Work Understanding: Discovery mode
Objective. Explain discovery mode as the early, learning phase where the app is still building the process picture, and read its findings as provisional with lower coverage and confidence.
Discovery mode is the phase where Work Understanding is still learning a customer's processes from observed work. It answers 'what are we seeing while the picture is still forming?' - a build-up view whose findings are provisional by design.
See it in the demo
What this screen answers
The customer question is: when we first turn this on, what does it show, and how much can we trust it yet? Sales people use discovery mode to set honest expectations for the first weeks; CSMs use it to manage the gap between go-live and a stable, trustworthy process picture.
The value is a managed ramp rather than a false promise of day-one insight. Discovery mode exists precisely so a customer understands that the system is learning the real workflow from observed work, and that early findings are the start of a picture, not a finished one.
The task
- Open discovery mode from the overview.
- Read it as a coverage-and-maturity view: which processes are being discovered, how much work has been observed, where confidence is still low.
- Use it to decide what is ready to act on and what needs more observation before a finding can carry a decision.
- Set the expectation explicitly with the customer: discovery findings are provisional and will firm up as coverage grows.
Describe the phase by its purpose - learning the process - not invented controls, durations or thresholds.
How to read it
Read discovery-mode output as provisional, with coverage and confidence still rising. The classic misread is treating an early, incomplete picture as the settled truth and acting on a finding before the process is adequately observed. An early 'variant' may just be unobserved normal work; an early 'gap' may be a process the system has not seen enough of yet.
Low coverage in this phase is expected, not a defect. The honest move is to let findings mature: a strong, repeated pattern over adequate observation is actionable; a thin, early one is a lead to watch.
Honest limits
This screen does not deliver a complete or stable process model, and its early figures should not seed a customer commitment. What 'adequate coverage' means, and how long discovery takes for a customer's processes, depend on their data and volume.
Confirm with product how discovery works, what maturity or coverage signals indicate a process is ready to trust, and when discovery transitions to steady-state reading. Position discovery mode as evidence of an honest ramp - a reason to trust the eventual numbers - not as a view to make decisions from prematurely.
Interpretation check
Discovery-mode findings are provisional while coverage and confidence are still rising; an early, incomplete picture is not settled truth and should not seed a decision or a customer commitment.
In the field
For an RCM rollout, saying 'in the first weeks we are learning your claim workflows, so early findings are leads, not targets' sets the honest ramp that makes the later days-in-AR evidence credible.
Practical task
On the demo discovery-mode screen, pick one process that is still being discovered, note what coverage or maturity signal you would look for, and write the one sentence you would use to set a customer's expectation about when its findings become actionable.
Self-check
A customer wants to act on a finding in the first week of discovery. What do you advise?
Treat it as a provisional lead to watch, not a decision. In discovery the process is only partly observed; let the pattern mature over adequate coverage before acting.
An early 'variant' appears. Does that mean the process is genuinely inconsistent?
Not necessarily. It may be unobserved normal work the system has not seen enough of yet. Low coverage in discovery is expected, not proof of a real variant.
Is low coverage in discovery mode a defect?
No. It is the nature of the learning phase. The honest framing is a managed ramp that firms up as coverage grows, which is what makes the eventual numbers trustworthy.
What should you confirm with product about discovery?
How discovery works, what coverage or maturity signals mean a process is ready to trust, and when it transitions to steady-state - these depend on the customer's data and volume.
Verified vs demo. Screen is live in the demo at /#/discoverymode; exact controls, maturity signals and figures are SME-validation-pending, and how discovery and its readiness thresholds work for a customer must be confirmed with product.