Accurate and compliant auditing practices have never been more important. The recent Wirecard scandal in Germany illustrates this. As the volumes of data generated by enterprises increases alongside the legal requirements imposed by regulators, compliance is becoming more complicated. A complete transformation of accountancy firms’ business models is needed if they are to become fit for purpose. Engine B is tackling this problem head-on with their AI-powered platform based on Common Data Models and Knowledge Graphs. In this interview with Engine B’s co-founder and CEO, Shamus Rae, we get an insight into how both structured and unstructured data can be managed at scale to ensure companies can confidently ensure compliance.
Martin: Can you tell us a bit about the history of Engine B, who founded it and why?
Shamus: Engine B was founded in 2019, in partnership with the Institute of Chartered Accountants in England and Wales (ICAEW), as a direct response to the burning platform for change within professional services.
Engine B’s origins came from a series of research projects at KPMG. Donne Burrows, Co-Founder and COO at Engine B, and I were working with Imperial College to look at the impact of artificial intelligence (AI) on the audit, tax and legal sectors. An integral part of the project was looking at the blockers that halted transformation. One of the biggest was around client data.
Having just celebrated its first birthday, today Engine B has rapidly increased its market presence as unique audit technology company, and successfully created a product portfolio that leverages common data, AI and analytics to help auditors improve the quality of audits, detect fraud and hidden risks faster and to comply with dynamic market regulations.
As well as having the ICAEW on our Board, we also work closely with Microsoft contributing to their industry-specific Open Data Initiative. We are creating Microsoft’s Common Data Models for professional services.
Martin: What are the company’s core product/service offerings?
Shamus: Engine B is rapidly building next-generation data and AI products for the audit, tax and legal sectors. The bedrock of our platform is our Common Data Models and Knowledge Graphs. When used together, this powerful combination allows the auditor, lawyer or tax expert to access one source of normalised/standard data then use our Knowledge Graphs on top to analyse the data with a context-driven view.
Martin: What makes you different?
Shamus: The combination of our Common Data Models and Knowledge Graphs makes us unique. This is the key to our differentiation. No one else in the market uses a Common Data Model combined with Knowledge Graphs to perform audits. Many audit technology companies still rely on extracting a limited set of ERP data (only taking tiny proportions of data from a small subset of ERP systems), one client at a time, then use a use linear database, that merely highlights risks with no context, and on a spreadsheet to analyse the data! Auditors, therefore, do not get a full picture and can miss risks and fraud.
We exist to solve this problem. Our Audit Common Data Model, which contains both structured and unstructured data, means you only extract data once, the data remains with the client or the audit firm, so any auditor can pick it up and can access it.
We are working with nine accounting firms (including the largest six) who contribute to our data model. We do not want to lock organisations into one ecosystem, Engine B wants to open up the market for interoperable tools working to open source standards.
Our Knowledge Graphs (AI-based) sit on top of the Audit Common Data Model and analyse structured and unstructured data to reveal complex relationships in data, using context, the same way a human brain works.
Non-graph search tools, like the majority in the audit market, work from flat, linear, static-relational schemes, with thousands/millions of rows of data in datasets. This makes discoveries of patterns or anomalies very challenging.
Martin: Why focus on the legal and accounting sectors?
Shamus: Both legal and accounting are sectors that have only just begun to be touched by technology. If you think about sectors like retail or professions like engineering, technology has already had a significant disruptive effect and using advanced technology is the norm. Audit and legal are more conservative professions, but that means that a gap is widening between what people need from accountants and lawyers and what accountants and lawyers can deliver.
For accounting, especially audit, the impact of this is really clear: well-publicised, high profile business failures, massive undetected frauds, and real market instability. Accountancy is struggling to keep up with the volume of data available now, let alone the advanced ways people can manipulate or interfere with that data. Accountancy needs to be brought up to date to make sure markets can still make use of financial information, and that stakeholders can still rely on it.
For law, the impact is similar: lawyers must hold huge volumes of information in their heads, considering the impact of a vast array of relevant cases to every legal judgement they make or each piece of advice they give. It’s just not possible for a single human or even a team of humans to incorporate all of the relevant knowledge and experience, or not without vast resources and cripplingly long work hours. Legal tech can help consolidate knowledge and make it available to the trained professional, summarising what’s relevant and presenting this information to the expert professional for them to apply judgement too. This should make legal professionals more efficient and effective and it has big implications for access to justice, as this knowledge can be pulled together now with technology rather than with armies of paralegals and junior staff which can only be afforded by the wealthiest.
Martin: There have been failed attempts at creating Audit Common Data Models before, why has yours succeeded?
Shamus: There have been attempts made by organisations in the past to create an Audit CDM, however, these attempts have failed. That’s because to create an Audit CDM that transforms professional services practices it takes a huge collaboration effort as well as innovation standards. Our Audit CDM has been successful and is driving the professional services industry forward through our collaboration with Microsoft, the ICAEW and thirteen audit firms. It’s also a non-competitive offering, with one revenue model that doesn’t compete with other professional services players – we are an enabler for transformation. Being Microsoft ‘plug and play’ friendly also helps. Our Audit CDM easily integrates with other Microsoft solutions such as Power BI and Power Apps.
Martin: Can you give some examples of how these tools have helped clients?
Shamus: We’re currently working with one of the Big Four firms, and piloting with mid-tiers, and we’re already seeing the value of the Audit Common Data Model in resolving those gnarly headaches that arise from getting data from multiple different clients which can be used consistently with standard tools. We’re also building tools with adjacent professions to bring fresh approaches to considering and analysing data, especially for audit. The biggest benefit our clients are seeing is the richer context they get in their audit tests when using a Knowledge Graph on top of our Audit Common Data Model – we’re already seeing how this context helps auditors apply their judgement, discarding information that’s not relevant and cutting right to the heart of assurance and risk.
Martin: You describe Engine B as an “AI and audit technology company”. How do you use AI?
Shamus: We use AI within our platform in a number ways, as part of our document understanding process we extract and relate key document elements that show relationships between data in our Audit Knowledge Graphs. We also use graph algorithms within the graphs to highlight anomalies such as potential fraud detection.
Martin: You are working with Microsoft and contributing to their Open Data Model Initiative to create open-source Common Data Models for the professional services sector. Can you talk a little about how this is progressing and how you envisage these data models will be used?
Shamus: Our success with the Audit Common Data Model is also its alignment to Microsoft’s ‘Open Data Initiative’ project. Microsoft’s Open Data Initiative is creating open-source Common Data Models in different industries, such as pharmaceuticals, healthcare, and professional services. Engine B is the only organisation contributing to Microsoft’s data project for audit, making us an integral component to Microsoft’s open data vision for professional services.
Our current Audit Common Data Model is available on GitHub in the Microsoft Common Data Model format. Clients who have Microsoft Azure Data Factory can use this within their own environments to process data and store in a data lake as a source for data.
We are also working on a data ingestion engine that leverages this Common Data Model in our platform.
Martin: Earlier in 2020, you raised £1.7 million from Innovate UK. How is this money being used to grow the firm?
Shamus: The Innovate UK grant has allowed us to accelerate our growth both in our technology and product build and also in team growth, growing to a team of 17 since we founded 18 months ago.
Martin: Does the company have plans to extend its offerings to other sectors beyond professional services?
Shamus: Our products deliver benefits in other sectors as well. For example, in financial services, such as tax, insurance and retail banking where assessments for loans need to be conducted. Tax is another great example of where our Knowledge Graphs can be applied.
The digitisation of tax is moving at a faster pace than audit and other accounting services and at the same time, the spotlight is on the advice given. The pressure on tax firms is increasing, and this demands greater use of the client’s data, both from finance systems and from contracts (suppliers, clients, leases etc) and other unstructured data sources. Our tax specific Knowledge Graphs allow tax advisors to access the data they need in unstructured documents and to create hypothesis views. For example, if you want to know which suppliers are based in China due to a change in import duty or you want to understand which lease contracts don’t have a termination clause then our Knowledge Graphs can present these for you.