From Periodic Reviews to Perpetual KYC: A 10-Step Rollout Plan for Banks
Most banks can't drop scheduled reviews, and supervisors rarely ask them to. Here's a 10-step sequence for adding event-driven KYC on top of your current policy, from cleaning the data to briefing your regulator.
What perpetual KYC actually changes, and what it doesn't
If your compliance team is behind on its refresh backlog, someone is pitching you perpetual KYC software as the fix. The idea is simple: stop reviewing every customer on a one-, three- or five-year cycle, and review a customer when something about them changes. That could be a new director, a sanctions hit, an unusual transaction pattern or a lapsed ID document. Done well, pKYC puts analyst time on files that have changed and takes it off the ones that haven't.
Vendor decks tend to skip one point. In most jurisdictions, event-driven review adds to scheduled reviews for higher-risk customers. It doesn't replace them. Some examples:
- European Union: the AML Regulation, which applies from mid-2027, sets maximum update intervals, with high-risk customers refreshed more often than the rest. Triggers come on top.
- India: the RBI's KYC Master Direction requires periodic updation at fixed intervals that lengthen as risk falls, from roughly two years for high-risk customers to around ten for low-risk ones. Check the current text before you design anything.
- UK, Singapore, UAE and the United States: the rules are more principles-based and lean on the FATF risk-based approach and ongoing monitoring. Even so, examiners routinely ask how high-risk relationships are reviewed on a fixed cadence.
So the realistic goal is a hybrid. Scheduled reviews stay for high-risk tiers, and for every tier where local law requires them. Triggers handle everything in between, and a low- or medium-risk file that monitoring keeps current can often be refreshed with a lighter touch. The ten steps below follow the order in which a compliance team actually does the work. Most of the risk sits in the first six, before any software goes live.
Setting a data quality baseline and drawing the scope
Step 1: Measure what your customer data can support
Triggers only fire on reliable data. Before you talk to vendors, sample customer files from each risk tier and segment and score them on these points:
- Completeness of mandatory CDD fields (identifiers, nationality, occupation or business activity, source of funds)
- Presence and age of beneficial ownership data for legal entities
- Structured versus free-text fields. You can't build a trigger on a scanned PDF
- Consistent customer linkage across core banking, cards, lending and wealth platforms
- Document expiry dates captured as data, not only stored as images
Take a hypothetical regional bank with 400,000 retail and 25,000 business customers. Suppose 30% of entity files have no structured ownership data. Ownership-change triggers would be blind for nearly a third of the business book. Fix that with a remediation sprint before launch, not after.
Step 2: Pick a first population and write down the exclusions
Don't start with the whole book. Most banks begin with low- and medium-risk SME customers. Volumes are high, the data sources (company registries, screening) are mature, and moving off fixed cycles saves the most effort there. Keep high-risk, PEP and correspondent relationships on their current schedule in phase one. Record every exclusion and its reason. Your regulator will ask.
Designing triggers that analysts can act on
Step 3: Build a trigger catalog tied to risk outcomes
Each trigger needs a data source, a threshold, a default action and an owner. A trigger that only says "something changed" produces a queue nobody can work. The table shows how a typical catalog is structured.
| Trigger | Typical source | Default action | Noise risk |
|---|---|---|---|
| Sanctions or PEP match | Screening provider | Immediate full review, escalation | Medium (name-matching false positives) |
| Change of director or beneficial owner | Company registry, data aggregator | Targeted review of ownership and control | Low to medium |
| Adverse media | Media screening | Analyst triage, then review if relevant | High without good filtering |
| Activity outside expected profile | Transaction monitoring | Profile refresh, possible risk re-rating | High |
| ID or licence expiry | Internal customer record | Automated outreach, document refresh | Low |
| Change of address to a higher-risk country | Internal record, payments data | Risk re-score, targeted review | Low |
Step 4: Tune the thresholds before anyone sees the queue
Run the catalog against six to twelve months of historical data before go-live and count the reviews each trigger would have generated. Say adverse media alone would have opened 8,000 cases a month for a team that can handle 1,500. Then the threshold or the source is wrong. Set materiality rules too. A registered-address change within the same city probably shouldn't open a review. A new shareholder above your beneficial ownership threshold should. This calibration decides whether KYC refresh automation lowers the workload or just moves it around.
Rewriting risk-tier policy without breaking the audit trail
Step 5: Rewrite the policy as a hybrid model
Your current policy probably says something like "medium-risk customers are reviewed every three years." The new version has to say what happens between those dates, what a trigger-driven review involves, and when a triggered review counts as the scheduled one. A common structure:
| Risk tier | Scheduled review | Event-driven review | Scheduled review satisfied by trigger? |
|---|---|---|---|
| High / PEP | Retained at current or regulator-set cadence | Yes, all triggers | Usually no; both run |
| Medium | Retained, possibly extended where law allows | Yes | Only if a full review was completed |
| Low | Statutory maximum, or lighter-touch confirmation | Yes | Often yes, if documented |
Be careful with the last column. If a targeted ownership review resets the clock on a full refresh, the policy must say so explicitly and the system must log the decision. Otherwise auditors will treat those files as overdue.
Step 6: Put the risk scoring model under formal governance
In a pKYC setup, risk scores are recalculated far more often, so the scoring model starts behaving like any other model the bank relies on. Document the factors and weights, who can change them, and how changes are tested. Most banks fold it into existing model risk management. Expectations for documentation and validation differ considerably between the EU, the Gulf states and Asia-Pacific, so confirm local requirements with your legal and compliance advisers.
Changing the analyst workflow and choosing the software
Step 7: Redesign the work, not just the screens
Ongoing customer due diligence changes the shape of analyst work. Big annual batches give way to a steady stream of smaller, targeted cases. Expect four consequences:
- Case types multiply. A targeted ownership review needs a different checklist from a full refresh. Build a template for each and don't force analysts through a full refresh every time.
- SLAs replace due dates. Set response times by trigger severity, for example same-day for a sanctions match and 30 days for an expired document.
- Customer outreach gets automated. Requests for documents and confirmations should go out through digital channels, not analyst emails.
- QA sampling changes. Sample by trigger type so you can spot triggers that keep producing low-value cases.
Plan for retraining. Analysts used to checking every field tend to over-review targeted cases, which wipes out the efficiency gain.
Step 8: Choose software on integration and evidence, not demos
The market splits roughly into three groups: client lifecycle management platforms with pKYC modules (Fenergo, Pega and others), entity-resolution and data specialists (Quantexa, Moody's, Encompass), and screening and monitoring vendors adding triggers (LSEG World-Check, Dow Jones, LexisNexis Risk Solutions, ComplyAdvantage, NICE Actimize, Napier). Coverage of local company registries varies widely. A vendor that's strong in the UK and EU may have thin data for Indonesia, Nigeria or Brazil, so test coverage in your own markets.
On cost, typical 2026 market ranges vary a lot by contract and region. A mid-sized bank might pay somewhere in the low hundreds of thousands to low millions of dollars a year in platform licensing, with implementation fees often equal to one to two times the first-year licence. Tier 1 programs can run well above that. Data feeds are frequently the cost that keeps growing, because registry, screening and media data are usually priced per entity or per check. Confirm all pricing with vendors directly.
Red flags: no audit log of why a trigger fired or was dismissed; trigger logic you can't adjust without a vendor change request; data coverage claims with no test against your own customer sample; and per-alert pricing, which gives the vendor no reason to cut noise.
Parallel running, regulator engagement and cutover
Step 9: Run both models side by side
For the population in scope, keep the scheduled cycle running while the trigger model operates alongside it for at least two quarters. Compare results. Did triggers catch risk changes that scheduled reviews missed? Did scheduled reviews find anything triggers missed? Every gap in the second category means a missing trigger or a data blind spot. Fix it before cutover. This comparison becomes your main evidence for the regulator and internal audit.
Say the hypothetical bank from Step 1 runs 6,000 medium-risk SME files through both models. If scheduled reviews turn up 40 material changes and triggers had already caught 37, you can show where the other three came from and how you closed the gap. That's a stronger case than any vendor benchmark.
Step 10: Brief the supervisor before cutover
Don't present pKYC to your regulator as a finished decision. Brief your supervisory contact early. Share the hybrid policy, the trigger catalog and the parallel-run results, and say clearly which tiers keep scheduled reviews. Some markets expect formal notification or prior discussion; in others an informal briefing is enough. Your compliance and legal advisers will know local practice. Then cut over one segment at a time, with documented criteria for reverting a segment to scheduled reviews if trigger performance drops.
A decision rule: if your Step 1 baseline shows structured, current data for most of your first population, and your Step 4 calibration fits within analyst capacity, start the parallel run. If not, spend the next quarter on data remediation. That will do more for your backlog than any software purchase.
Common questions
Can perpetual KYC software fully replace scheduled periodic reviews?
Rarely, at least for high-risk customers. Many regimes, including the EU's AML Regulation and India's RBI KYC Direction, set maximum review intervals, and supervisors elsewhere usually expect high-risk relationships to be reviewed on a fixed cadence. Most banks run a hybrid: scheduled reviews stay for higher-risk tiers and triggers handle changes between them. Check local requirements with qualified compliance counsel.
How long does a pKYC rollout typically take?
For a mid-sized bank starting with one customer segment, a year to two years is a realistic planning range. That covers data remediation, trigger calibration, policy rewrites, at least two quarters of parallel running and regulator engagement. Banks with poor beneficial ownership data or fragmented customer records should plan for the longer end.
What usually drives the cost of KYC refresh automation beyond the licence?
Implementation and integration work often costs one to two times the first-year licence. Third-party data, such as company registry, screening and adverse media feeds priced per entity or per check, grows with the size of your book. Internal costs for data remediation, model governance and analyst retraining are often left out of the business case. Get each quoted separately when you compare vendors, and treat all figures as indicative.
Written by the Wealth CRM Direct editorial team. Prices, features and rules change often - confirm details with the vendor or a licensed professional before you decide.