Feature
White Paper
The Future of Leadership Development: AI-Driven Manager Enablement
As agentic AI enters the mainstream, it's bringing with it a new version of an old problem: AI agent, chatbot and Copilot sprawl. Gartner predicts that by 2028 the average Fortune 500 company will have more than 150,000 agents in operation, causing “significant agent sprawl, IT complexity and management challenges.” And yet only 13% of businesses feel they have the right governance in place to manage AI agents.
Table of Contents
- A New Variation of an Old Problem
- The Rapid Rise of Agents and Bots
- Isolated Agent Adoption Adds to Agent Sprawl
- Agent Governance vs. Agent Adoption
- The Battle for AI Agent Orchestration
- The Wider AI Agent Governance Required
- Preparing for Agent and Bot Sprawl
A New Variation of an Old Problem
The problem will sound familiar to anyone who has worked in the digital workplace or digital employee experience (DEX) space in recent years. Lack of governance has caused sprawl across the digital workplace, whether through content bloat, site sprawl or app overload.
App sprawl in particular has been a recurring culprit behind bad DEX. Research suggests that, on average, employees use 11 apps or more per working day, as they switch multiple times across tools to complete tasks or respond to the ping of a notification. The result is context switching, fragmented user processes, inconsistent interfaces and reduced productivity.
Are we in danger of adding to this fragmented DEX with with the addition of multiple AI agents?
“How many agents is the right number? The technology is too immature for one agent to encompass the whole digital workplace,” said James Robertson, founder of Step Two and a DEX veteran. “But when a hundred bots become a thousand, organizations risk fragmenting the employee experience and adding, rather than reducing, complexity.”
The Rapid Rise of Agents and Bots
While it’s early days for introducing AI agents at scale, signs suggest the rise in the number of agents will be fast.
“Agentic AI is the hot theme of the day,” said Ross Dawson, futurist and strategy advisor. “Many organizations are experimenting with agentic AI, with a smaller number introducing them more broadly across the workplace.” Early adopters are already applying agents in areas such as software development, while others are close behind.
Roberston also sees agent adoption across lines of business. “We’re seeing the fastest growth at the margins, with local business areas spinning up small-scale agents rather than big, strategic solutions,” he said. Whether this is where AI delivers the greatest value is yet to be seen, he added.
Isolated Agent Adoption Adds to Agent Sprawl
Gartner’s estimates of hundreds of thousands of agents running in large enterprises within the next 18 months heightens the need for orchestration, governance and an aligned deployment approach to avoid sprawl’s detrimental effects.
But that looks to be a tall order, and something IT leaders are worried about: Recent Salesforce research indicates that 83% of IT leaders fear that agents will drive more complexity than value.
Part of the problem is the proliferation of AI features and agents in every class of enterprise software, such as AI-first intranets and AI-powered learning platforms.
“Like other periods of innovation, every enterprise vendor wants to ‘own’ the employee experience,” said Roberston. “So, every major product is pushing its agentic capabilities, which is great in isolation, but potentially terrible for employees unless deployment is guided by a clear vision of the desired digital employee experience.”
As teams increase their adoption of agents within the products they already use, they risk creating a multitude of standalone agents, with the potential for overlap, confusion and loss of value. Indeed, the Salesforce research also found that 50% of agents being deployed are standalone and disconnected, effectively running in silos.
“Implementing single agents for individual use only gives limited benefit,” said Dawson. “Agents have the greatest value when they’re interacting with other agents and sometimes human teams.” Simply setting up agents is unlikely to yield immediate value, he added.
Agent Governance vs. Agent Adoption
Moreover, while the window to add controls to control sprawl is still open, pressure to show ROI on AI investments may push organizations to take a light touch approach to governance.
The push for wide adoption of AI agents at the potential expense of DEX is consistent with some of the findings from Reworked’s 2026 State of the Digital Workplace, where only 19% of organizations have a DEX that is “frictionless and productive.”
“We can’t slow AI progress down, but we also can’t expect employees to keep up with half-baked products or features,” said Lesley Crook, a change management consultant and Microsoft MVP focusing on responsible AI adoption.
The Battle for AI Agent Orchestration
How should organizations start to control the proliferation of agents? One possibility is AI agent orchestration to help implement governance and controls to reduce sprawl.
Several tech vendors are rushing to release AI orchestration capabilities that provide a range of features for a more controlled and connected AI agent landscape. These include offerings from the major tech providers such as Microsoft's Agent 365, Google's Gemini Enterprise and ServiceNow's AI Control Tower. Even EXP and intranet software provider Unily has introduced AI orchestration features.
While each solution is different, they all provide specific control features to help reduce agent and bot sprawl and orchestrate multiple agents:
- A registry or inventory of all the agents running within an organization.
- Automated scanning to detect and update this inventory.
- Monitoring and reporting agent behavior to track value.
- Interfaces that build workflows and handoffs between agents, and establishing meta-agents to centralize the bot experience for employees.
- Integrations and connectors, including support for interoperability, such as the MCP protocol.
- Other governance aspects such as access control, security and policy enforcement, and lifecycle management.
Orchestration may reduce complexity for employees by reducing the number of agents they need to interact with, or by creating a chat interface as the starting point for users that hands them off to the right bot, agent or human. New offerings, such as ServiceNow’s EmployeeWorks, are being positioned as a “conversational front door for the enterprise.” However, these products are all new and seem unlikely to solve immediate proliferation and sprawl issues.
The Wider AI Agent Governance Required
As with other aspects of the digital workplace, AI agents need governance to deliver value while supporting DEX. Governance needs to cover several elements.
Webinar
Jun
17
Engaged, But Not Convinced: What's Eroding Beneath Your Strongest Engagement Numbers
Perceptyx EX Benchmark '26: leadership trust, change confidence and conviction gap truths at large enterprises.
Register
Webinar
On demand
Retention Is a Communication Problem: New Data from 6,000+ Employees
This session gives leaders a clear diagnosis and practical path to treat retention as an infrastructure problem.
Watch Now
Webinar
On demand
Do Learning Programs Really Work? How to Turn Education Into Engagement In Healthcare
See how leaders are using learning programs to build trust with healthcare professionals and create measurable engagement.
Watch Now
Webinar
On demand
Fuel Your Front Line: Workforce Trends Shaping 2026
Get expert insight into the tools and tactics top frontline teams will need in 2026.
Watch Now
Webinar
On demand
Fix the Disconnect Between Projects and Profit
Your margins are hiding in disconnected data. Let's find them.
Watch Now
Webinar
On demand
AI in Customer Service: Faster Resolutions, Happier Customers
Don’t let rising demand burn out your team. See how to build a smarter, more resilient support org.
Watch Now
Clear Accountability
“Accountability needs to be defined,” said Dawson. “In any agentic AI implementation, you need clarity on who has approved the system, how it functions and any human approvals required by the system.”
Data Permissions
AI agents need specific data permissions, as agents with too much access break data governance structures.
Identifying Higher-Risk Agents
“Graduated governance will be crucial for AI agents and bots, distinguishing between local solutions that are low-risk and low-complexity with light-touch governance, to key agents that have higher risk and complexity that should be governed strategically,” Robertson said. This helps businesses avoid issues with strategic AI agents going straight from pilot into production without introducing suitable management practices and governance controls.
Approval Process
Like sites and apps sprawl that gets controlled through an approval process, creation and deployment of agents could be controlled. That may mean permitting staff to create agents only through an approval workflow or system to see whether a similar agent has been created, Crook said.
However, in practice, approval might be difficult to implement, especially when multiple applications and platforms deploy agents. The potential volume of local agents may also make this impractical.
Lifecycle Management
Management across the agent lifecycle, not just at creation, is also needed, from reviews at regular intervals to moving to a different governance model if they assume more importance, to decommissioning.
Guidance and Training
Any governance framework for AI agents that supports DEX also needs to consider change management and guidance for users, which reduces agent sprawl confusion. Some of this work may be in educating users about the basics, given that research from BCG found that only a third of employees understand how agents work.
“Understanding the scope of each agent or bot is critical, not just the content or tasks they cover, but how they can be easily explained to users so they know when and how to use each individual solution,” Dawson said.
Preparing for Agent and Bot Sprawl
The time to act is now. The possibility of AI agent creation exploding is real, particularly in Microsoft 365 organizations where the capability may already be launched across your workforce.
While orchestration capabilities are being launched and will help, wider governance measures are required — and required soon — to prevent a repeat of the content, site and app sprawl that has already hurt DEX.
Editor's Note: If agent sprawl, bad DEX and unclear accountability is your jam, read:
- Don't Let Your Company's Digital Tools Sabotage Employee Experience — It’s time for organizations to overhaul their approach to digital employee experience.
- Your AI Workflows Will Outlast the Leaders Who Approved Them — AI agents don't create accountability problems, they inherit them. When autonomous systems outlast the teams that built them, ownership disappears.
- The Teammate With No Manager: Who's Accountable for AI? — Why cross-functional leadership, accountability and employee experience design are central to AI’s success at work.