Nudges to Knowledge: Transforming Workplace Learning with AI

Removing Friction Through Smart AI Experience Design

Mar 20, 2025

Executive Summary
The workplace learning landscape is undergoing a fundamental shift. Companies are abandoning traditional learning platform systems in favor of AI-driven approaches that deliver knowledge precisely when employees need it.
By combining modern push and pull tech, we see "smart nudges" for behavioral guidance combined with just-in-time knowledge from AI systems to create a force multiplier. Start implementations with high-value targets like sales enablement, manager effectiveness, and onboarding acceleration. These quick wins build momentum and create champions for broader adoption.

Today's workplaces face two big problems that hurt the bottom line.

People aren't working together as well as they used to, and they're not getting the knowledge they need. Employee engagement is down with only 29% of workers satisfied with their work last year (2024) - down from 36% in 2021 (Gartner).

Meanwhile, nearly 60% of employees report inadequate on-the-job coaching.

We have declining collaboration with inadequate knowledge transfer.

Traditional approaches to learning aren’t working. Training sessions pull people away from their real work. Learning platforms require yet another login to remember. Knowledge bases force people to hunt for answers they need right now. None of these fit into how people actually work each day.

There has to be a better way. And there is.

In this report, I'll share two tools that offer a fresh approach:

Smart Nudges: AI-powered behavioral intervention tools (“NudgeTech”) that reduce friction in workplace collaboration through timely, contextual nudges
Just-In-Time Learning: LLM-Driven Learning Platforms (AI) that reduce friction in knowledge acquisition through personalized, "push"-based content delivery

Together, these technologies represent a cohesive solution to the learning friction problem. Personalized delivery provides the right guidance, to the right person, at the right time, in the right context.

Early users see workers getting 4% more done. For sales teams, that's about $2 million more yearly sales per 100 reps. Team happiness jumps by almost a third. New hires get up to speed 40% faster.

The best part is that for every dollar spent, companies get about four dollars back. This is rare for training programs.

Adoption Rates: AI vs Traditional Learning Models - Sean Wood AI

Hidden Productivity Drains

The Cost of Workplace Friction

I worked with a big tech firm that couldn't figure out why their training wasn't helping workers do better. People finished the courses. The content looked good. Yet somehow, they weren't using what they'd learned.

The reason was friction in the learning experience. People didn’t have access when they needed it most.

We watched employees' daily work and found that knowledge and guidance was consistently unavailable at the moment of need. When a sales rep needed to know about a rival's new product during a client call, that info was buried in a report on the company website. When a team leader needed advice on a tough talk with a worker, that guidance was in a course they took months ago.

The business paid a heavy price. Teams made decisions 45% slower when they couldn't find info easily. Mistakes rose by a third when people worked without good guidance. Customers grew unhappy when workers couldn't share knowledge well.

In financial terms, these workplace friction points cost them about 3% of operating expenses annually. For a company their size, that translated to over $40M a year.

Why is this problem getting worse now? Three factors are happening a once to create a perfect storm:

First, teams now include people with vastly different backgrounds, ages, and work styles. No single approach works for everyone.
Second, work models have fragmented beyond recognition. Some people are remote, some hybrid, some in-office. The days of everyone absorbing the same information in the same way are long gone.
Third, the pace of change has accelerated dramatically. Markets that once shifted every few years now change in weeks. What you know gets outdated quicker.

These factors create different kinds of friction, each with its own cost:

Time (Temporal) friction hits when help isn't there when you need it. You have a tough client call at 2pm, but the training on handling them is next Tuesday. The cost? On average, a 3-day gap between needing and getting knowledge. Each incident costs about $380 in wasted work and slow choices.

Access friction happen when information exists but is hard to find. I watched a project leader spend 47 minutes searching through three systems to find a simple form. Think about that happening to every worker, every day.
Cognitive friction happens when guidance is too generic to be useful. That excellent training on "giving feedback" suddenly seems incomplete when you're facing a high-performing but disruptive team member with unique cultural expectations.
Engagement friction happen when content doesn't match how people like to learn. Some leaders still wonder why no one reads their 40-page PDF plans.d.

Behavior by Design

How AI Nudges Drive Measurable Performance Gains

Remember when Gmail started suggesting better ways to phrase your emails? That simple feature fundamentally changed how millions of people write. It's a perfect example of a digital "nudge" - a suggestion that guides behavior without restricting freedom of choice.

Smart Nudges bring this idea into the workplace, using AI to deliver personalized, contextual nudges to employees at key decision moments. It's built on Richard Thaler and Cass Sunstein's Nobel Prize-winning behavioral science framework - the idea that small adjustments in how choices are presented can significantly influence decisions while preserving autonomy.

What makes this different is how well it fits into how people already work. Instead of making people search for guidance, Smart Nudges bring tips right to the tools they already use - Slack, Outlook, Teams, and more.

The Kraft Heinz Company saw this firsthand. Leaders who used AI nudges got much higher ratings from their teams. Those who ignored the nudges were 2.5 times more likely to need extra help. This saved them about $3.8 million in training costs in the first year - far beyond what they hoped for.

Not all nudge systems work equally well. Many early ones gave generic tips that ignored each person's needs. The newest ones use AI to tailor nudges based on how you communicate, your team, and cultural factors. The difference is night and day.

Just-in-Time Knowledge:

How LLMs Are Replacing Traditional Learning Platforms

For nearly 20 years, Learning Management System (LMS), Content Management Systems (CMS), and Learning Experience Platforms (LXP) were the center of company training. But they were built on the idea that workers would actively seek out knowledge when they need it. But the reality is they don't. At least not consistently.

The traditional "pull" model of learning creates too much friction. Workers must:

Realize they have a knowledge gap
Remember the learning platform exists
Log into a separate system
Search for relevant content
Try to apply general guidance to their specific situation

That's too many steps. Companies waste about $4,200 per worker yearly on learning that never gets used. More than three-quarters of training budgets deliver less than 15% actual business value.

Large Language Models (LLMs) are enabling something fundamentally different: a "push" model where knowledge finds the employee, not the other way around.

I saw this transformation at a financial firm last quarter. They'd spent years building an extensive knowledge base about their products, regulations, and procedures. The problem? Nobody used it. When I asked their advisors why, the answer was telling: "It's faster to ask a coworker than to find the info in the system."

Their solution combined LLMs with smart delivery mechanisms. They implemented an AI platform that integrated directly with their CRM. Now, when an advisor opens a client account, the LLM analyzes the client's portfolio, life stage, and recent market events, then automatically delivers relevant information through the tools advisors already use. Key knowledge appears right in their workflow - no searching needed.

The impact was immediate and measurable:

Client meetings became more productive because advisors were better prepared
Compliance violations dropped by 72%
The help desk saw 58% fewer basic knowledge questions
Most tellingly, advisors who previously avoided the knowledge base became power users of the new system

This shift from "pull" to "push" learning fundamentally changes knowledge dynamics. The most valuable aspect of corporate training over the next five years won't be content creation - LLMs can generate that instantly. It will be effective delivery of mission-critical updates, AI upskilling resources, and behavioral nudges at exactly the right moment.

The financial impact is huge. A typical company with 10,000 workers can save about $18.7 million yearly in time no longer wasted finding info. That's not theory - that's real savings from cutting wasted time.

What makes this approach so powerful is that it uses existing systems. There's no new platform to learn, no separate app to open, no extra login to remember. Knowledge appears in the tools people already use every day.

One word of caution: implementing LLM-driven learning isn't just about the technology. While LLMs excel at generating personalized content quickly, human oversight remains essential for ensuring quality and compliance. Companies that simply layer LLMs on top of poorly organized information end up with sophisticated systems delivering mediocre content. Do the foundational work first.

The Competitive Landscape:

Balancing AI and Delivery

Here's where things get really interesting - and where I've seen the most dramatic results. The corporate learning space is evolving rapidly, with several key players emerging:

Internal IT teams building custom LLM solutions
Platforms like Sana and Glean focusing on knowledge management
Microsoft's Copilot integrating AI assistance across office tools
Specialized AI push platforms like Arist focus on delivery mechanisms

It’s not just having AI capability that makes it work, it's effective delivery. Smart Nudges and LLM-driven learning are powerful alone, but together they create multiplication, not addition.

Think of it this way: Smart Nudges help with how to work (behaviors), while AI learning helps with what to know (knowledge). Together, they create a support system that greatly reduces workplace friction.

The multiplier effect in action: Using either tech alone typically deliver 2.5-3.5x ROI, integrated implementations consistently achieve 5-7x ROI.

This teamwork creates several key benefits:

First, knowledge-backed nudges become far more effective. When product info changes, behavior guidance automatically reflects those changes without needing manual updates. This ensures consistency between what people know and how they apply it.
Second, behavior-reinforced learning greatly improves memory. We've known for decades that using knowledge is key to remembering it, but old learning approaches struggle with this. Integrated systems close the loop by prompting specific uses of newly acquired knowledge.
Third, sharing infrastructure cuts total costs by nearly half compared to using these techs separately.

Most companies use these tools in silos because different departments own different parts of the employee experience. That's a mistake. The real value comes from careful integration.

Strategic Implementation:

Shifting Your Focus from Content Creation to Delivery

As LLMs continue to automate content creation - generating personalized training, summarizing documents, creating quizzes - the focus for training teams must shift dramatically. The key differentiator won't be who has the best content (LLMs level that playing field), but who delivers that content most effectively.

If you're convinced these technologies could transform your company (and you should be), the natural question is: how do we implement them effectively?

From my experience, massive, digital transformation implementations that hope to solve every problem at once don’t work. Too many ambitious projects fail under their own weight.

The better approach is targeted, step-by-step, and focused on quick wins that build momentum. Here's my tested playbook:

Map your specific friction points. Don't assume you know where knowledge and guidance breakdowns occur - actually observe and measure them. I worked with a service firm that thought their knowledge sharing problems came from poor documentation. When we studied actual workflows, we found the real issue was that relevant knowledge was scattered across seven different systems with inconsistent taxonomies.

This discovery phase typically identifies $500K-$800K in immediate improvement opportunities before you implement any technology.

Next, identify high-value use cases with direct business impact. The best starting points are typically:

Sales enablement, particularly competitive positioning and objection handling
Manager effectiveness, especially for new or developing leaders
Cross-functional collaboration, particularly in matrix organizations
Onboarding acceleration for complex roles

Start small with a focused test. Identify a specific group, train them well, gather detailed feedback — test and refine before expanding. Pay close attention to workflow integration - that's where most implementations stumble.

I've seen companies rush deployment only to create new kinds of fricion they didn't foresee. Take time to get the user experience right.

As you expand, maintain a relentless focus on business outcomes. Gradually extend to additional use cases and employee groups. Communicate early successes to build momentum and address concerns. The most successful implementations share one trait: they measure impact in terms that matter to the business, not just learning metrics.

Create learning loops with regular reviews that assess impact and refine your approach. The best implementations evolve based on real usage patterns and worker input.

Finally, establish clear oversight around content quality and ethical use. These AI systems are only as good as the info they access and the rules they follow. Without proper oversight, you risk amplifying existing biases or delivering inconsistent guidance.

This last point deserves special emphasis. While these tools have enormous potential, they also raise real ethical questions about privacy, freedom, and potential surveillance. The companies that implement most successfully address these concerns openly and honestly.

Trust me on this: your workers will embrace systems that truly help them perform better, but they'll actively fight anything that feels like digital micromanagement. The line between helpful guidance and intrusive oversight can be thin - be thoughtful about where you draw it.

The Competitive Advantage of Frictionless Work

We're at an inflection point in how organizations support employee performance. The old models - traditional LMS platforms, static knowledge bases, periodic training - simply can't keep pace with today's dynamic work environments. The future belongs to companies that master the shift from "pull" to "push" learning.

Looking ahead five years, we'll see a dramatically different learning landscape:

Traditional LMS, CMS and LXP platforms will be largely replaced by internal LLM solutions
Companies will retain only minimal "pull" systems for compliance requirements
The competitive advantage will belong to those who master "push" delivery mechanisms
Training teams will shift from content creation to delivery optimization

This isn't just incremental improvement - it's transformation. When implemented well, these technologies don't just slightly improve existing processes - they fundamentally change how work happens. Decisions accelerate. Collaboration improves. Knowledge flows freely. The entire organization becomes more adaptable and responsive.

In every industry I’ve seen, companies that systematically reduce workplace friction outperform those that don't. They make better decisions faster, respond to market changes more effectively, and create places where talented people can contribute their best.

The technology is ready now. Forward-thinking companies are already using these approaches and seeing significant value:

Productivity gains of 3.8-5.2% across various employee segments
Revenue increases about $20K per salesperson annually
Nearly one-third improvement in workplace collaboration metrics
Onboarding speeds up by 2.5 months for complex roles
Return on investment approaching 4× initial outlay in the first year alone

But the most compelling evidence comes from the workers themselves. I recently talked with a manager who said it well: "For the first time in my career, I feel like the company's systems are actually helping me do my job better rather than just creating more work."

That's the promise of these technologies - better metrics for people and better work experiences to drive business results.

The question is when, not if, you'll start removing workplace friction.

This report draws on my research and consulting work with multiple organizations implementing AI-powered performance support solutions. It's intended to provide strategic guidance for executives considering how these technologies can transform workplace effectiveness. Various AI tools were used for research and analysis, with human fact-checking and triangulated LLM outputs for accuracy and trustworthiness.

Sources: