Demo build on sample data

Testing every transaction, not a sample, for a group audit team

The whole audit cycle in one place, from risk scoring and planning to testing, findings, follow-up and the audit committee pack.

This is a demo build. It runs on sample data. We keep the group's name private.

At a glance

Demo build, sample data

Business
Group internal audit
What we built
  • An audit command center
  • Full-population testing
  • Payroll audit
Internal audit demo build: entities, open exceptions and exposure, a risk heat map by company and a list of priority exceptions.
Audit command center demo.Demo build, sample data.

The problem

Internal audit at a large group is stretched thin. There are many companies in very different industries, and a small team that can visit a few of them each year on a fixed rotation. Fieldwork means pulling a sample into working papers. Anything outside the sample goes unchecked.

Group payroll is one of the riskiest areas and one of the least audited. Findings take weeks to write up. Overdue management actions slip. The audit committee gets a slide deck pulled together the week before the meeting.

What we built

An audit command center

Every entity, open exception and quantified exposure on one screen. A risk heat map scores each company on financial, operational, compliance and fraud risk, and the priority exceptions open line by line.

Full-population testing

The audit steps run over every invoice, asset and stock item. Each exception links to the evidence behind it.

Payroll audit

HR records, the payroll register, social insurance records and the bank payment file are cross-checked for ghost employees, shared bank accounts, overtime outliers and payment file mismatches.

What's in it

  • Risk and annual plan: the whole audit universe scored from live data, with high-risk entities missing from this year's rotation flagged.
  • Planning: pick an entity and it drafts the scope, the risk and control matrix, the audit program and the data request, based on the team's own prior-year method.
  • Continuous monitoring: nightly tests for duplicate payments, unapproved invoices, manual journals, vendors that match employees and segregation-of-duties conflicts.
  • Findings: exceptions written up in the standard five-part format of condition, criteria, cause, impact and recommendation.
  • Issue tracking with owner, due date, age, reminders and escalation.
  • Reports: a drafted engagement report and a quarterly audit committee pack.
  • Communications: data requests, finding notices and reminders drafted with evidence attached, waiting for approval.
  • AI governance: an inventory of the group's own models, checked for bias, privacy, explainability and human oversight.

What it would take to go live

The audit function. Its methodology, the group's ERPs, the entities in scope and how findings are reported today.

The data and programs. Ledgers and payroll connected, the team's audit programs loaded and the monitoring rules set.

A first version in about 10 days, usually one entity and payroll.

The group owns its data and the platform. We add tests as the plan grows.

See also AI for financial services and workflow automation.

Lead an audit team?

Book a call and we'll show you the demo, from the heat map to a finished finding.