Work Understanding - Certification assessment
Objective. Test whether a learner can use Work Understanding correctly under realistic pressure: presenting a finding as an inferred candidate, refusing to sum overlapping findings, naming the reference path behind a conformance figure, and reporting measured released time rather than modelled opportunity - at the certification standard.
This is the practical assessment for the Work Understanding certification gate. It presents two realistic scenarios the learner must work through, a scored checklist of what a correct answer contains, and the critical errors that fail the assessment outright regardless of score. The self-check items are the actual assessment questions with model answers. The gate is the grounding standard: practical score at or above 80, no critical error, manager review.
See it in the demo
Scenario 1 - The stacked savings demo (Sales-led)
You are demoing Work Understanding to an operations leader at an RCM provider. On the friction screen the demo shows four findings on their claim process, roughly 2 hours each. The prospect leans in: 'So that's about 8 hours a week we're losing - if you can get that back across the team, the ROI is obvious. Put that number in the proposal.'
Work the scenario. Produce:
- Exactly what you say in the room, out loud, in response.
- How you reframe the four findings honestly.
- What you offer to put in the proposal instead of the 8-hour number.
- The next step you agree with the prospect.
This scenario tests whether you hold the line on the app's core interpretation check under direct commercial pressure - the moment where the easy win is to accept the number and the correct move is to refuse to stack.
Scenario 2 - The disputed conformance figure and the 'nothing improved' account (CSM-led)
You are the CSM on a live account. Two things land in the same week:
(a) A process owner emails: 'Your tool says our prior-auth process is 64% conformant. That's insulting - our team is excellent. The number is wrong.'
(b) The sponsor says: 'We've had Work Understanding for a quarter, we've got dozens of friction findings, we even automated a data-entry step - but days-in-AR hasn't moved. What are we paying for?'
Work the scenario. For each, produce:
- The likely cause and the first checks you run.
- What you say to the customer.
- What - if anything - you escalate, and to whom.
- For (b), the concrete plan to get to a measured result.
This tests conformance literacy (the reference path drives the number), coverage/confidence diagnosis, the validation loop, and knowing that the days usually sit in waiting, not the busy active step.
Scored checklist - what a correct answer contains
Score out of 100. The learner must reach 80 and commit no critical error (next section).
Scenario 1 (50 points):
- States the findings are observed contributions with a confidence, not proven waste (10)
- Explicitly refuses to sum the four findings, and explains overlap - they may touch the same time (15)
- Offers a validated before/after on one finding as the proposal number instead of the 8-hour stack (10)
- Pairs the friction finding with the experiment (baseline, one change, owner, re-measure) (10)
- Agrees a concrete next step: one process, one owner, a real finding-and-experiment (5)
Scenario 2 (50 points):
- (a) Names the reference path and re-agrees it with the customer before defending the 64% figure; treats low conformance as a prompt to investigate deviations, not a grade (15)
- (a) Checks coverage, detection confidence and taxonomy as first diagnostic steps (10)
- (b) Identifies the missing validation loop and/or finding backlog as the cause (10)
- (b) Checks the target is waiting where the days sit, not the longest active step (10)
- Draws the escalation line correctly - method, capture scope and privacy go to product/specialist, not defended from the demo (5)
Critical errors - automatic fail
Any one of these fails the assessment regardless of the points scored, because each one would damage a real customer relationship or put an indefensible claim in front of a buyer:
- Summing overlapping findings into a savings number (accepting or producing the 8-hour stack). This is the app's core interpretation failure.
- Reporting a modelled opportunity as a realised result - presenting an estimate as achieved released time.
- Quoting a conformance figure as an absolute grade without naming or agreeing the reference path.
- Presenting an inferred finding as proven causality or proven waste - treating discovery, friction or rework as settled fact rather than a candidate to validate.
- Overclaiming capture or access - promising a capture scope, confidence, or that a manager can drill to an individual, as fact from the demo.
- Defending a detection method, metric or privacy guarantee that should be escalated rather than saying 'confirm with product'.
A learner can give a fluent, confident answer and still fail on any of these. The gate rewards disciplined honesty over polish.
The certification gate
Per the Academy standard: practical score at or above 80, no critical error, and manager review.
The manager review exists precisely because the critical errors above can be committed smoothly. The reviewer checks not just that the learner reached the right conclusion, but that they reached it for the right reason - that they would hold the line under pressure, not just when asked a clean exam question. A learner who scores 85 but stacked the findings fails; a learner who scored 82 and refused the stack, named the reference path, and escalated what should be escalated passes.
Tie-back: this assessment is the application-level counterpart to the certification simulations in the curriculum (S10 for Sales, C12 for CSM). A rep who passes here can run the evidence-led demo and the adoption loop for Work Understanding without supervision; one who does not repeats the misinterpretations and data-setup modules before re-sitting.
Interpretation check
Mastery of Work Understanding is the discipline to treat every output as an inferred, confidence-bounded hypothesis: discovery and variants confirmed with the customer, conformance stated against its reference path, friction and rework as candidates not proven waste, and only measured before/after released time reported as a saving. Summing findings, reporting opportunity as result, or overclaiming capture are automatic fails.
In the field
KnackGlobal RCM is the natural setting for both scenarios: a 6,000-seat customer under a ProHance displacement attack will push hard for a headline savings number and will dispute a conformance figure - the rep who stacks findings or defends an unagreed reference loses credibility, while the one who offers a measured before/after on a waiting step and names the payer-driven reference path holds the account.
Practical task
Complete both scenarios in writing to the scored checklist, then self-mark against it and flag any critical error you nearly committed. Submit for manager review. To pass: score at or above 80, no critical error, and a reviewer confirmation that you held the line (refused the stack, named the reference path, reported only measured results, escalated what should be escalated).
Self-check
ASSESSMENT Q1: A prospect sums four 2-hour friction findings to 8 hours and asks you to put it in the proposal. What is a passing answer?
Decline to stack. Say the findings are observed contributions with a confidence, not proven waste, and that overlapping findings may touch the same time so they cannot be summed. Offer instead a validated before/after on one finding, paired with the experiment (baseline, single change, owner, re-measure), and agree a next step to run it on one process with a customer-side owner. Accepting or producing the 8-hour number is a critical error and an automatic fail.
ASSESSMENT Q2: A customer calls a 64% conformance figure 'wrong' and insulting. What is a passing answer?
Do not defend the percentage. Name the reference path it was measured against and re-agree it with the customer - the reference drives the number, so 64% is meaningless without it. Treat low conformance as a prompt to investigate which deviations add time, not a grade. Check coverage, detection confidence and taxonomy as first diagnostics. Quoting conformance as an absolute grade without the reference is a critical error.
ASSESSMENT Q3: An account has dozens of findings, automated a data-entry step, but days-in-AR has not moved. What is a passing answer?
Identify the missing validation loop and/or finding backlog, and the likely wrong target - effort on an active step while the days sit in waiting. Confirm whether acted-on findings have a baseline and re-measure in the improvement workspace, check that the target is the waiting that holds the days, cap live experiments, and run one clean before/after. Do not report any prior opportunity estimate as a result.
ASSESSMENT Q4: A buyer asks whether a manager can drill into one employee's screen. What is a passing answer, and what fails?
State what the evidence shows and at what aggregation, and say the real capture and access configuration is customer-owned, jurisdiction-dependent and confirmed with product and the customer's security owner. Promising a specific capture or access capability as fact from the demo is overclaiming and an automatic fail.
ASSESSMENT Q5: Name three things that fail the assessment regardless of score.
Any of: summing overlapping findings into a savings number; reporting a modelled opportunity as a realised result; quoting conformance as an absolute grade without the reference path; presenting an inferred finding as proven causality or waste; overclaiming capture or access; or defending a method or privacy guarantee that should be escalated to product.
Verified vs demo. Scenario screens are demo-shown on seeded data; figures are illustrative and SME-validation-pending. The assessment tests the app's own interpretation discipline, not product specifics - detection method, coverage, available metrics and privacy configuration for a customer remain confirm with product.