Investor onboarding & AML/KYC document chasing
Can AI speed up investor onboarding and AML/KYC document chasing? Yes for the reading and the first-pass triage — classifying documents, spotting gaps, drafting the chase, thinning screening alerts. No for the sign-off: the money-laundering judgment stays a named person's regulatory call.
| The pain | Every new subscriber arrives as a folder — subscription document, passports, proof of address, structure charts, source-of-funds evidence, for every beneficial owner. Someone reads each, works out what is missing, drafts the "please also send…" email, re-checks on return, and runs the names through screening — then reads the screening hits to sort the real matches from the false ones. Multiply by every investor in a close, against the closing deadline. |
|---|---|
| What AI does today | Classify the documents (what is this, whose is it, does it match the sub-doc), extract the identity and ownership data, spot the gaps and draft the chase email, and run a first pass over screening hits — clustering the obvious false positives so a human reviews a shortlist instead of hundreds of raw alerts. The boundary: AI reads and sorts; the deterministic screening match and the human disposition stay exactly where they were. |
| Proof it's real | Vendor-claimed. Gen II Fund Services — one of the largest private-capital administrators — began rolling out an automated onboarding and AML/KYC platform in November 2025 (announced via the vendor, Fenergo), with verified-KYC reuse across funds the headline gain. The document-reading and hit-triage layer on top is where the AI claim sits; we found no independent measurement of it in production. Newest evidence: 2025-11. |
| What it can't do | It cannot make the AML sign-off — that is where the human sits. A screening-hit disposition and the source-of-funds judgment are the money-laundering reporting officer's call and stay a named person's regulatory responsibility. A model saying "this is a false positive" is an input, not a decision — and treating it as a decision is exactly the failure an examiner will find. |
| The real alternatives |
|
| What you need in place | Access to the investor register and your screening provider; a data-protection sign-off before identity documents touch any model (this is personal data at its most sensitive); and named ownership — the MLRO owns dispositions, compliance owns the audit trail of what the model suggested versus what the human decided. That trail is your examiner answer. Identity documents also raise the bar in a group: data-protection and security sign-off usually sit centrally, so expect the approval path to run through the centre even for a desk-level pilot. |
| Effort & cost |
|
| What to watch | Check the false-negative behaviour, not just the time saved — a tool that clears onboarding same-day is only good if it is not also clearing the hit it should have escalated. Ask to see every case it auto-dispositioned, and sample them yourself. |
Prompt to try: build a KYC document-gap register for a batch of investors
The prompt checks each investor's KYC file against your onboarding checklist and the subscription form, and returns a linked gap register with draft chase points.
- 1Map the sourcesSubscription forms, KYC files, checklist
- 2Trace the preparationWho provided what, when, and in which form
- 3Match and reconcileInvestor IDs, names across documents
- 4Build the linked workbookGap register, all formulas
5. Controls (must pass)Every checklist document trackedExpired or out-of-date copies rejectedNames match the subscription formChecklist fits investor type and riskChase list matches the gapsFour-eyes sign-off
6. Measured over timeBack-test finds every known difference0 corrections after sign-offCycle time vs manualRolls forward with 0 manual edits
Before you start:
- Use your firm's approved AI tool, not a personal account.
- Mask investor names and any data your policy keeps out of AI tools.
- Replace the parts in [brackets] before you run it.
You are helping an investor onboarding analyst check the KYC files for [batch or closing] against the onboarding checklist. A human reviewer checks everything you produce. You do not approve an investor, you do not clear a screening hit, and you do not decide risk ratings.
I attach:
- Subscription forms: [the signed forms, with investor type and amounts]
- KYC files: [the documents received per investor]
- Checklist: [your firm's document requirements by investor type and risk rating]
Work through the six steps below. Show the result of each step before you move on.
STEP 1. MAP THE SOURCES
For each file, record: who produced it; the as-at date it states (not the date it was sent); the currency; what one row or section represents; the fields that identify an investor; the value fields and how each is labelled; and any subtotals, hidden rows, merged cells or footnotes that change how it should be read. Where the files differ in as-at date, period or currency, list each difference and say how you will handle it: convert at an FX rate from the Inputs tab with its source, or compare like periods only. Stop and ask me only if a difference means two figures cannot validly be compared.
STEP 2. TRACE THE PREPARATION
For each investor, say which documents arrived, from whom (investor, adviser, nominee), in what form (certified copy, original, electronic) and with what date. Note expiry dates, certification dates and any document that covers a different entity from the one subscribing.
Label each statement "stated in the file" or "my inference".
STEP 3. MATCH AND RECONCILE
3a. Keys. List the identifiers used: investor number, LEI, registration number, passport or ID number (masked). Say which appear on both the subscription form and the KYC documents.
3b. Names. Describe how names differ across documents: legal name vs trading name, nominee or trustee wording, name order, transliteration, abbreviations.
3c. Standardise. Write numbered rules for names and apply them in new columns, keeping the original text.
3d. Map. Build a table linking each subscription form to its KYC documents, with match method and reason. A suggested link is never confirmed.
STEP 4. BUILD THE LINKED WORKBOOK
Build an .xlsx file with these tabs, in this order:
- Index: a link to each tab, one line on what it holds, the as-at date, and the steps to refresh the workbook next period.
- Inputs: checklist version, as-at date, document validity periods, certification validity periods as the checklist sets them; plus the as-at date of each source file, with any differences between them and how each is handled; any FX rates needed, with their source and date, and the tolerance for agreement with documents (0.00 wherever exact agreement is required). Each value carries its source beside it, and every formula that needs one points here.
- _Source tabs: each file exactly as received, one tab per file. Never edit these tabs.
- Map: subscription forms linked to KYC documents. Each row carries its match method ("shared ID", "standardised name" or "suggested, needs review"), the reason, and the person who confirmed each suggested match and the date.
- Checklist: the requirements by investor type, as given.
- Register: one row per investor and required document, with the investor type and the name from the subscription form, and the risk rating as the KYC file records it ("Not found" where it records none). Required, received, valid on the as-at date and status ("Received", "Missing", "Expired" or "Not required") are formulas against the Checklist, Map and Inputs tabs. For each received document: the source document (file name), the page, the date on the document, the certification date where the checklist requires certification, the name as it appears on it, and a name-match flag (TRUE where that name matches the subscription form, FALSE where it does not).
- Chase: a formula-driven list of missing or expired documents per investor, with a draft one-line request for each.
- Exceptions: one row per failed control, untraced or disagreeing figure, name mismatch, unconfirmed match and open query, with a type column ("failed control", "untraced or disagrees", "name mismatch", "unconfirmed match", "query" or "timing difference"), its owner, status ("open" or "closed"), expected clearing date and the approver's sign-off (name and date).
- Checks: the controls in step 5.
- KPIs: the measures in step 6.
- _Log: the audit trail in step 6, with the names of the preparer and the approver.
Every derived number is a formula that leads back to a _Source tab or the Inputs tab. Nothing is typed in as a value. Use Excel tables and XLOOKUP or INDEX/MATCH. Wrap lookups in IFERROR and show "Not found", never zero. Dates as yyyy-mmm-dd, negatives in brackets, units in every value header, header rows frozen, no decorative formatting.
STEP 5. CONTROLS (EVERY ONE MUST PASS BEFORE SIGN-OFF)
On the Checks tab, write each control as a formula that returns TRUE or FALSE, with the count or difference beside it. A FALSE is not a score to improve: it blocks sign-off until it is resolved, by correcting the entry with a note or by getting the source corrected. An explanation alone does not clear it. The only item that may stay open at sign-off is a timing difference with an expected clearing date: it is listed by name on the Exceptions tab as a reconciling item and re-checked next period. A document tie-out failure is never a timing difference and can never be carried. Show every FALSE at the top of the Index tab under "Checks failing".
a. Complete: every investor on the subscription list appears on the Register, and every checklist requirement for that investor type has a row. Target: 0 unexplained rows.
b. Chase list agrees: the count of missing documents on the Chase tab equals the count of "Missing" and "Expired" rows on the Register. Target: a difference of 0.
c. Formula-linked: every calculated cell is a formula (test with ISFORMULA), and there are no #N/A or #REF errors. Target: 100%.
d. Tied to documents: on the Register tab, for each "Received" row, the document name, page and date are cited, and the name on the document matches the subscription form or is flagged. Every quote and figure must be traced (it cites its document and page) and must agree: a quote matches the document word for word, and a figure equals the document within the tolerance on the Inputs tab and no wider. Any item that is untraced or disagrees goes to the Exceptions tab at once. If the document itself looks wrong, for example a certified copy whose certification date is before the document's issue date, raise it as a query; never change the workbook to make it agree. Target: 100%.
e. Matches confirmed: count the rows matched by shared ID, by standardised name and by suggestion. Target: 0 suggested matches left unconfirmed at sign-off.
f. Right checklist: every investor's type on the Register matches its subscription form, and its rows follow the checklist version on the Inputs tab for that investor type and the risk rating the KYC file records.
g. Validity: no document counts as received if it has expired, or its certification is out of date, on the as-at date.
h. Not required is justified: every row reading "Not required" follows the checklist for that investor type and risk rating. Where the risk rating reads "Not found", the full checklist applies.
i. Open exceptions: count the Exceptions rows that are not closed, leaving out only named timing differences with an expected clearing date. Target: 0.
j. Four-eyes: the _Log names the preparer and the approver, who must be different people, and the approver signs off every Exceptions line.
STEP 6. REUSABLE, AUDITABLE AND MEASURED
- Reusable: next period, a person replaces the _Source tabs and updates the Inputs tab, and every other tab recalculates. Keep the Map tab so confirmed matches carry forward.
- Auditable: on the _Log tab, list each source file (name, producer, as-at date, date received, row count, control total), every rule and match you applied with the reason, every assumption, and who prepared and approved. Never overwrite source data. A correction goes in a separate column with a note.
- Measured: add a KPIs tab that the team fills in each period: back-test recall (run on a past period with known answers; share of known differences found; target 100%, and do not adopt the workbook below that); items agreed first time with no manual fix; items flagged that turned out fine; corrections made by the reviewer before sign-off; corrections found after sign-off (target 0); manual edits outside the Inputs and _Source tabs at refresh (target 0); and cycle time in hours from receiving the files to sign-off, against the manual baseline.
If your tool cannot create an .xlsx file, give me each tab as a table plus the exact formula for the first row of every calculated column.
At the end, give me investors complete, investors with gaps, and documents outstanding, the result of every check, and a list of anything you were unsure about.What good output looks like:
- Every required document shows as received, missing or expired, with the reason.
- Name mismatches are flagged, not silently matched.
- Chase points name the exact document, not "outstanding KYC".
- Every control on the Checks tab reads TRUE before sign-off. Only a timing difference with an expected clearing date is carried as a reconciling item.
How to know it worked:
- Back-test: run it on a past period where the answers are known. Adopt it only if it finds every known difference.
- After sign-off: no corrections are found later.
- Over time: fewer reviewer corrections and false flags, and a cycle time below the manual process.
Who signs off: The onboarding analyst sends the chases, and the MLRO or delegate signs off each investor in the usual control. The workbook is the working file, not the record.
In practice: test it on a closing you have already completed. If it finds a gap your team missed, the checklist needs a rule, not just the prompt.
Questions operators ask
Is AI-assisted KYC acceptable to regulators?
As an assistant, yes — regulators' concern is that a named person owns each disposition and the firm can evidence the decision trail. What fails examination is delegating the judgment: an auto-cleared screening hit with no human record is a finding, whatever the tool's accuracy.
Should we buy an onboarding platform or leave KYC with our administrator?
If the admin or a managed service already runs it and your closes are occasional, stay put — you're paying for exactly this. A platform earns its place on repeat closes at scale, where verified-investor reuse and e-subscription workflows compound; the AI reading layer is a feature of that decision, not the reason for it.
How much time does AI actually save in onboarding?
The honest answer: it compresses the reading and chasing (document classification, gap letters, alert thinning), which is most of the elapsed time in a messy close — but the human review and the investor's own response time remain, so "same-day onboarding" claims assume a clean file. Test on your messiest structure chart, not the demo.
Related: investor-services inbox triage and the Fund AI desk.