued that employees currently pay a substantial "consumption tax" simply to participate in work: reading, searching, processing, and making sense of information before they can act.
AI agents could take on more of that burden. But they introduce a different challenge: what Cambon described as a "delegation tax," as employees learn to direct, supervise, and coordinate intelligent systems.
It's an important reminder that removing one form of complexity doesn't automatically eliminate complexity altogether.
For CIOs, the opportunity is to design experiences that genuinely reduce the effort required to move from intention to outcome.
But the broader architectural challenge involves connecting knowledge, identity, permissions, applications, and workflows in ways that support reliable action.
4. Governance Will Become The Foundation of AI Acceleration
The organizations that scale AI most effectively won't have fewer controls. They'll have better ones.
Governance is often presented as the tension at the heart of enterprise AI.
Business teams want to experiment. Technology leaders need to protect sensitive information, manage cybersecurity risk, and maintain control.
It's tempting to see those priorities as competing forces. But as AI becomes more autonomous, effective governance will be what allows organizations to move faster.
At Unite 26, Norman Nowak described how Deutsche Telekom learned this firsthand. An early HR chatbot showed how quickly answers can drift when information isn't managed consistently across employee experiences. His message: innovation needs clear ownership and centralized standards, especially when intelligent systems operate across a large, complex organization.
The same principle becomes even more important as agents move beyond providing answers.
An agent that retrieves information requires appropriate access controls. An agent that initiates transactions, updates records, or coordinates workflows requires additional safeguards. Who authorized the action? Which permissions apply? What happens when something goes wrong? And how can the organization explain or reverse an automated decision?
Gartner's 2027 trends list identifies AI agent identity as a strategic priority, reflecting the need to govern systems that can act independently and access enterprise resources.
Stale content is now a liability, not just clutter
Outdated content used to be an annoyance. Once AI is answering questions and taking action on people's behalf, an out-of-date policy becomes a risk. In 2027, ownership, review dates, and provenance (being able to trace an answer back to a trusted source) will become core AI controls, not housekeeping. Dentsu's Neil Marshall summed it up at Unite 26: "AI amplifies poor content as well as it amplifies good content."
From probabilistic to deterministic outcomes
Most enterprise AI today runs on probability. General-purpose assistants, from ChatGPT to the copilots built into every app, make a best guess from whatever data and tools they can reach. Even when you give them rules in plain language, the model decides how to apply them. So the rules will probably be followed, but nothing guarantees it. That works in a simple organization. In a large, complex one, that's exactly where trust breaks down.
In 2027, the shift will be to deterministic orchestration, meaning AI that follows set rules every time. The organization decides the outcomes, limits, and policies it needs. The platform then enforces those rules around the model. It controls which sources the AI can use, filters results by permissions, and picks the right agent for each request, based on what it knows about the person asking.
The AI still works out how to get there, but always within rules that don't bend. And those guardrails will go beyond identity and permissions to cover brand, tone of voice, and policy too.
"AI moves at the pace of trust, and trust needs predictable outcomes. Large organizations can't run on rules the model will probably follow. They need rules that always hold." - Sam Hassani, CTO, Unily
This suggests that traditional approaches to application governance will need to evolve. CIOs will increasingly need to manage not just what systems can access, but what intelligent agents are authorized to do. That includes defining boundaries for autonomy, maintaining visibility into agent activity, and establishing appropriate human oversight.
Governance isn't the brake on AI. Done well, it's what gives organizations the confidence to accelerate.
In 2027, we expect this distinction to become central to the CIO agenda.
5. Purpose-Built Digital Experiences Will Define Adaptable Enterprise Architecture
The future of the digital workplace won't be one experience that tries to serve everyone. It will be many purpose-built experiences, assembled quickly on a foundation that stays consistent.
The pace of organizational change is already exposing the limits of digital experiences that take months to build and years to evolve.
An acquisition closes, and thousands of new colleagues need to understand policies, systems, and leadership by Monday.
A new operating model lands with requirements nobody planned for in last year's roadmap.
A workforce initiative needs its own mix of information, communication, and services, and it needs it now.
ASML knows how quickly this can happen. Hiring more than 11,000 people in two years was enough to overwhelm an intranet that had been built for a smaller, slower-moving organization.
Traditionally, responding to each change might require a dedicated development project, additional integrations, custom components, or another application. When organizational needs are changing continuously, that approach becomes increasingly difficult to sustain.
In 2027, we expect CIOs to place greater emphasis on technology environments that can be assembled, adapted, and retired as business requirements evolve.
Gartner's 2027 trends list identifies disposable applications as an emerging architectural concept: software portfolios designed to evolve and retire as business needs change, rather than becoming permanent additions to the technology estate.
The broader principle is directly relevant to the digital workplace. Rather than treating every new employee requirement as a long-term software development commitment, organizations will create purpose-built experiences using reusable services, trusted knowledge, and existing enterprise capabilities.
Imagine applying the same approach to an acquisition: a dedicated employee environment that brings together the relevant communications, policies, applications, and support services, and evolves with employees' needs as the integration progresses. The same principle applies to onboarding, organizational change, workforce development, and emerging business initiatives.
The same discipline will apply to emerging technologies themselves. Not every wave, from blockchain and web3 to the metaverse, will earn a permanent place in the estate. And just as DevOps changed how software is shipped, CIOs will expect employee experiences to be built, improved, and retired continuously.
Innovations such as Indi by Unily illustrate how creating purpose-built digital experiences will become faster and more accessible.
But speed alone isn't enough. The real opportunity is to combine flexibility with architectural consistency. Experiences may change rapidly, while the underlying standards for identity, integration, accessibility, information quality, and governance remain dependable.
6. The Frontline Will Become The Real Test of Every AI Strategy
If your AI strategy only works for someone at a laptop in Teams, it misses a large share of the workforce.
Much of the AI conversation still assumes a desk, a corporate laptop, and a full Microsoft 365 license. But nurses, drivers, store colleagues, cabin crew, and engineers don't work that way. Deskless employees need mobile-first access, content in the right language, and information relevant to their site and shift. Many have no corporate device at all, and their AI will often need to work alongside connected IoT equipment rather than a desktop.
Industry 4.0 has already put IoT sensors, edge computing, digital twins, and 5G technology on factory floors and in the field. Wireless networks reach the most remote sites.
Yet the engineer standing next to a digital twin of a pump still can't always get the right safety procedure on their phone. The gap isn't connectivity. It's the experience.
Unily and Gallup's research shows the cost of that gap. Frontline workers are less likely than their non-frontline colleagues to have a good digital culture (30% versus 36%), and more than twice as likely to say they never see recognition through digital communications (23% versus 11%).
The State of the Digital Workplace Experience
The organizations closing that gap are proving it can be done. Etihad Airways put just over 14,000 staff, around three-quarters of them frontline, onto a single app and retired mass email as its main channel for employee news. Within three months, 99% of employees had joined. Alshaya Group built shift-sized, five-minute learning modules for around 30,000 frontline colleagues, with a bilingual English and Arabic assistant that answers only from approved content.
In 2027, we expect the frontline to become the clearest test of whether an AI strategy is real. If AI can deliver a trusted, relevant answer to a deskless employee, in their language, on their device, mid-shift, it will work anywhere.
7. The AI Productivity Debate Will Move From Time Saved to Organizational Outcomes
An AI assistant saving an employee 30 minutes doesn't automatically mean the organization has created 30 minutes of value.
For CIOs, demonstrating the return on AI investment is becoming increasingly urgent.
Organizations are investing in licenses, infrastructure, data centers, integration, security, and training, often while existing technology budgets remain under pressure. Gartner's CIO Planning for 2027 research forecasts average IT budget growth of 3.7%, while funding for agentic AI is expected to increase by an average of 31.8%. However, just 13% of respondents reported significant value from AI tools so far.
That creates a difficult equation. The appetite for AI investment is growing faster than the resources available to support it.
Traditional measures of success aren't sufficient. Licenses deployed, employees activated, and prompts submitted can tell us whether technology is being used. They don't tell us whether work has improved.
Microsoft's 2026 Work Trend Index argues that organizations need to redesign work as AI agents take on more execution. Deloitte makes a similar distinction in Tech Trends 2026 emphasizing the importance of redesigning processes rather than simply automating existing ones.
At Unite 26, Alexia Cambon challenged the tendency to frame AI productivity primarily in terms of time savings. Her argument was that organizations are unlikely to run out of work. Instead, the nature of that work will change as people spend less time consuming information and more time directing intelligent systems, exercising judgment, and making decisions.
For CIOs, this means a more demanding definition of value.
Did AI reduce the time required to complete a process? Did it improve the quality of a decision? Did it remove unnecessary approvals or duplicated effort? Did employees gain capacity for work that matters more?
Measure understanding and action, not usage
The most meaningful measures form a closed loop. Did people see it? Did they understand it? Did they act on it? That loop will matter far more in 2027 than prompts, logins, or "hours saved" estimates.
Productivity and resilience from the same foundation
There's another outcome that rarely makes the business case until it's needed. The same foundation that saves employees time day to day is how organizations reach everyone, fast, in a crisis. When a site goes down or a safety incident unfolds, the ability to get a trusted message to the right people, on the right device, and confirm they've acted on it, is mission-critical resilience.
These questions require closer collaboration with business leaders. They also require a willingness to challenge investments that generate impressive adoption figures without delivering meaningful improvements.
8. The CIO's Next Challenge Will Be Orchestrating People, Knowledge, and Technology
The intelligent enterprise won't emerge from technology decisions made in isolation. It will require a different approach to organizational leadership.
The traditional boundaries between information technology (IT), HR, communications, and business operations are becoming harder to maintain.
IT may own the technology decisions, but business teams own much of the information those platforms depend on. HR understands employee needs and workforce processes. Internal communications helps create shared understanding. Operations knows where work slows down.
AI increasingly connects all of them.
At Unite 26, Norman Nowak described how Deutsche Telekom had previously struggled with shared responsibility for its intranet. HR, IT, and communications were all involved, but there was no clear decision-maker. The organization subsequently established clearer ownership while maintaining cross-functional collaboration.
It's a practical example of a challenge many CIOs will recognize. A digital workplace can have excellent technology and still fail if nobody has clear accountability for the overall employee experience.
Forrester analyst Cheryl McKinnon reinforced the importance of bringing together technology, information, and employee context during the same Unite 26 discussion. Her perspective highlights why an intelligent workplace cannot be reduced to a single platform. It's the combination of reliable systems, relevant information, employee data and data analytics, integration, and the organizational decisions that make them work together.
Unily and Gallup's research adds another dimension. Only 20% of U.S. employees have a good digital workplace experience, combining effective access to technology and information with digital experiences that support connection, alignment, and growth. At organizations with more than 10,000 employees, that falls to just 15%. Employees with good experiences were 3.1 times as likely to be engaged as those with poor experiences, and almost twice as likely to trust the information AI tools provide for their work (47% versus 26%).
The findings establish associations, not causation. But they reinforce why CIOs should consider the human experience of technology alongside traditional performance indicators.
In 2027, we expect more technology leaders to recognize that enterprise AI success depends on cross-functional operating models as much as technical architecture. That means establishing clear ownership, shared objectives, and ways of working that connect technology decisions to employee and business outcomes.
What does this mean for CIOs in 2027?
The CIO agenda is becoming more complicated.
AI capabilities are accelerating. Business expectations are rising. Technology environments are becoming harder to govern, and organizations are being asked to demonstrate value faster than ever.
But the IT trends that will matter most in 2027 aren't about another platform, application, or intelligent agent. It's making the capabilities organizations already have work together more effectively.
That starts with context. Employees and intelligent systems need access to reliable information, relevant context, and a clear understanding of what matters. A system of record for enterprise knowledge provides the trusted foundation. Knowing who each employee is, what they do, and where they work makes that knowledge relevant.
It continues with action. Technology should make it easier to complete work, connecting information with the applications, services, and processes that turn intention into outcomes.
And it depends on governance. Organizations need the confidence to scale intelligent capabilities without sacrificing security, accountability, or control, with rules that hold every time, not just most of the time.
These aren't separate priorities. They're interconnected foundations for an intelligent enterprise.
Knowledge creates understanding. Understanding enables action. Governance makes both trustworthy and scalable.
For CIOs, the opportunity is to bring these foundations together around the people using them, all of them, from head office to the frontline.
Because the organizations that succeed in 2027 won't be the ones with the most AI.
They'll be the ones that turn AI capability into human understanding, understanding into action, and action into measurable enterprise value.
Explore the Research and Conversations Behind These Trends
The State of the Digital Workplace Experience: Unily and Gallup
Discover how digital workplace experiences relate to employee engagement, well-being, and retention in new research involving 2,305 U.S. workers.
Access the insights. Move work forward.
Unite 26 On Demand
Hear directly from the leaders navigating workplace transformation, including stories from The Adecco Group, Etihad Airways, Alshaya, and more on Unite On-Demand.
Explore the trends for your function
Discover what the changing employee experience means for HR, internal communications, and IT leaders.
Get started. Get your personalized demo.
