Following the recent publication of the Governments AI Opportunities Plan we thought we would put it to the test. The plan states amongst many objectives that:
‘Government needs to develop a more sophisticated understanding of the value of the data it holds, how this value can be responsibly realised, and how to ensure the preservation of public trust across all its work to unlock its data assets.’
‘Government should look to collect data in strategically significant areas’
So with the Payment Intermediaries sector facing yet another unnecessary shake up due to HMRC incompetence, and a long list of Governments who are unable to follow through on their stated objectives with little, if any, accountability, we thought we would test out Labour’s big plan for AI.
We posed 3 questions to AI and its answers/analysis were quite enlightening!
Setting the scene:
Company A employs 100 workers and pays them as employees. The employees carry out work at Biz B and Biz B pays money to Company A for that work [BBP]. Company A then makes a series of deductions on the money received and pays the balance to the employees [CAP]. For each of the employees paid by Company A an electronic report showing a complete breakdown of all deductions, after Company A has retained a small amount for themselves, is provided to Body H [CAPR]. Body H is responsible for ensuring the correct deductions are made.
For the purposes of this example let’s assume that deductions are at 40% and that Company A retains £20 as their share before making deductions [CAMD].
Body H also receives a report from Biz B that confirms the total amount sent to Company A for each of the employees [BBPR].
What we did not reveal to AI was the following, but it will help you put this in context:
Company A = Umbrella Provider
Biz B = Recruitment company
Body H = HMRC
BBP = Gross invoice value
CAP = Employees Net Pay
CAPR = RTI
CAMD = Deductions, what in real world would be tax, NI, pensions etc.
BBPR = Employment Intermediaries Report
40% Deductions = Tax and NI Rates, actual rates can easily be replaced
£20 Deduction = Umbrella Margin
Question 1
In the example above how should Body H check that the correct amount was being paid by Company A to its employees?
Please include the calculations in your answer.
BBPR (report from Biz B) confirms the total amount paid to Company A for each employee.
CAPR (report from Company A) provides the breakdown of deductions and final payment to employees.
Step-by-step Calculation:
1. Determine the total amount sent per employee
- Suppose Biz B pays Company A **£X** for an employee.
2. Deduct Company A’s share (£20 per employee)
- Amount left after Company A retains its share:
**£X - £20**
3. Apply the 40% deductions
- Amount deducted:
**40% of (£X - £20) = 0.4 × (£X - £20)**
- Amount paid to the employee:
**(£X - £20) - (0.4 × (£X - £20)) = 0.6 × (£X - £20)**
4. Cross-check with CAPR reports
- The final amount received by the employee in CAPR should match **0.6 × (£X - £20)**.
**Example Calculation:**
If Biz B paid **£200** for an employee:
- Amount left after Company A’s share:
**£200 - £20 = £180**
- Deductions at 40%:
**0.4 × £180 = £72**
- Employee's final payment:
**£180 - £72 = £108**
Body H should verify that the CAPR report for that employee shows a payment of **£108**.
**end**
What AI confirms here is that HMRC already has all the data it needs to ensure compliance in the market. I refer back to a stated objective of the AI Plan - Government needs to develop a more sophisticated understanding of the value of the data it holds, how this value can be responsibly realised, and how to ensure the preservation of public trust across all its work to unlock its data assets.
Question 2
How would Body H identify any Company failing to make the correct deductions?
Body H should compare CAPR and BBPR reports for inconsistencies.
Step-by-step Checks:
1. Compare BBPR (Biz B’s report) with Company A’s declared total income
- If **BBPR total ≠ CAP total income**, Company A may be underreporting income.
2. Verify deductions using the formula from Q1
- Expected employee payment = **0.6 × (BBPR amount - £20)**
- If the CAPR report shows employees receiving **more than expected**, deductions are too low.
- If the CAPR report shows employees receiving **less than expected**, Company A might be over-deducting.
3. Check for discrepancies
- If **BBPR total - CAP total income > 0**, Company A is not reporting all money received.
- If **actual employee payment in CAPR > expected employee payment**, Company A is under-deducting.
- If **actual employee payment in CAPR < expected employee payment**, Company A is over-deducting.
**Example of Incorrect Deductions:**
If Biz B paid **£200**, and CAPR shows the employee receiving **£120** instead of **£108**, the extra **£12** suggests deductions were not made correctly.
**end**
Wow, so HMRC and HM Treasury are saying that this is beyond them as they have been unable to work this out for 10 years and AI did it in seconds. This is embarrassing for HMRC and HM Treasury.
As we have said for years if this had been implemented the market place would look very different now.
Now to Question 3, let’s see if Rachel Reeves decision making skills stands up to the test of AI.
Question 3
Discuss - As Body H you are currently not using any of the data described above but want to ensure that more Company As are making correct payments you have 2 choices:
Ignore the data and start again with a new set of rules or use the data provided and act on any discrepancies highlighted.
Please also note:
Option 1 holds costs and significant business disruption for both Company A and Biz B.
Option 2 there is no cost or disruption to Company A or Biz B.
And AI said:
Option 1: Ignore the data and create new rules
- **Pros:**
- Ensures a fresh and standardised approach.
- Could be more comprehensive.
- **Cons:**
- **High cost** and **significant disruption** to both Company A and Biz B.
- Loss of valuable existing data.
- Longer implementation time.
Option 2: Use the data and act on discrepancies
- **Pros:**
- **No cost or disruption** to Company A or Biz B.
- Allows real-time monitoring and gradual improvements.
- Immediate action can be taken on incorrect deductions.
- **Cons:**
- Requires efficient data analysis and monitoring.
AI Conclusion
**Option 2** is the most efficient and practical approach. By leveraging existing data, Body H can improve compliance without creating unnecessary business disruptions.
**end**
Professional Passport has been calling on Government for over 6 years to use their existing data to stamp out non-compliance in the payment intermediary sector.
In 2014 HMRC applied RTI reporting to all employers, as I am sure most of you know, this provides Real Time Information on every employers payroll by employee.
Then in 2015 Employment Intermediary Reporting came in which gathers data on workers payments, the gross amount, sent to intermediaries. This report is made every 3 months. In the Governments own summary of what these new regulations will do it states:
- support intermediaries that comply
- penalise intermediaries that do not comply
- make sure the right tax and National Insurance is paid by people working through intermediaries
- reduce unfair commercial advantage
HMRC does this by making certain intermediaries provide details of the workers they supply, and the payments they have made to those workers, where they did not operate Pay As You Earn (PAYE).
This cost businesses in the sector significant amounts of money and business disruption to build the required data sets and implement new processes, all of which now appears to be a complete waste as HMRC failed to deliver on any of its stated objectives.
So over the last 6 years we have been pressing HMRC to use the data to achieve their stated objectives. Analysis across both sets of data would immediately give them EVERY provider who failed to comply. Furthermore, the data held by HMRC tends to be correct, data used and shared in the general marketplace is often wrong, or fraudulent.
With the UK economy under pressure and HM Treasury looking to fill the coffers as quickly as possible, AI shows us we don’t have to wait until April 2026. The extra money can start flowing from non-compliant providers from April 2025 without all the complexities and costs in making these proposed changes.
Labour, and Rachel Reeves, have missed a great opportunity to demonstrate they were business friendly, and clamping down harder on tax avoidance than their predecessors by making the right choices.
So why they did make the current choices is anyone’s guess, mine would be: those who have allowed this farce to continue for 10 years are trying to cover up their total incompetency.
This isn’t rocket science.
