How Can Digital Strategy Help Companies Be Compliant?

The rapid acceleration of modern technological advancement portends exciting things for the world of compliance, even as it outpaces existing regulations.

Russell Sheldon, chief operations and technology officer at TMF Group, shares his predictions for the future of digital technology and its implications for the legal sector.

Artificial intelligence (AI), big data and blockchain are changing everyone’s world – and that is not even the half of it. Virtual reality, digital biotechnology, nanotechnology, digital manufacturing, the internet of things (IoT) and ‘connected everything’, limitless computing power in the cloud, and robotics are all standing in line.

The short answer to the titular question is ‘because it must’. Compliance professionals cannot refuse the hurricane of technological change sweeping across business, finance and wider society any more than we can hide from the regulatory blizzard itself.

Compliance on the front line

Legal and regulatory compliance should be on the front line of digital transformation because there is a compelling reason for regulator and regulated alike to embrace this new world. In a word, that is complexity.

The idea that human experts can always provide definitive answers is little more than a comforting illusion. In our hearts – and from our experience – we know this. Even they have human limits to their knowledge, experience and information processing. As Edgar Fiedler famously said of economists: ‘Ask five and you’ll get five different answers – six if one went to Harvard’.

For regulators, the sheer volume and pace of modern human activity has made the scale of their task almost impossible. The French tax authorities now use AI to scan online platforms like social media, Airbnb and classified adverts for signs of fraudulent under-reporting. They combine AI with aerial photography to find tax cheats by locating their undeclared swimming pools. In 2019, AI-driven data mining helped France collect an extra €640 million in personal taxes. In South Africa, data-driven, self-learning AI now holds the key to modernising the tax system, cutting fraud and rebuilding public trust.

For regulators, the sheer volume and pace of modern human activity has made the scale of their task almost impossible.

Meanwhile, for the regulated and their advisors, the growing weight and complexity of law and regulation (especially for cross-border businesses), as well as the speed with which the landscape changes, has had a comparable effect. Fortunately, many of these challenges are almost tailor-made for AI.

From mindless chat to deep learning

AI is the buzzword of our times, covering anything from chatbots to a self-driving cars, so we should define our terms. The potential to automate the routine and complex alike is at the heart of this buzz.

At the simplest end of this automation is robotic process automation (RPA) which excels at repetitive, high-volume, data entry-type tasks like credit card applications processing. But when its rules run out, so does the ‘thinking’.

Chatbots make use of natural language processing (NLP) to mimic human speech and text-based interactions. We are all used to these in our personal lives with the likes of Siri or Alexa, but we are seeing increasing examples of how chatbots are helping to provide answers or guided advice quickly and consistently (and 24/7). Examples include providing responses to frequently asked questions for KYC activities or automating the generation of NDAs using keyword detection.

Autonomous vehicles (AVs) use ‘deep learning’, the pinnacle of modern AI. Their neural networks draw on vast quantities of often real-time data to acquire and evolve an expanding and ongoing awareness of what they need to do next. For all their bad press, AVs already do things that can seem, as Arthur C. Clarke famously put it, ‘indistinguishable from magic’.

In between these extremes sit the many and varied forms of ‘machine learning’, which also combine great processing speed, large and multiple data sets and the capacity to learn and self-improve, but at a lower level of sophistication. Insurance companies already use machine learning AI to assess the likelihood that a transaction is fraudulent by monitoring 50-odd different data points, all in near real time. Perhaps most encouragingly, machine learning AI systems are now able to identify tumours in lung cancer patients much faster and earlier than expert radiologists.

For all their bad press, AVs already do things that can seem, as Arthur C. Clarke famously put it, ‘indistinguishable from magic’.

In compliance, where regulatory information load is such a big problem, this kind of augmented decision-making has a natural home. A combination of natural language text processing and a machine learning engine can help keep clients on top of legislative, regulatory and tax changes across many countries. These solutions scan any data source they are trained for (including internet sources) and consolidate the relevant information. Adding natural language generation capabilities can even create summaries that a human can work with.

The right combination of solutions can free compliance professionals to work smarter by helping them frame better questions, including some they never realised needed asking. What if one person is a director across a very large number of jurisdictions and her statutory duties clash? Perhaps country X will not accept her precisely because she has a role in country Y? The traditional, static, two-dimensional picture of entity compliance information is of limited use here because the risks hide inside the dynamic relationships. By bringing data together from a variety of sources (the entities themselves, KYC information, due diligence information, regulatory and legal updates) and using AI to perform complex analyses across jurisdictions, persons and roles, it is possible to pinpoint long-buried risk factors.

Bigging up data

In 2018, Gartner grabbed the headlines by saying that 85% of AI projects do not deliver because a poor understanding of AI led to it being fed poor quality or insufficient data. Today, data is nothing like the obstacle to digitalisation it once was. Better tools enable us to find meaning in data that is messy and imperfect.

Data privacy is now a process we know how to manage. Data sovereignty is a growing concern, especially if you are trying to do your global administration on a single consistent platform. Governments like Russa, China, Singapore and Switzerland are stopping certain data from leaving their jurisdictions. But will other governments follow suit? And what will that mean for digitalisation? Time will tell.

Today, data is nothing like the obstacle to digitalisation it once was.

Blockchain unblocked

In confidential matters of trust, the chain of security and confidence has only ever been as good as the weakest, and leakiest, human link. All that is changing.

Distributed ledger technology is the foundational technology of blockchain. It is best known for its cryptocurrency applications, but it will soon usher in a fundamentally different approach to many aspects of transparency and trust in transactions. The consequences for compliance and legal professionals will be profound.

Blockchain manages the complexity of trust, transparency and certainty by granting all parties on the network simultaneous and secure access to precisely the same data. This removes the need for trusted intermediaries to facilitate interchange and exchange. Time currently spent making and remaking complicated documents, and establishing the evidence of their authenticity, will be replaced by a one-off need to codify the information and knowledge into the blockchain technology landscape. ‘Smart contracts’ will be able to automate agreement outcomes between two parties and grant auditors, advisors and even the authorities the ability to see all that they need to see about a transaction, with certainty and in real time.

The ability to prove that someone is who they say they are, and that their source of wealth is legitimate, is compelling. Regulators and governments will start to apply blockchain capabilities (which includes the ability to combine multiple blockchains) to tackle money laundering and financial crime. In taxation, blockchain will form the basis of future record-keeping, leading to fundamental changes in how we collate data and make submissions.

Taking the helm or missing the boat

These are exciting times. We have already come such a long way and continued digitalisation is inevitable and accelerating. We all need to be thinking about how we can make this new world work for us.

Today, success in digital transformation initiatives, like any business change, often boils down to whether we can make change ‘stick’. It is never easy, but it is less hard with some guiding lights. Here are mine:

  • Always start with the business problem you want to solve, not the technology.
  • Change and transformation is not a destination. It is a journey with no end. Not only is the technology in constant flux, but so too is the business landscape we are applying it to.
  • Keep sustainability front-of-mind. Think in terms of end-to-end processes. And do not expect to implement and walk away – plan to stay close and connected to the project because you will always be learning.
  • Start small. A pilot or proof-of-concept model first, then proceed incrementally (country-by-country or function-by-function). Large scale and top-down? It hardly ever works these days.

Finally, sometimes you must accept that the change we seek just will stick, even though we have done everything right and worked with colleagues in the business who are enthusiastic and fully engaged. Do not worry. Another one will be along in a minute. Good luck on your adventures.


Russell Sheldon, Chief Operations and Technology Officer

TMF Group

Tel: +44 (0) 20 7832 4900



Russell Sheldon boasts over 20 years’ worth of experience leading operational and technological change in global professional services firms, including AON Hewitt, Axiom Law, NGA Human Resources Nd PricewaterhouseCoopers. Prior to joining TMF Group, he was chief client officer at technology solutions firm Sabio.

TMF Group is an international provider of administrative services. Its 9,100 experts provide legal, financial and employee administration through the Group’s 120 offices worldwide.

Leave A Reply