error
Subscriptions are not available for this site while you are logged into your current account.
close
Skip to main content

Loading

The page is loading.

Please wait...

      Client
      Anon

      Industry
      Tech, Media & Telecoms

      Primary Goal
      Simplify transfer pricing reporting


      MAKE AI SCALE FAST

      Making transfer pricing reporting easier

      Changes made by the UK Tax Authority to transfer pricing reporting meant that our client – a global tech, media and telecoms business – was having to analyse tens of thousands of rows of data manually.

      We helped the business implement a machine leaning model that gives them a first-draught analysis in a matter of minutes, saving them a huge amount of manual effort. 

      Hitesh Rajani

      Partner, Head of Global Compliance and Transformation for UK

      KPMG in the UK




      90%

      classification accuracy

      2 hours

      to run analysis rather than 3 months

      Monthly insights

      instead of a year-end process





      MAKE AI SCALE FAST

      From manual to AI-enabled reporting

      MAKE REPORTING FASTER

      A bespoke machine learning model


      To pull together the documentation required on transfer pricing, our client was running a line-by-line analysis in Excel. The process of analysing tens of thousands of rows of data could take as long as three months.

      The client approach us to help it simplify the process – taking out time and manual effort. Our solution was to build a bespoke machine learning model to radically cut down the need for manual review of data. 

      woman holding phone
      We built a bespoke machine learning model that was trained on their historical data. This machine learning model was then used to classify their line items and also transform their data, perform validation checks, and then also overlay rules, which then produced the output ready for them to review.

      Navjeet Puwar

      Tax Manager, GCT

      KPMG in the UK

      MAKE DATA WORK

      Training the machine learning with transfer pricing data


      Creating a bespoke machine learning model was not something we could do in isolation. We needed to rain the model on historical data. And that started with working with our client to review their data and determine where we faced issues.

      We ran weekly catch ups, which enabled us to ensure the model was trained on a much data as possible. We transformed the data, performed validation checks and overlaid rules to create transfer-pricing reports that were easily digestible.

      While we’d been tasked with making the process faster, providing accurate reporting was, of course, essential. And we recognised that meant keeping a human in the loop to check results. But that human is now faced with a much less burdensome task.

      Happy young female business product manager presenting analytic data on whiteboard to team of colleagues, partners, standing at large screen with graphs, smiling, laughing, holding meeting
      I think the biggest hurdle was data transformation. I think it was a journey for us and also the client to review their data and actually realise the state that it was in and where certain columns had to be taken away or had to be added in.

      Priyanka Patel

      Senior Manager, GCT

      KPMG in the UK

      MAKE RESULTS FASTER

      From three months to two hours


      The machine learning model can be executed with a simple click of a button, generating results that can be seen within minutes. Any anomalies are flagged and reviewed by the client’s tax team.

      Where before, the client might be analysing tens of thousands of line items, they now only need to check the rows that are flagged. That could mean they’re now reviewing 1,000 line items – it depends on their risk appetite.

      That’s saving our client valuable time it can deploy elsewhere. And what was once a year-end process can now be run monthly.

      people laughing
      Not only are we able to save our client valuable time that they can then redeploy and use elsewhere in the business, we’re also able to save them the monotony of going through line items on a line-by-line basis.

      Priyanka Patel

      Senior Manager, GCT

      KPMG in the UK


      How we’re making the difference

      Our client now has a system that enables them to produce transfer pricing documentation faster. And they have greater confidence in the data they’re presented with.

      As the machine learning system is now firmly embedded within their operating environment, the client is equipped with the foundations necessary to implement additional technologies as part of a future-ready tax function.


      The bespoke machine learning model that we built is over 90% accurate on the classification of the line items. They have more confidence in their data sets.

      Navjeet Puwar

      Tax Manager, GCT

      KPMG in the UK



      Our people

      Hitesh Rajani

      Partner, Head of Global Compliance and Transformation for UK

      KPMG in the UK

      Charlotte Daw

      Partner, Global Compliance & Transformation

      KPMG in the UK

      Kevin Tipton

      Partner, Global Compliance & Transformation

      KPMG in the UK


      How we’re helping our clients drive performance, stay compliant and build sustainable businesses.

      Your essential guide to navigating global minimum tax compliance in 2026 and beyond


      MTD

      Get in touch


      Discover why organisations across the UK trust KPMG to make the difference and how we can help you to do the same.