Prompt fatigue is draining teams. Read how ambient agents and zero UI eliminate workflow friction while keeping humans in control.
Sydney Scott
Editorial Strategist, AI
Workday
A doctor talks with a patient, never once picking up a notebook or tablet to take notes. As they chat, an agent listens in the background, organizes clinical notes, and enters data straight into the electronic health record system. No typing, no searching menus. The doctor is able to give the patient their undivided attention.
In the ongoing discussion about what AI can accomplish, we often forget to talk about major improvements to the traditional user experience. Tech is moving past traditional graphical user interface (GUI) buttons and chat boxes toward ambient agents running in the background.
This is Zero UI. And it could be the answer to cutting workflow friction.
Employees lose up to seven hours per week to fragmented systems.
Report
Many businesses today are running on fragmented systems. A report from Okta shows the average enterprise runs 101 distinct SaaS applications, while larger firms average 131.
In a survey of over 3,700 leaders and employees, WalkMe found 60% waste time juggling multiple applications just to complete one task. Workday research shows employees lose up to seven hours per week to sprawl.
This constant juggling takes its toll. Research found that when workers are interrupted, it takes 23 minutes to get back on track. And software clutter costs companies $21 million, on average, in unused or redundant SaaS licenses.
Companies are trying to fix this, using chatbots to make finding information easier. But, this has introduced a new problem—prompt fatigue, mental exhaustion from constantly typing detailed context and refining AI answers.
To truly eliminate friction, companies need to start looking for software that quietly works in the background.
In May, during the Sana AI Summit, Sana CEO and Workday Chief AI Officer Joel Hellermark shared a vision of the kind of agents we’ll see in the future:
“Today, all of our agents are sleeping until we start talking to them. I want agents that never sleep, that run 24/7 in the background, checking compliance, updating the rules, developing new workflows on my behalf. This is what ambient agents enable.”
To understand where enterprise technology is going, leaders must think of digital systems in two clear parts: a perception layer and an action layer.
Ambient agents handle perception, watching how businesses work 24/7. They listen to ambient audio, scan calendar schedules, and read updated documents. They also check application states and review infrastructure logs. These tools gather context and lower mental burden without requiring screen interactions.
Agentic AI acts as the execution motor. It takes the context gathered by ambient sensors and builds multi-step plans. Then, it calls an API and runs background actions across corporate databases.
The two systems need each other to get work done. Agentic AI without ambient sensing forces users right back into the tiresome working of manually adding context and details. But, ambient AI without agentic tools just creates notifications without finishing the job.
Together, they create a continuous loop with ambient agents capturing context and passing structured goals to agentic tools to finish the task.
“I want agents that never sleep, that run 24/7 in the background, checking compliance, updating the rules, developing new workflows on my behalf. This is what ambient agents enable.”
Joel Hellermark,
Workday Chief AI Officer and Sana CEO
There are clear benefits to ambient agents.
In healthcare, ambient AI listens to doctor-patient conversations in real time, converting dialogue into structured notes. This saves doctors 13 to 22 minutes daily and overall documentation time falls by 20% to 30%. In fact, The Permanente Medical Group’s ambient AI scribes have saved over 15,000 hours on documentation and, in some cases, led to less burnout.
Sales operations have shown similar improvements when using ambient agents. Background agents passively monitor deal emails, call transcripts, and customer interactions to automatically update CRM pipeline stages and draft status reports. Cutting out manual data entry and automating daily tasks helps sales teams close deals 40% faster.
In IT, incident response moves fast as ambient agents gather operational anomalies in real time, instantly showing summaries of root causes when an engineer receives an assignment. This cuts meant time to resolution by 20% to 40% by eliminating manual log searches.
And supply chain logistics see speed gains with employees wearing smart audio gear and using visual sensors that track inventory picking automatically. Employees then confirm choices with quick voice commands. This process boosts warehouse output by 20% to 40% while cutting entry errors by over 30%.
Companies love the promise of ambient agents, but invisible software running in the background can introduce big risks without the right guardrails in place. Employees can feel like they are being constantly surveilled leading to a collapse in employee trust. Users can also trust automated outputs too much until it makes a mistake.
To fix this issue, product teams must keep humans not just in the loop but in control.
UI designers can use clear confidence bands to guide worker actions, skipping fake percentages and tying trust levels directly to what the interface allows. High confidence puts the focus on "accept" for a fast single click. Medium confidence shows a quick preview before the system allows an approval. And low confidence limits the system to a "copy only" mode, where employees must edit the draft before saving the work.
Teams must also make mistakes easy to fix. Apps should apply small changes and offer a clear "undo" button to stop or fix automated choices.
Source tags should be used to show where data comes from. Hover cards should explain what the LLM generated and what came from source documents. Brief, inline hints should give updates without blocking the screen. Clear physical rules—like hardware status lights that turn on whenever mics record sound—can calm privacy fears. Finally, edge devices should process data locally to keep raw files off cloud servers and strict zero-data-retention policies wipe the data away.
These boundaries build trust before companies launch new tools.
Companies love the promise of ambient agents, but invisible software running in the background can introduce big risks without the right guardrails in place.
Flashy new dashboards and chat windows asking for better text prompts aren’t going to eliminate workflow friction, but software that gets entirely out of the way will.
By mastering the user experience, businesses can break the cycle of software sprawl and rebuild trust. Employees can reclaim their focus and their time.
Software running quietly in the background must remain transparent, reversible, and secure. Product leaders who build strong human-in-the-loop guardrails, clear source tags, and local privacy controls will come out ahead.
Invisible software is already reshaping enterprise work and the companies that balance background automation with human control will take the prize.
Report
A gap between effort and impact is the real ROI problem. Discover the solution.