Increasing AI adoption with agents built to serve ALL employees

The pattern is remarkably consistent across industries: executive enthusiasm at the top, isolated pockets of experimentation in the middle and stalled adoption everywhere else. Despite billions in AI investment, only 5% of firms worldwide have achieved AI value at scale, according to Boston Consulting Group. A 2025 UKG global study found that just 38% of frontline workers use AI in their daily roles — a stark reminder that technology alone doesn’t drive transformation.

The real obstacle isn’t capability. It’s trust.

Adoption breaks down when employees feel like AI is happening to them, not for them. They want to know: Will this replace me? Who controls it? What’s in it for me? Until those questions are answered — through actions, not just messaging — enterprise AI will keep stalling at the pilot stage.

At UKG, we made a deliberate choice that changed everything: We started with an AI agent built to serve everyone.

A case study: The agent for all

In October 2025, UKG launched a global brand refresh — a pivotal moment to reintroduce ourselves to the market and to our more than 14,000 employees who serve 80,000+ customers worldwide. The challenge: How do you get an entire global workforce aligned on a new brand identity, quickly and consistently?

We built the UKG Brand Communicator, an AI agent designed to help every employee — not just marketing or communications — apply our new brand voice across everything they write, whether a customer email, an internal memo, or a social post.

This wasn’t a theoretical proof of concept. A cross-functional team built it, stress-tested it, broke it, refined it and deployed it. The agent was trained extensively on our new brand: what we say, how we say it and — just as critically — what we don’t. It reviews drafted messages, suggests tone adjustments and rewrites content to be clear, warm and people-focused.

The results were immediate. In the first 60 days:

• ~7,300 employee sessions
• ~13,000 AI-assisted rewrites
• ~1,500 hours saved — redirected to higher-value work and customer service

Why “agents for all” work

The Brand Communicator succeeded because of deliberate design choices that most AI rollouts skip:

• Low risk, high relevance. Brand guidance is something nearly every employee needs. It’s a safe sandbox to learn — people can experiment without fear of breaking a critical system or exposing sensitive data.
• Built-in simplicity. Employees didn’t need to learn prompt engineering. The agent was structured and guided from day one, meeting people where they are — not where technologists wish they were.
• Psychological safety. Pre-configured guardrails give employees permission to try, fail and try again without consequences. That experimentation loop is exactly what turns first-time AI users into daily users.

One lesson we learned the hard way: If it’s not in the workflow, it won’t stick. Sidebars, sandboxes and innovation labs are useful for discovery, but they don’t drive sustained adoption. The agent has to be the path of least resistance to getting the job done. It’s not just about launching. It’s about landing.

From one agent to an AI-native organization

The Brand Communicator was the first domino for UKG. By making AI useful — immediately, tangibly — for thousands of employees across functions, it converted curiosity into habit. It lowered the psychological barrier for every AI initiative that followed.

Today, 80% of UKG employees use AI in their daily workflows. We have more than 11,500 agents built by employees, for employees, generating approximately 155,000 agent-supported actions per month and saving 24,000 hours monthly. This is what human-AI collaboration looks like at scale.

That kind of adoption doesn’t happen by accident. Three systems made it possible:

1. An AI hub and idea-to-implementation (I-2-I) framework. Our internal AI Hub is the centralized place where employees explore tools, submit ideas, collaborate and experiment. It prevents duplication, surfaces promising work for scale and feeds governance. Functional champions shepherd ideas from concept to production, keeping momentum without chaos.
2. A portfolio model for experimentation. We run AI like a venture capital portfolio. Tier one is scale — prioritized use cases with clear ROI and defined outcomes. Tier two is growth — building capabilities that don’t yet exist, always with the customer at the center. Tier three is exploration — time-boxed, 90-day pilots and AI Demo Days that help us quickly decide what to grow, pivot, or archive. Velocity with discipline.
3. Lightweight, principled governance. Guardrails are baked into the flow of work — not bolted on at the end. Security, privacy, legal checkpoints and a risk checklist are standard. We route teams toward validated enterprise tools and away from ad hoc point solutions. When teams bypass the guardrails, we restrict access — not to punish, but to protect the trust we’ve worked hard to build.

The bottom line

Trust is the invisible infrastructure of AI adoption. It’s built through transparency about intent, honest conversations about job impact, visible upskilling opportunities and letting employees see their peers genuinely benefit. When those conditions exist, adoption doesn’t need to be pushed — it pulls itself forward.

If you want AI to scale across your organization, start simple and start broad. Choose problems that are safe, relevant across roles, and immediately useful. Embed guardrails from the beginning. Measure what actually changes how work gets done — not just what gets launched.

Do that, and AI stops being a mandate. It becomes how your organization works.

This article is published as part of the Foundry Expert Contributor Network.
Want to join?