3
Applicants, three different outcomes
7
Chapters in this walkthrough
2
Real reviewer approvals, live
1
Shared workflow, three outcomes
Daniel Okoye
A new applicant with a clean health profile. His case moves through the fast track and reaches an issued policy, approved by a reviewer before it's finalized.
Applicant — Fast Track, Issued
Patricia Doyle
Applies twice in this walkthrough — once with a more involved but insurable profile (standard review, issued), and once with a serious cardiac history that leads to a documented decline.
Applicant — Standard Review + Decline
Fast track — issued
Clean lab and physician evidence — the system classifies and routes automatically, a reviewer signs off, and the policy is issued.
Standard review — issued
A more complex profile takes extra review steps on the same workflow, then the same reviewer sign-off before it's issued.
Declined — documented
A serious, physician-documented medical history correctly leads to a decline, with the specific reasons on record and a formal notice sent.
About this walkthrough
Every screen below is a real screenshot from an automated, end-to-end run of Regisseur, captured live — not a mock-up. The people are fictional and the data is illustrative. Three applicants, three outcomes: two policies issued (one on a fast track, one on a closer standard review) and one coverage decline, each reached by the same 27-step workflow responding to real underwriting evidence. How much human sign-off is required at each step is a setting your team controls per step, not a fixed platform behavior — this walkthrough shows both: most steps run fully automatically, while the moment a policy is actually issued always pauses for a named reviewer to approve it first.
Two requests go out automatically: an order to the exam vendor for lab results, and a request to Daniel's physician for a statement of health — each with a secure link so the recipient can respond without creating an account.
The physician request letter is generated and ready to send the moment it's needed — populated with the right applicant, the right request, no template editing required.
The exam vendor's view: a clean, no-login upload page. This is exactly what they see when they click the secure link they were sent.
A real file upload, captured live: the exam vendor's lab results land in the case the moment they're submitted, and the workflow moves forward on its own.
With the lab results and the physician's statement both received, the case has everything it needs to move into review.
The lab results are read and summarized automatically, producing a clear analysis document ready for the underwriting step.
this step APS Medical Review — Medical Review Prep Agent assembled lab + APS → clinical review packet. Output gated at awaiting_review (always_review autonomy — PHI-appropriate).
With a clean lab result and a straightforward physician statement, the case is automatically classified into the best available rate class and routed onto the fast track — the standard, more manual review track is skipped because it isn't needed.
The policy schedule is generated automatically, but issuing it is treated differently from everything before it: the case pauses here and waits for a named reviewer to sign off before anything is finalized. This is a deliberate setting — Woodgrove Life chose to require a human decision at the moment real financial commitment happens, even though earlier steps run fully automatically.
A reviewer opens the case, reviews the drafted policy, and approves it — a genuine action by a real person, captured live, not a simulated click. Only after this approval does the case move forward to issuance.
Daniel uploads his signed receipt through the same kind of secure, no-login link — a real action by the applicant himself, captured live.
The case reaches its final state — Policy Issued — and the case header shows a clean, settled "Complete" status.
The full path Daniel's application took, end to end: every step that ran shown in green, and the steps that weren't needed for his profile clearly marked as skipped rather than hidden.
Every fact about Daniel's case — his coverage, premium, and the evidence gathered along the way — in one record, populated automatically as the case progressed, all correctly attributed to Daniel himself.
Activity History: AGENT_TASK events + DOCUMENT_DELIVERED + AGENT_REVIEW_APPROVED (medical@woodgrove.demo at this step, policy reviewer at this step). actor attribution: human approvers resolved to real member identities.
A second, independent application arrives for Patricia Doyle — a different applicant with a different health profile, on the exact same workflow.
Patricia's evidence isn't as clean as Daniel's, so the system routes her case onto the standard review path instead of the fast track — the same workflow, adapting to what the evidence actually shows.
The additional review steps that only run on this path — the ones the fast track skipped for Daniel — all complete for Patricia's case.
Just like Daniel's case, Patricia's drafted policy pauses here for a named reviewer to approve before it goes any further — the same governance point applies no matter which path the case took to get here.
A reviewer approves Patricia's policy through the same review screen Daniel's case used — a genuine, live action, and proof that the sign-off requirement holds consistently across every path through the workflow, not just the simplest one.
Patricia's case reaches Policy Issued too — by a different route through the same workflow, proving the platform handles more than one kind of applicant on a single template, with the same reviewer sign-off in place either way.
A third application arrives — a different case for an applicant with a significantly more serious medical history documented by her own physician: a prior heart attack, ongoing heart disease, and continued tobacco use. This scenario tests whether the system correctly recognizes when a case should NOT be approved.
Based on the physician's documented findings, the system declines this application and routes it to a formal notice process rather than to policy issuance — the same evidence-driven logic that approved Daniel and Patricia's earlier cases here correctly withholds approval instead.
this step Adverse Action Delivery sent the decline notice email to the applicant (ross.gordon48@gmail.com, test substitute). DOCUMENT_DELIVERED event in History. After delivery, this step auto-advanced.
The application reaches its final state: Closed, Declined, no premium — a clear, unambiguous outcome rather than a case left hanging in limbo.
The full path this case took, end to end: the decline and notice steps completed (shown in green), and every policy-issuance step correctly skipped because they no longer applied.
The case record shows "Declined — no premium" side by side with the physician's own documented findings — the specific medical evidence that drove the decision is on record, not a black-box outcome.
Activity History: AGENT_REVIEW_APPROVED at this step (medical) and this step (underwriter FCRA gate), DOCUMENT_DELIVERED (decline letter), NODE_COMPLETED for denial-path nodes. actor attribution. Regulatory audit: specific FCRA reason + human sign-off + delivery timestamp on record.