Role clarity is the least glamorous number in Gallup's latest workplace read, and probably the most important one. In data published on 21 July, 31% of US employees were engaged at work in the first half of 2026, unchanged from 2025. Eighteen percent were actively disengaged. Buried under those headline figures sits a quieter statistic: only 49% strongly agreed that they know what is expected of them at work.
Slightly more than half the workforce is not sure what their job is. Not what the company strategy is, not what the five year vision looks like. What their own job is, this quarter.
Worth noting where these numbers come from, since not all workplace research deserves equal trust. This is Gallup's probability-based panel, 43,262 responses collected 4 to 19 February and 6 to 20 May 2026, weighted to Census population targets, margin of error plus or minus 0.7 points. The methodology is published alongside the findings. Gallup sells consulting, but it does not sell the HR software the numbers might push you toward, and it shows its work.
Engagement has been flat for two years, and the money is real
The stall is the story. Engagement peaked at 36% in 2020, fell to 31% by 2024, and has not moved since. Gallup translates that into roughly eight million fewer engaged employees than in 2020, and estimates the cost of not-engaged and actively disengaged staff at about $2 trillion a year to the US economy.
Cost estimates at that scale are always a modelling exercise, so treat the trillion as an order of magnitude rather than an invoice. The flatline is the harder fact. Four years of engagement programmes, listening tools, wellbeing budgets and manager training have produced no net movement at the national level.
That should prompt a question most HR teams avoid asking out loud. If the interventions are not moving the number, maybe the interventions are aimed at the wrong thing.
Clarity is upstream of everything else you measure
Here is why the 49% figure deserves more attention than it gets.
"I know what is expected of me at work" is the first item in Gallup's engagement index, and it sits first for a reason. It is a precondition. An employee who cannot say what good performance looks like in their role cannot be meaningfully engaged by anything else you do. Recognition lands oddly, because they are not sure what they are being recognized for. Development conversations drift, because there is no agreed baseline to develop from. A pulse survey asking whether they feel valued collects an answer shaped mostly by their last conversation with their manager.
This is the part that makes clarity such an unglamorous problem: it is not a feelings problem, so it does not respond to feelings solutions. You cannot fix it with an offsite. It is closer to an information architecture problem. Somebody has to write down what the job is, agree it with the person doing it, and keep it current when priorities move.
We have argued before that an eNPS score is only as useful as what sits underneath it. Role clarity is a good candidate for what sits underneath. If half your people are unsure what is expected, your sentiment metrics are measuring the fog rather than the weather.
Why clarity decays even when someone wrote it down
Most organizations have documented roles. They also have a 49% problem. Both things are true because clarity is not a document, it is a state that decays.
Three things erode it in ordinary companies. Priorities shift mid-quarter, and the shift gets communicated in a meeting rather than written back into the goal, so the recorded version and the real version drift apart. People pick up work that belongs to a vacancy nobody has filled, and the borrowed responsibilities never get formalized or removed. And reorganizations move reporting lines faster than anyone updates what each team is accountable for, which leaves two managers each assuming the other owns a thing.
None of these are failures of effort. They are what happens when the description of the work is stored somewhere different from the work itself. A job description in a shared drive, goals in a spreadsheet, actual priorities in Slack. Each of the three is current at a different moment.
The practical test is cheap and slightly brutal. Ask five people in the same team to write down, without conferring, the three things they will be judged on this quarter. Then ask their manager to do the same for each of them. Compare. Most teams that try this find at least one significant mismatch, and the mismatch is usually invisible to both sides until the review conversation, by which point it is a performance discussion rather than a clarity one.
The AI finding is a clarity finding in disguise
Gallup's headline framing is about AI, and the AI numbers are interesting, but read them carefully and they repeat the same lesson.
Engagement runs six points higher at organizations that have adopted AI, and eight points higher among employees who use it at least weekly. Those gaps are modest. The bigger ones are about direction rather than technology. Employees whose organization gives a clear plan for integrating AI show a 15 point higher engagement rate. Where a manager actively supports AI use, engagement sits at 48% against 30% where employees do not report that support. With all three conditions present, it reaches 53%.
Strip the subject matter away and the pattern is: tools produce small gains, explicit direction about the tools produces large ones. The differentiator is not whether the company bought something. It is whether anyone told people what it is for and what is now expected of them because of it. That is role clarity wearing a different hat.
There is a correlational caveat here, and it should be said. Organizations with clear AI plans are probably better run in general, so some of that 15 point gap is the underlying competence rather than the plan itself. But the direction of the finding is consistent across every cut of the data.
What to fix, in order
Start with goals, because that is where clarity is cheapest to establish and easiest to keep current. Every person should have a small number of written objectives, agreed rather than assigned, with an explicit statement of what completion looks like. Small matters. Fifteen objectives is another way of saying no objectives.
Then attach the conversation to the object. The reason clarity decays is that goals live in one place and the discussions that change them live in another. In DTPulse, objectives sit alongside recurring one-on-ones, so when a priority shifts in a conversation the record shifts with it, and the version everyone can see stays the version that is true.
Then make the org chart honest. A surprising share of "I do not know what is expected of me" is actually "I do not know who decides." When reporting lines and team accountabilities are visible and current, a chunk of the ambiguity resolves without a single conversation about goals.
Finally, measure clarity separately from sentiment. Put Gallup's first item into your own pulse survey, verbatim, and track it as its own line. It gives you something an engagement score cannot: a number you can act on directly, because the fix for low clarity is writing and talking, not culture change.
The flat national engagement number will not move because of anything a single mid-size company does. But the 49% is not a macroeconomic fact. It is the sum of a lot of individual conversations that either happened or did not. That one is yours to change.