Resources

Most leaders already sense that friction exists in their organization. The hard problem is prioritization: across dozens of workflows and hundreds of potential improvement opportunities, where do you focus first?

The workflow scorecard is how we answer that question.

The idea is straightforward. For every workflow you want to understand, say handling a customer escalation, onboarding a new hire, or prioritizing an engineering backlog, you collect structured feedback from the workers running it: how much time the workflow takes, how much effort it demands relative to the outcome, where things get stuck, and what is helping.

That data feeds a scorecard for each workflow: a quantified picture of where the workflow stands across dimensions like time spent, friction level, tool effectiveness, and clarity of process. Every workflow gets a score, and because the methodology is consistent, workflows can be compared directly.

That is where the prioritization power comes from. Instead of relying on the loudest voice in the room or the most recent anecdote, you have evidence that one workflow carries high friction and high strategic importance and should come first, while another carries moderate friction but low impact and can wait.

The scorecard also tells you why a workflow is struggling, not just that it is. Root causes surface in the data, whether the issue is a tool that does not work as needed, an unclear process, an under-resourced role, or a data gap that creates constant rework. Each root cause routes to a different owner: the AI team, IT, Ops, or HR. That replaces siloed dashboards where each function sees a different slice of the same underlying problem.

And because the scorecard is based on worker feedback that can be collected repeatedly, it becomes a tracking mechanism over time. Intervene on a workflow, remeasure six weeks later, and see whether the score improved. That is your evidence the intervention worked, and your signal to scale it.

Most organizations make transformation investments without this kind of feedback loop. They deploy and wait for lagging indicators to confirm what they suspect. The workflow scorecard makes the feedback loop continuous.

There is a framing issue at the center of most AI evaluation efforts, and it is distorting the signal.

Leaders are asking whether the model is performing: accuracy, latency, usage.

The more useful question is whether the work is performing better. For every workflow being transformed by AI, and for every role involved, the real question is: did this make it easier and faster for workers to reach a better outcome?

These are different questions that require different data, and they lead to very different conclusions about whether an AI investment is working.

A CIO magazine article on rescuing failing AI initiatives put it plainly: leaders need to shift from model performance metrics to workflow performance metrics. The technology can be working perfectly and the work can still be worse. Employees may be using the tool, as clicks and logins confirm, but if they are also doing more manual review, navigating more unclear handoffs, or spending more time reconciling AI outputs with reality, adoption is not translating into value.

The organizations making genuine progress on AI ROI have learned to separate these two signals. Model performance tells you whether the technology is functioning. Workflow performance tells you whether it is creating value in the context of real work.

Workflow performance is harder to measure. It requires getting inside the work itself: how effort is distributed, where time goes, and what has improved or gotten worse since the AI was introduced. System data captures some of this, but much of it requires direct input from the workers running the workflows, who know the full picture in a way no dashboard can reconstruct.

The shift also matters for how organizations diagnose problems. When AI underperforms, leaders often look at the tool first: model quality, prompt engineering, integration. Those are worth checking, but more often the diagnosis points to something surrounding the AI, such as a workflow that was never redesigned to accommodate it, a role left unclear, or a data source the AI cannot access.

Those problems are invisible at the model level. They only become visible when you measure the workflow.

For every AI deployment worth measuring, build in a workflow performance baseline before go-live, then remeasure at regular intervals. The delta between those measurements, not the model metrics, is where your ROI signal lives.

Every organization we work with has plenty of data.

They have system telemetry, process mining outputs, engagement survey results, and productivity dashboards. What they typically do not have is a clear answer to this question: in the workflows that matter most to our AI transformation, what is actually getting in the way?

System data tells you what happened. It does not tell you why effort is high, where time goes, or what workers do off-system to compensate for a process that does not quite work. Process mining shows patterns in structured flows and misses the informal coordination, judgment calls, and workarounds that never touch a system. Engagement surveys reveal that workers are frustrated, but rarely which workflows are most broken or what would fix them.

The gap is a data philosophy problem rather than a technology problem. We have gotten comfortable measuring what systems can easily log, and uncomfortable relying on what workers know firsthand.

What changes that equation is asking workers directly: specific, brief questions about the workflows they actually run.

A head of AI at a Fortune 500 insurance company described the challenge well. His team had undertaken a massive workflow mapping effort across HR, finance, and product, a critical input to their AI strategy. “The workload is heavy, it’s slow, and we rely heavily on external consultants to drive it. At the end of the day, we still don’t have quantitative data behind it — so I can’t measure progress.”

That is the problem with qualitative methods at scale. Interviews and workshops provide depth, but they are slow and expensive, and they produce insights that are hard to prioritize and impossible to track over time.

The alternative is a structured, workflow-specific feedback mechanism that takes workers under two minutes to complete, reaches a statistically meaningful sample, and produces quantified data on where effort is high, where friction concentrates, and what causes it, across every workflow you care about.

Within three weeks, that insurance company had visibility across workflows and roles, quantitative data on effort and friction, root causes behind the biggest issues, and clarity on which workflows to prioritize first.

More importantly, they had a baseline, which means they can measure whether the changes they make are improving work over time instead of assuming they are.

Most organizations can identify friction. Fewer can track whether they are fixing it, and that is the piece that matters.

There is a consistent gap between what enterprise leaders say about AI deployment and what they actually do.

The language is familiar: “People first.” “Human in the lead.” “AI needs to work for workers.” These phrases appear in strategy decks, town halls, and press releases across nearly every major organization deploying AI.

Then you get on a call with the leaders actually running those deployments, and you hear: “I don’t need input from employees on their workflow reality. I have telemetry. I have process mining. I have task intelligence.”

We trust scraped system data more than the people doing the work. We prefer click-tracking, an echo of what is happening, over asking directly for the actual thing.

When you probe why, two objections come up consistently.

The first: “We can’t burden employees. There’s survey fatigue.” This is real, but it is being misapplied. Survey fatigue is the result of asking people too many questions about things that do not change anything. If you ask workers about the specific workflows they run, what gets in their way, what is improving, what they would change, and then act on the answers, that is not a burden. Workers generally want to be heard on how their work is evolving. The challenge is asking the right things, briefly, and following through.

The second: “Employees aren’t objective. They’re biased.” This one is harder to accept. Workers are the only people who know where work actually breaks down, what they work around daily, and what AI genuinely helps with. For decades, Lean and Kaizen were built on exactly this belief: the people closest to the work are best positioned to improve it. That principle did not stop being true when AI entered the picture.

Yes, some leaders collect AI tool feedback, and some run change readiness surveys. But those center on the technology rather than the workflows workers run through it.

The red thread in effective AI transformation is that you cannot successfully evolve work without involving the people doing it. It happens to be the right thing to do, and it is also the most effective strategy available, because workers hold the map of the work you are trying to redesign.

Closing the say-do gap requires only that you actually ask.

Ask any large organization who owns the CRM, and you will get a name. Ask who owns the staffing policy, the onboarding process, or the new AI assistant, and you will get names. Then ask who is accountable for whether the workflow those things support actually got better last quarter, and you will not get an answer.

The gap is not one of effort or talent. Every function around a workflow is doing its job. IT shipped the tool, HR updated the role, Ops revised the process, and the AI team deployed the agent. Each function has metrics that say it is succeeding, and those metrics are not wrong. They just do not show whether the workflow the worker performs improved after all those changes landed in it.

The worker is the only person who experiences the combined effect. She runs the workflow across all of it: the tool, the policy, the process, the data, the supporting teams, the handoffs, and now the AI. When the pieces do not fit, she reconciles them in the flow of the day, and the combined friction appears on no function’s dashboard.

Design accountability is the missing piece. It means each functional owner can see, and is answerable for, how what they own affects whether the work improves. Not whether the tool shipped or the policy updated, but whether the workflow that runs across them got faster, easier, and better at producing the outcome.

This is a different demand than asking functions to coordinate more. Coordination without a shared measure of the work produces alignment meetings, not alignment. Design accountability requires an instrument: a quantified, recurring picture of how the workflow performs from the perspective of the person running it. With that picture, each owner can see the effect of their piece on the whole. Without it, accountability has nothing to attach to.

AI raises the stakes. Every function is now changing the work faster, with more autonomy and more capital behind it. The organization is already redesigning work. What it has not decided is who is accountable for whether the work gets better. Until someone is answerable for that question, AI investment will keep improving the pieces without improving the work.

Workflow intelligence exists to make the question answerable. Deciding who must answer it is a management choice, and it costs nothing to make.

‍

Work friction is anything that makes it harder for an employee to reach a better outcome.

The definition is intentionally broad because friction takes many forms: a tool that does not connect to the data a worker needs, a policy that requires three approvals for a decision one person could make, unclear handoffs between teams, conflicting information from different systems, or a workflow designed for yesterday’s process that has not caught up with today’s reality.

Individually, any one of these might seem minor. Collectively, they compound. A worker who spends fifteen extra minutes reconciling data, ten navigating an unclear escalation path, and five waiting on an approval they should not need has lost half an hour of productive time to friction, in a single workflow, on a single day.

Multiply that across thousands of employees and hundreds of workflows, and work friction becomes one of the largest hidden costs in any large organization. No P&L line captures it and no dashboard displays it, but it shows up in productivity that falls short of expectations, in AI tools that get adopted but do not deliver, and in employees who work hard without reaching the outcomes they are capable of.

Friction is also one of the primary reasons AI deployments underperform. When AI is introduced into a friction-heavy workflow, one of two things tends to happen: the AI accelerates the parts it touches and leaves the friction untouched elsewhere, or the AI itself becomes a new source of friction, requiring manual review, producing output that does not fit the surrounding process, or creating handoffs nobody planned for.

This is why removing friction and deploying AI are the same problem rather than separate workstreams.

The organizations seeing the strongest results from AI transformation made work friction visible first. They identified where it concentrates and what causes it, and that visibility let them redesign workflows before deploying AI into them, rather than discovering the friction after go-live through flat productivity numbers.

Work friction is specific and measurable, and when surfaced properly it is actionable. The challenge has always been surfacing it at scale, and that challenge is now solvable

‍

Agentic AI is going to make a lot of measurement frameworks obsolete. Understanding which ones, and why, is worth your attention before it happens to yours.

The dynamic is this: agentic AI will increasingly execute individual tasks inside workflows. Tasks that once required a human will be automated, accelerated, or disappear entirely: searching for an answer, completing a form, updating a system, generating a report. This is already happening in early deployments, and it will accelerate.

What will not disappear is the need to accomplish goals through workflows: handling a customer escalation, conducting a sales call, coordinating patient care, hiring into a team. These workflows will keep evolving as AI becomes involved, but the underlying goal remains, and workers will still be accountable for the outcome. They will just get there differently.

The workflow, not the task, becomes the unit of work.

That has a major implication for measurement. When tasks change or disappear, task-level measurement becomes far less useful. You cannot benchmark productivity against activities that no longer exist in the same form. What remains stable, and therefore measurable over time, is the workflow and its outcome.

Organizations will need visibility into whether the workflow itself is improving: whether it is faster and easier, whether outcomes are better, and whether the experience of running it is improving.

There is a second implication, and it is bigger. As AI takes over more tasks, leaders across technology, digital, AI, operations, and HR become responsible for something new: the experience employees have performing workflows alongside AI, and whether the AI in the loop is helping or creating new friction of its own.

Without this visibility, AI transformations may look compelling on paper but fail in the field. The worker becomes, in effect, the customer of every leader deploying AI, and of the AI agents themselves.

Soon, organizations will feed AI agents data about how employees experience their workflows: the friction they encounter, where they slow down, and what creates rework. Those agents will use that context to improve how the end-to-end workflow gets done.

That feedback loop only works if organizations capture workflow-level data in the first place, and most do not.

The question worth asking now is how your organization will build that capability before it becomes critical.

There is a measurement question most organizations are getting wrong, and it is undermining their AI investments.

The question they are asking: does AI improve performance on this task? Does it draft an email faster or generate code more quickly?

The question they should be asking: does AI improve the workflow? Does it make it easier and faster for workers to reach a better outcome, end to end?

These are not the same question, and they do not lead to the same answers.

Recent research from MIT Sloan captures it well: leaders “should focus less on whether AI excels at each individual step and more on whether it improves the efficiency of the entire workflow.” A productivity bump on one task does not automatically translate into a faster, better workflow. Sometimes it makes things worse, adding handoffs, review loops, and friction between AI outputs and the humans who still own the surrounding work.

This is exactly what shows up in the data. An AI tool improves the speed of a specific step, but that step was never the bottleneck. The bottleneck sits two steps later, in a handoff that nobody redesigned. The AI accelerated the input, the friction stayed in the output, and net improvement was zero.

Task intelligence, the measurement of AI’s impact on individual tasks, tells you something. It tells you much less than you need to know.

Workflow intelligence asks different questions: is the end-to-end workflow faster, is it easier for the worker to reach a better outcome, and where are time and effort concentrating now that AI is in the picture?

These questions require different data. System logs alone cannot answer them. You need input from the people running the workflows on where they get stuck, what has improved, and what has become harder than it used to be.

As the MIT Sloan research puts it: “It’s not about how I’m going to introduce AI in my existing workflow. It’s about how I can redesign my workflow in such a way that is more AI-friendly.” That means grouping AI-compatible steps, reducing handoffs, and designing around what AI does well and where humans add the most value.

Otherwise, you are paying for AI and still paying the friction tax.

The leaders making the most progress on AI ROI have made this shift. Task intelligence has a role, but the signal that matters lives at the workflow level.

The distinction matters more than it sounds.

A process map shows how work was designed to happen, step by step and system by system. It is built from the organization’s point of view and reflects intentions.

A workflow shows how work actually happens, across tools, teams, data gaps, exceptions, and human judgment. It is built from the worker’s point of view and reflects reality.

The gap between the two holds most of the friction: unclear instructions that workers resolve informally, conflicting data that someone reconciles by hand, tool-switching that adds fifteen minutes to a task the process map assumes takes two, exception handling that nobody wrote down, and informal coordination between roles that are not supposed to interact but always do.

For years, people absorbed this friction. They worked around it and built tacit knowledge about how to navigate what the process could not describe. The system worked, imperfectly, because people filled the gaps.

Now AI is being asked to step into those workflows, and AI does not absorb friction the way people do. It hits a gap in a data source and stops. It produces output that requires manual review nobody planned for, or creates a handoff nobody thought to redesign.

This is why AI so often fails to deliver the productivity gains the business case promised. The technology gets deployed into the process map version of work, and the messier reality of the workflow defeats it.

The needed shift is in the starting point. Before asking what AI can automate or augment, ask how work actually unfolds today. Where does time go, where do people get stuck, and where does effort concentrate in ways no dashboard shows?

You cannot transform work you cannot see. Process maps, useful as they are for system design and compliance, show the intention rather than the work.

The organizations making the most progress with AI have closed that gap. They hold a working picture of workflows as they actually exist, and they use it to make better deployment decisions, redesign ahead of problems, and measure whether AI made the work better.

That picture is what workflow intelligence provides, and it is where the next stage of progress starts.

‍