Compliance investigations with the evidence gathered, ordered and cited before you start
Every investigation starts with its evidence gathered, ordered and cited, and closes in a median of 18 days.
See one case, screen by screen ↓A Speak Up report lands and the case is open. Before the investigator can weigh a single allegation, someone has to place a legal hold, pull email, CRM call notes, expense reports, event records and training history for every custodian in scope, and work out which of thousands of items actually matter.
Then the facts have to be put in order by hand. The contradiction that decides the case — a sign-in sheet with 8 signatures against a receipt for 5 covers — sits across three systems, and only shows when the dates line up. Interview outlines get written from memory rather than from the gaps, and the report is drafted at the end, when every statement still has to be traced back to its source before counsel signs.
Estimated split for a typical investigation, in days, by hand and with the solution.
Six specialist agents triage the report, collect the records, build the timeline, weigh each allegation, plan the interviews and draft the report; the investigator decides each finding and counsel signs.
An anonymous web report says the Corventa speaker dinners with Dr. Victor Hale in Atlanta North repeat the same doctors, list people who were not there and split the bar bill. Here is IC-1042, screen by screen, in the working solution.
Lena opens the solution to 8 open investigations, 4 items that need her today and 4 cases past the 28-day median. IC-1042 is at the evidence stage on day 18. A new named report, SU-2291, is waiting in Speak Up: the intake agent read it, matched it to IC-1042 at 91 % and sent the reporter the anti-retaliation notice. Lena adds it to the case as exhibit EX-16.
Needs you: “IC-1042 · Build the case file — 2,318 items collected from 6 sources.”
IC-1042 is high risk and privileged, run under the direction of employment counsel Omar Haddad. The scope names two subjects, territory manager Ryan Castell and district manager Mia Torrance, 4 custodians and Jan 1 – Oct 7, 2026. The evidence collector already holds 2,318 items: 1,684 from the email archive, 412 CRM notes, 96 expense lines, 58 event records, 4 training records and 64 policy documents.
“About a minute · nothing leaves the case workspace.”
The evidence collector reads the 2,318 items in 14.2 s and keeps 16 exhibits. The timeline builder puts 21 events in order across 6 sources and flags 10 where sources disagree. The policy analyst lays 5 allegations against 5 policy clauses and proposes a finding for each. The interview planner writes 4 outlines with 16 questions tied to 4 evidence gaps, with the counsel notice added.
“Waiting for you — each finding is yours to decide · scope changes go to counsel.”
Zoomed to the flagged week, the timeline shows Jul 16–17 marked “sources disagree”. The sign-in sheet carries 8 signatures; the restaurant receipt shows 5 covers and $1,610.40, including $488 at the bar. Expense report ER-55821 claims $998.40 for 8 attendees, and a second report, ER-55822, files the remaining $612.00 as an “office lunch”. Each dot opens its exhibit with the passage highlighted.
ER-55821: “Per attendee: $124.80 — just under the $125 cap, so no extra approval was triggered.”
A3 — meal cap exceeded and the charge split across two reports — is laid against SOP-HCP-014 §6.1 and §6.4. Four pieces of evidence for, each with its exhibit: the receipt, both expense reports and two card transactions seven minutes apart. Nothing found against it. The proposal is Substantiated at 96 %, with the reasoning written out.
“Three records agree. The real cost was $322 a head — $985.40 over the cap for 5 people. A similar case was substantiated in 2024.”
A4 — the sign-in sheet lists HCPs who were not there — now carries the second report as evidence: Dr. Lund was at a conference in Denver that week. Against it, signatures cannot be verified from the scan, and gap G1, who signed for the HCPs who were not there, is on an interview outline. Lena marks it Substantiated; the decision is recorded with her name and the time.
The analyst never records a finding. Proposals below 70 % are never pre-selected.
Building the timeline turned up something nobody reported: needs assessments are missing for the 4 programs added after Jun 4, and the district manager wrote “skip the needs assessments for the add-ons”. Because the case runs under counsel direction, A5 cannot be decided until Omar Haddad adds it to the scope. Lena asks him from the allegation; he adds it.
“Found while building the timeline — it is outside the reported allegations, so counsel decides whether to add it to the scope.”
Four outlines, witnesses first and the subject last: specialty representative Alana Pierce, the venue coordinator at Lakeline Meeting Services, Mia Torrance and then Ryan Castell. Mia Torrance’s outline opens with the counsel notice and asks why the four add-on programs were approved and the needs assessments waived. Every question names the gap it closes and the exhibit to show.
Counsel notice: “This conversation is privileged; the privilege belongs to the company, which may decide to share it.”
The report writer drafts the investigation report in the house format: allegations and scope, how the evidence was gathered, chronology, findings, root cause and corrective actions. It works from Lena’s decisions, not the proposals, and marks every page “Privileged & confidential · prepared at the direction of counsel”. 37 statements cited, 0 uncited.
When Lena first sends it, the check stops her: “1 finding still proposed — a person decides every finding before counsel signs.”
With all 5 findings decided, the report goes to Omar Haddad, who signs it. Six corrective actions become tasks with owners and due dates: a referral to the discipline committee, recovery of the $985.40 over-cap spend and the $612.00 mis-filed charge, corrected transparency records for the Jul 16 dinner, a pause on new bookings for Dr. Hale, a manager accountability review, and a monitoring rule for split charges. The case moves to Remediation.
The same investigation, seen by the five people who work it — what their weeks looked like, and what they look like now.
Six specialist agents triage the report, collect the records, build the timeline, weigh each allegation, plan the interviews and draft the report; the investigator decides each finding and counsel signs.
Reads each Speak Up report, sets category, risk and privilege, and matches it to open and closed cases.
Collects email, CRM, expense, event, HR and training records for the custodians and dates in scope, under legal hold.
Orders every event across sources, links each to its exhibit and flags dates where sources disagree.
Lays each allegation against the policy clause, weighs evidence for and against, and proposes a finding with confidence.
Writes interview outlines whose questions close specific evidence gaps and show the right exhibit, witnesses first and the subject last.
Drafts the investigation report from the timeline and the investigator’s findings — every statement cites its exhibit.
Decides each finding; counsel signs. The agents propose; a named person decides.
Ask about any case, exhibit or rule — or say what to change. Answers cite the exhibits they rest on.
What is the strongest evidence for the sign-in sheet allegation?
The receipt and the venue booking put the Jul 16 dinner at 3 or 4 HCPs at most, against 8 signatures — and a CRM note the next morning records Dr. Brooks saying she missed it. Still open: who signed for the HCPs who were not there (G1) — on the interview outline for Ryan Castell on Oct 14.
How much money is involved?
For Jul 16 alone: the real bill for 5 covers was $1,610.40; at $125 × 5 the allowed amount is $625.00, so $985.40 is over the cap, and $612.00 was mis-filed as an “office lunch”. Speaker fees for the nine programs total $22,500. The recovery action is with Priya Raman, due Oct 31.
Do we need to correct transparency reports?
Yes. The Jul 16 meal was reported at $124.80 for each of 8 HCPs, but the receipt shows room for 3 at most — so at least 4 of those records name someone who was not there. Those records must be corrected, the value per HCP who did attend restated, and the other 8 programs reviewed the same way.
Add a rule: flag split charges at one merchant on one day
Done. New monitoring rule: flag two or more expense charges at the same merchant on the same day by the same employee when either one is a program meal. Run over the last 12 months it finds 6 more pairs in 3 territories — queued for Priya Raman. The rule is on in Settings and recorded in the audit trail.
13 screens from the working solution, on its sample data. Pick one to see it large.
Privilege, privacy, sign-off and the rules the agents apply are settings.
| Setting | Default | Choose from |
|---|---|---|
| Start these cases under counsel direction | On | HCP interactions, anti-bribery, retaliation, patient support |
| Mark interview memos and report drafts privileged and confidential | On | On · Off |
| Scope changes need counsel approval | Always on | Required by the investigations procedure |
| Show the agents’ proposed finding | On | Turn off to decide each allegation without seeing a proposal first |
| Call out proposals below this confidence | 70 % | 60 · 70 · 80 % |
| EU employee data stays in the EU region | On | On · Off |
| Default collection window | Allegation dates ± 90 days | ± 60 days · ± 90 days · Calendar year to date |
| Report sign-off for HCP interactions and anti-bribery | Omar Haddad | Omar Haddad · Priya Raman · Dana Okafor · HR counsel |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample case data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (whistleblowing benchmark, 2025, 4,000+ organisations: median closure 28 days). People named on this page are characters in the working solution; Northwind Bio, Corventa and the doctors in the case are its sample names.
Software that takes a compliance concern from the Speak Up report through triage, gathering the evidence, interviews, findings and a signed report to corrective actions, with every step recorded. The Investigation Case Builder adds agents that collect the evidence, build the timeline, weigh each allegation against policy, plan the interviews and draft the report.
The timeline builder creates one event per fact, with its date, its source and the passage of the exhibit it comes from. When another source on the same date contradicts it, the event is flagged and the date is ringed — in the sample case, 8 signatures against a receipt for 5 covers on Jul 16.
No. The policy and substantiation analyst proposes a finding with a confidence and the evidence for and against; the investigator decides. Proposals below the confidence bar you set are flagged and never pre-selected, and you can switch proposals off to decide without seeing one.
Categories you choose start under counsel direction. Interview memos and report drafts carry the privileged and confidential label on every page and in the export, every employee interview opens with the counsel notice, and new allegations go to counsel before they are investigated.
Yes. Evidence is collected over read-only connections, only for the custodians and dates in scope, and under legal hold, with a fingerprint and chain of custody on every exhibit. EU employee data stays in the EU region, and cross-border access asks for the data protection officer’s approval.
Your email archive, CRM, expense system, event system, HR and learning systems, and your Speak Up channels — web, hotline and email. It also reads your policy library, so each allegation is laid against the clause it concerns.
Yes. Every finding is decided by a named person, scope changes need counsel’s approval, and the report goes to the signer you set for each category. Every agent step and human decision is written to the case activity.
The Agentic Solution Engine builds and deploys it from your requirements — your investigations procedure, policies, report format and a sample of closed cases — and it goes live once every quality gate has passed. We will walk you through it on your own cases first.
Every speaker program scored against the OIG's eight red flags, before and after it runs.
1-min filmPayment RecipientsEvery payment to a clinician or teaching hospital matched to the right registry record, with the evidence behind each match.
1-min filmThird-Party Due DiligenceEvery distributor, agent and vendor checked for bribery risk before the contract is signed.
1-min filmWe’ll run the Investigation Case Builder on a sample of your own closed investigations.