Designing Digital Coaches for the Flow of Work

Insights from a recent Digital Learning Institute (DLI) webinar with Chris King

Our first public webinar of 2026 at the Digital Learning Institute kicked off January’s Continuing Education theme: Learning Design. In the session, Chris King (former Executive Director of the Five Moments of Need Academy) challenged a familiar default in L&D and higher education:

If learning doesn’t change what people do, what’s the point?

What followed was a practical, performance-first way to think about learning design. One that zooms out from courses and content and zooms in on workflow, application, and support in the moment of need. This blog captures the core ideas from Chris’s talk, rewritten as a guide you can use.

Most of us have seen it: a course launches, attendance looks fine, completion rates are decent… and then nothing changes on the job.

Chris named two common reasons:

• The training wasn’t tied to business outcomes.
  If the learning isn’t clearly connected to what the organization is trying to improve, it becomes activity rather than impact.

• Learners understood the content, but couldn’t apply it.
  This is the classic transfer gap: the space between “I get it” and “I can do it when it matters.”

To underline the difference between knowledge and application, Chris shared a story from a government agency. The organization had plenty of training... webinars, e-learning, in-person sessions, and (of course) binders. But the work itself involved complex workflows and deadlines. When someone missed one deadline, an issue escalated into formal arbitration and resulted in a $1 million fine.

The takeaway was blunt: training can’t fix every performance problem, especially when the problem is about execution in real conditions like time pressure, complexity, nuance, and consequence.

Chris argued that a performance-first approach requires a shift in how learning professionals see their role:

• From order taker → to partner → to performance partner

• From “delivering content” → to solving problems

• From measuring completion → to measuring impact on outcomes

This is also why he deliberately uses the term performers instead of learners.

It’s not just semantics. Language shapes priorities. Calling people “performers” keeps the focus on what the organization actually needs: competence in real work, not just understanding in a classroom.

Many learning teams excel in two of the Five Moments of Need:

• New: learning something for the first time

• More: deepening knowledge

That’s our traditional comfort zone: courses, modules, instruction.

But Chris emphasized that the real performance breakthroughs often happen in the other three moments:

• Apply: doing the thing in context

• Solve: when something goes wrong or doesn’t work as expected

• Change: when processes, tools, or expectations shift

These are the moments where people feel friction and where learning teams often step back and unintentionally say, “That’s informal learning… someone else will handle it.”

Chris’s point: you don’t have that luxury anymore if you’re serious about performance.

So what does help close the transfer gap?

Chris’s answer: digital coaches a modern expression of performance support designed for the flow of work.

He traced this back to Gloria Gery’s foundational work on Electronic Performance Support Systems (EPSS) and then updated it with a more current lens from Bob Mosher and Conrad Gottfredson:

Performance support should be immediately available, intuitive, and intentionally tailored, delivered at the moment of need.

And crucially: a digital coach is not “a piece of software.”
It’s a set of design principles that can be implemented using the tools you already have.

Chris even admitted he has built performance support tools in PowerPoint, PDFs, wiki tools, and Confluence... imperfect, but effective when grounded in the right methodology.

One of the most useful models from the webinar was the pyramid Chris uses to structure a digital coach. The idea is simple: support people closest to the work first.

1. Context
   Make it easy to find the right support in the right situation. Chris used an aspirational benchmark: two clicks, ten seconds to reach task-level help.

2. Tasks (task-level support)
   Support tasks at two levels:

   • Quick steps: a checklist for experienced performers (fast reminders)

   • Detailed steps: the nuance and “edge cases” for newer performers or infrequent tasks

3. Supporting knowledge
   Curated “just enough” explanations that help someone complete the task correctly without turning the coach into a textbook.

1. Reference resources (policies, templates, procedures)

2. Learning resources (courses, modules, deep links, not LMS wandering)

3. People resources (peer support, help desk, communities of practice)

As you move down the pyramid, you move further away from the workflow which is fine, as long as the performer gets to choose how deep to go based on time, risk, and need.

Copyright © 2024 by APPLY Synergies, LLC All Rights Reserved. Used with permission.

A standout discussion in the Q&A was a familiar frustration:
“We have learning materials in the LMS. It’s easy to access. People still don’t use it.”

Chris’s response wasn’t anti-LMS. It was anti-dumping.

If people land in a giant repository with hundreds of items, they will:

• forget it exists

• feel overwhelmed

• search YouTube instead

• do it their own way (not your organization’s way)

Performance support narrows the choice architecture:
“Here are the six things you need for this task, right now.”

Access to the information is not enough you add value when it’s curated, relevant, and easy to use.

To design support for the flow of work, you need to understand the work at the right level.

Chris outlined a method called Rapid Workflow Analysis (RWA). It’s similar to job task analysis, but deliberately faster and more focused.

In RWA, you identify:

• Tasks: collections of steps that produce a specific performance outcome

• Supporting knowledge: what someone needs to know/understand to do those tasks

Then you organize tasks into processes and create a workflow map so performers can see how the work fits together.

Chris stressed a simple but powerful distinction:

• Processes = collections of tasks

• Tasks = collections of steps that achieve an outcome

• Steps = actions that don’t stand alone as outcomes

He even had participants do a quick exercise: list everything your family does to plan a vacation, then notice what’s actually a task versus a step. It’s a great way to practice how easily we drift into the wrong level of detail.

Chris anchored the design decisions in two cognitive principles:

1) Chunking

A long list of tasks is intimidating and hard to remember.
Grouping tasks into meaningful “process chunks” makes the workflow easier to hold in working memory.

Chris referenced the classic guideline: 7 ± 2 items per chunk as a rough target.

2) Encoding and retrieval

People retrieve information in the way they encoded it.

If we teach in tidy topic sequences, but the job is messy and nonlinear, we create a mismatch and the learner has to do heavy mental reordering under pressure. That increases cognitive load and reduces performance.

So instead of teaching topics about the work, Chris argued we should teach how the work actually flows, using workflow maps as the organizing structure.

One of the most practical parts of the webinar was how to prioritize.

Chris described using a Critical Impact of Failure scale (it’s a 7-point scale measuring risk to the organization, sometimes referred to as a CIF scale). The question is:

If someone fails at this task, what happens?

• Low impact (minimal/moderate): recoverable, safe to learn through trial-and-error

  → keep it in the digital coach (don’t spend classroom time)

• High impact (significant/catastrophic): costly, risky, reputational, legal, safety-critical

  → prioritise in training, simulations, provide expert support

This is a clean way to stop trying to teach everything and start designing an ecosystem where training and workflow support complement each other.

To close, Chris shared a consulting-firm case study: three global centres trained support staff to build slide decks, but training differed across locations and average time-to-competency was 18 months and many people left right as they became competent.

The goal was to cut time-to-competency by 50% (18 → 9 months).
The outcome exceeded expectations:

• Year 1: hit 9 months

• Year 2: reduced to 5 months

• plus improvements in retention, decision-making, productivity, and reduced oversight

The key wasn’t “more training.”
It was workflow support + targeted training for high-risk tasks - this is a Five Moments of Need solution in practice.

If you’re designing learning in 2026, the question isn’t “How do we build another course?”

It’s:

• What does good performance look like in the workflow?

• Where do people struggle to apply what they know?

• What support would help in the moment... fast, contextual, trusted?

• Which tasks are high-risk enough to justify structured training time?

Digital coaches are one answer but the bigger shift is this: design for doing, not just knowing.