The operational playbook for running a workforce-level AI Career Index reading across a real team: what leadership has to decide first, how to frame privacy, the invite sequence, and what to do with the first reading.
Most teams that run a workforce-level AI Career Index reading do so for the first time without a playbook. The tool itself is simple to deploy. What is less obvious is how to frame the engagement internally, how to set participation expectations, and how to make sure the output becomes a real structural decision rather than a deck that sits in a shared drive. This is the operational playbook for getting that right on the first attempt.
Decide what the reading is for before you launch
The single most common failure mode is running a workforce reading without a clear internal agreement about what the output will change. The reading generates specific, measurable recommendations. If leadership has not pre-committed to acting on those recommendations, the exercise produces anxiety without change. Before any invite goes out, the leadership team should agree on three things: which structural questions the reading is meant to answer, who owns the follow-up decisions, and what the timeline for acting on the reading is.
In practice this means a 30 to 45 minute conversation between the CEO, COO, and head of people before launch. The questions to resolve: are we willing to restructure scope if the reading flags it, are we willing to freeze specific hires, and are we willing to open migration paths for staff in high-exposure roles. If any answer is no, the reading still has diagnostic value but the follow-up will be limited. Better to know that going in than to discover it after the output lands.
Frame the privacy model explicitly
The second operational failure mode is an under-communicated privacy model. The reading is only useful if employees answer honestly. They will not answer honestly if they believe the output could affect their role directly. The framing to use, which is also the operational reality: individual reports are private to the employee who completes the assessment; leadership sees only aggregated workforce-level metrics with no mechanism to identify individual responses. That is not a communications line. It is the design.
The internal announcement should say this explicitly. It should explain what leadership will see (Workforce Index, exposure distribution, authority band mix, archetype distribution) and what leadership will not see (any individual's response, any individual's score, any individual's archetype). Organisations that frame this clearly see participation rates in the 75 to 90 percent band. Organisations that do not frame it clearly see rates below 50 percent, which degrades the statistical validity of the reading.
The invite sequence that works
A working invite sequence has four touchpoints. First: a CEO or COO memo 48 hours before launch explaining why the reading is being run, what it will produce, and the privacy model. Second: the launch email itself with the assessment link, clear time expectation (seven to ten minutes), and a completion window (typically two weeks). Third: a reminder at the one-week mark, ideally from the same sender as the launch email. Fourth: a final reminder at the end of week two, with the framing that the reading will close shortly and the individual will not be able to access their personal report after it does.
What does not work: HR-only sponsorship without executive backing, open-ended completion windows, and anonymous reminder campaigns. Assessments with strong executive sponsorship and a clear close date complete at nearly double the rate of assessments that rely on internal HR communications alone.
Participation rates and what they mean
Target a 75 percent participation rate at minimum. Between 75 and 90 percent is the healthy range: high enough to make the workforce metrics statistically solid, low enough to reflect that the assessment is genuinely voluntary. Above 90 percent typically signals that employees felt implicitly required to participate, which can erode the honesty of answers. Below 60 percent means the framing was under-communicated or the privacy model was not trusted. The reading is still valid but the non-response bias widens.
If participation lands below target, do not rerun the assessment immediately. Wait a quarter, reframe the communications, and run it then. Immediate reruns almost always produce the same underwhelming result because the structural trust issue is unchanged.
Reading the output
The workforce output has four key metrics to read in this order. First, the Workforce Index score itself, which sits on a 0 to 1000 scale and tells you where the aggregate sits relative to other organisations in the same sector. Second, the exposure distribution: the share of roles in the high, moderate, and low compression bands. This is the forward indicator. Third, the authority density: the share of roles in the authority band. This is the insulation layer. Fourth, the archetype distribution: the split of roles across structural archetypes, which reveals scope concentration patterns.
For the conceptual grounding see workforce AI readiness: the structural diagnostic and measuring structural workforce risk. For worked examples with real before-and-after metrics see the workforce case studies.
Converting the reading into structural moves
A workforce reading that does not translate into specific operational decisions within 60 days usually ends up in a drawer. The operational pattern that works: leadership team reviews the reading within two weeks of close, identifies the top three structural moves (typically: one scope redesign, one hiring-strategy change, one migration-path creation), and commits to implementing all three inside one quarter. The re-read then happens at the six-month mark to measure whether the moves actually shifted the structural position.
The three most common structural moves after a first reading: repositioning production-layer roles into scope-owning roles through explicit authority migration, creating a senior escalation tier in high-exposure functions (support, coordination, junior analysis) so that AI-assisted workflows are supervised rather than unmonitored, and freezing hiring on flagged coordination roles while redirecting the backfill budget into the authority layer. Each of these appears across the workforce case studies.
What to do between readings
The period between the first reading and the re-read is when most of the structural value is created. Leadership should treat it as an operating programme, not a one-off exercise. The specific moves implemented should be tracked against the structural questions the reading raised. If the reading flagged high exposure in the marketing function, the re-read should show a measurable shift in that function's exposure distribution six months later. If it does not, the moves were performative rather than structural.
For the enterprise pricing that supports annualised or quarterly re-reads, see the company pricing tiers. For the product framing and what leadership receives, see the company overview. For the questions leadership teams most commonly ask before committing to a workforce engagement, see the company FAQ.
Run your structural assessment
The AI Career Index is free to take. Get your structural snapshot in 10 minutes, including your durability score, archetype, and projection.
Take the Free Assessment