5.
How does technology support creating accounts?

5.1 Introduction

The role of information technology (IT) in how accounts are created and maintained has been significant in recent decades. From the introduction of relational databases in the 1970s to collect, store and report financial transactions, to the emergence of enterprise resource planning (ERP) systems in the 1990s and recent developments in generative artificial intelligence (AI) and robotic process automation (RPA), accounting systems continue to evolve to produce more timely and relevant information for decision-makers.

The key role of IT is to facilitate the transformation of data into information.

• Data refers to facts such as the revenue earned from a product or customer.
• Data becomes information when it is analysed, such as when calculations are performed that allow us to understand the business’s most profitable products or customers.
• An accounting information system (AIS) is an increasingly computerised and automated system that collects, processes and stores accounting data for decision-making.

Increasingly, accountants draw on non-traditional or ‘unstructured’ data sources such text, images and video to complement reporting, auditing and decision-making. However, new technologies present challenges to accountants, introducing risks and necessitating the development of new skills to use these technologies and tools effectively. Questions are being asked about whether technology will take over accounting and how the profession can maintain its legitimacy.¹

This unit examines the fundamental and emerging technologies that facilitate the creation of information for decision-making, including the implications and challenges for the profession. It begins with an introduction to accounting software and ERP systems as critical technologies that enable the collection, processing and storage of accounting data. An overview of blockchain technology is also provided to explain how it can support accounting information systems, including its implications for accounting. Subsequent sections focus on understanding RPA, big data analytics, and how AI offers the possibility to automate accounting processes. In the final section, the implications and challenges of these technologies for accounting and the profession are considered.

5.2 Recording, processing and storing accounting transactions

Desktop and cloud accounting systems

Many small and medium-sized enterprises (SMEs) use accounting software to process financial transactions and support accounting processes like preparing income statements, statements of financial position and cash flow statements. However, it must also be recognised that many enterprises still do not use software for their accounting. In the UK, where the government is attempting to digitise the tax system, uptake of digital systems has been seen as a barrier. Read the article Client tech adoption single greatest hurdle for accountants for more information.

Accounting software has existed for decades and was traditionally installed on desktop computers. Today, many SMEs have switched to cloud-based accounting systems like QuickBooks, Xero or Sage. Due to their cloud service delivery model, these systems are accessed remotely and do not need to be installed on a local computer.

The cloud service model has gained popularity recently due to its advantages over traditional desktop systems. One advantage is that businesses can pay a monthly or annual subscription fee. This is different from the past, when up-front payment was required for software licences. These were a significant cost to many firms and could prevent them from operating with up-to-date versions.

Cloud-based accounting systems enable SMEs without an internal accounting function to access and use accounting software managed by an external accountant. These systems can be accessed from any device, including a smartphone, and integrated with various apps available from third parties on the vendor marketplace. In principle, they also allow organisations to scale their operations more easily if they grow than an on-premises system.

Organisations that use traditional desktop solutions miss out on cloud-based accounting systems’ added mobility and flexibility, which allows them to be used by those who do not work in an office and may not usually carry a laptop. A main limitation of cloud services is that they may be subject to downtime if disruptions to the internet or the cloud service occur. They also provide less control over how and where data is stored and security, and when the software is upgraded, and so on, compared to desktop solutions because cloud services are managed externally.²

Cloud-based accounting systems have significantly improved efficiencies in data processing and tasks like account reconciliations. This improvement is attributable to the increasing automation embedded in accounting software, which eliminates the need for manual transaction entry. For example, optical character recognition (OCR) software can automatically read and enter invoices with minimal human intervention. In addition, bank feeds can import transactions directly into accounting software, saving customers and accountants time and helping businesses keep their financial data up to date. This allows accountants to focus on value-added services, as clients perform some of the tasks that accountants previously performed themselves.

The functionality of cloud-based accounting systems can also be extended through integration with apps. For instance, apps can extend cloud-based accounting systems’ capabilities to include forecasting and dashboards; some can even calculate a user’s carbon footprint. The Xero App Store, for instance, offers over a thousand apps for various needs, such as invoicing, inventory management, payments and analytics. However, not all apps may work as intended or seamlessly, which can cause issues in reporting if undetected.

Automation in cloud-based accounting systems has allowed accountants to expand their roles, moving away from transaction data entry and account preparation to advisor roles, including technology and app advisory.

Enterprise resource planning systems

ERP systems from vendors like SAP and Oracle emerged in the 1990s and are used by larger organisations to support their processes, including accounting. They were initially designed for large enterprises with complex operations and scheduling problems to integrate disparate systems. ERP systems use a single database to ensure organisational data remains current and is accessible to various users for decision-making.

Today, ERP systems are also used by medium-sized organisations and can be accessed through the cloud using the cloud IT delivery model described in the previous section. Unlike accounting software, ERP systems include modules that support various functions beyond accounting, such as production, human resource management, customer relationship management, supply chain management and marketing. Since ERP systems provide a data repository from all functions of an organisation, they can also be used to report on sustainability. For example, metrics can be developed to monitor waste, energy consumption and water pollution.

Implementing ERP systems is complex and can take several years, costing many organisations hundreds of millions of pounds. Organisations implementing ERP systems must make many decisions affecting implementation costs like how the system operates. For example, organisations can undertake business process reengineering (BPR) when implementing a new ERP system, which changes how tasks are performed. ERP systems come with ‘best practices’ and organisations may change their processes to align with ERP system workflows. BPR aims to remove non-value-adding activities that may be redundant or unnecessary.

An example of BPR is invoiceless trading (also known as evaluated receipt settlement), where payments for goods and services are based on the quantities received rather than a three-way match between a purchase order, a receiving report and an invoice. BPR can simplify and speed up processes and remove bottlenecks from business processes.

Other issues include data migration, levels of user access to ensure data security, internal skill sets of users, level of technical support available and ongoing training for system updates.

It may take years for organisations to reap the benefits of an ERP system because they must test them and then learn how to use them effectively. Sometimes, this is not straightforward. For instance, Revlon, a multinational manufacturer of beauty products, began implementing SAP in 2018. Soon afterwards, Revlon ran into trouble because manufacturing and warehousing in the new system were not well integrated and employees were not effectively trained to use the system. This resulted in massive backlogs, causing the company to spend significant amounts to expedite shipping. They also lost sales. Revlon also failed to file their 2018 financial reports on time. The company eventually filed for bankruptcy and was delisted from the New York Stock Exchange in 2022.

In the UK public sector, some of Birmingham City Council’s recent financial problems have been linked to the failed implementation of an Oracle cloud-based computer system. Read more about the associated problems and questions raised in relation to value for money and accountability in the Audit Reform Lab report.

While ERP systems improve business processes, financial processes and management control,³ studies suggest that it may take several years for organisations to use the system properly. Research by Flyvbejerg and Budzier (2011) has also shown that up to one in six projects overspend on average by 200%. Also, key financial ratios like return on assets and return on investment may initially deteriorate after implementing an ERP system and may only improve after several years of use.⁴

Blockchain

Blockchain is a distributed ledger technology that stores data. It is maintained by a network of computers rather than a single entity. Users participate in the network and can interact with the blockchain but do not have ownership over the system. Users may own specific digital assets or tokens recorded on the blockchain, such as cryptocurrencies or non-fungible tokens (NFTs).

The records on a blockchain are immutable, which is a significant advantage over other methods of storing data, including relational databases used in conventional accounting and ERP systems. Blockchain processes are complex, involving digital signatures and encryption to verify and add transactions. Once a new block is added to the chain, all blockchain participants’ local copies of the ledger are updated to reflect the new transaction.

To learn more about blockchain, read the article ‘What is blockchain?’ by IBM.

Early applications of blockchain technology focused on creating cryptocurrencies like Bitcoin. However, many other potential applications of blockchain exist. Recently, some vendors have started developing solutions to store invoices on the blockchain to prevent tampering with invoice details like bank account numbers. Blockchain solutions have also been used to trace inventory in the supply chain and enable smart contracts. Smart contracts are useful and cost-efficient because they can automate contractual rights and obligations. They automatically enforce contracts when certain predefined conditions are met (for example, when a customer pays a fee), thereby reducing transaction costs.⁵

Organisations can use blockchain solutions in accounting for different purposes affecting the scope of the implementation. They can use private or public blockchains.⁶ Private blockchains are controlled by a central authority (for example, a business), while public blockchains like Bitcoin are decentralised. For example, a private blockchain can be used to store source documents to prevent the manipulation of records. A private blockchain can also be used to record accounting transactions. However, the benefits of blockchain technology are more significant when shared with other organisations or when blockchains interact with each other. Such blockchains can eliminate certain accounting tasks such as reconciliation because the accounts are transparent and validated by all participants on the chain as transactions occur. With triple-entry accounting, where organisations share ledgers and sign-on transactions, blockchain can significantly reduce accounting fraud and financial statement manipulation.⁷ Blockchain is, however, more expensive and energy-intensive than a traditional database and if errors are found, data can be challenging to correct.⁸ Full-scale applications of blockchain in accounting are yet to be developed as organisations and vendors face significant technical and economic challenges to develop relevant solutions.⁹

5.3 Technologies for accounting automation and the analysis of data

Robotic process automation

Robotic process automation (RPA) is a technology that mimics human action. RPA bots can read emails, copy data into a form, enter data into a database and send emails to solicit human input.¹⁰ Through RPA, routine tasks that are repetitive and time-consuming can be automated. Bots can also perform tasks more quickly and accurately if configured correctly. Accounting tasks like financial close, consolidation, accounts payable, billing and bank reconciliations are highly automatable.¹¹

To automate processes, tasks need to be defined and standardised so that rules can be programmed to execute them. RPA fails when bots encounter unexpected data or when the data quality is low. In contrast, intelligent RPA can handle exceptions and workflows when judgement is required. RPA has been deployed by firms who have sought to automate routine tasks where possible to save costs and free up time for other tasks and decision-making.

Before a process can be automated, it first needs to be defined and mapped at the keystroke level. Process mapping techniques include data flow diagrams, system flowcharts and process maps, which visually depict the tasks, the data and the actors involved in a process, as well as any decision points to consider (for example, amounts reconciled or not). Process mapping techniques are also useful for evaluating existing processes, for example, determining whether internal control weaknesses are present.

Reported benefits of RPA include lower costs and better report quality.¹² RPA also improves work satisfaction because employees can focus on tasks that require creativity and critical thinking. RPA can also bring back outsourced processes in-house.¹³

Big data, data analytics and data visualisation

Organisations are increasingly looking to develop more sophisticated forms of data analytics to ensure evidence-backed decisions. To some extent, these systems rely on big data. Big data is the term used to refer to data sets that are distinguished by six Vs: volume, velocity, variety, veracity, value and visualisation.

• Volume refers to the significant amount of data firms can collect and analyse because of connectivity provided by the internet.
• Velocity refers to the speed at which technology can generate and process data.
• Variety refers to the various available data types ranging from highly structured data like transactional data to unstructured data like social media images and videos.
• Veracity considers the trustworthiness and accuracy of data; for instance, do you know its origin, and can you judge its credibility?
• Value refers to whether the data permits meaningful insights to be extracted and used to help the organisation achieve its objectives.
• Visualisation as a sixth dimension or characteristic has recently been proposed¹⁴ to highlight the importance of meaningful representations when making decisions.

Considering the rapidly changing business environment, organisations of all types are now considering the potential of data analytics to provide new insights. This includes identifying patterns in data and shifting the focus from historical data to forward-looking data to improve decision-making. This represents a shift in thinking from collecting data to considering how it can be processed and used to help organisations make decisions. Computers can perform data analytics more effectively than humans; they can quickly process large data volumes and different data types.

There are various software programs that firms can use for data analytics and visualisation, such as Microsoft Power BI and Tableau. Microsoft Excel still offers features capable of performing data analysis when the data set is not overly large. To work best, the data needs to be cleaned, which means that you need to ensure there are no spelling mistakes or duplication of cells, remove spaces, fix number signs (that is, negative signs), and so on. You also need to consider the data quality.

Data analytics can be descriptive, diagnostic, prescriptive and predictive:

• Descriptive and diagnostic analytics use historical data to identify patterns and trends. Descriptive analytics can calculate performance, whereas diagnostic analytics are concerned with explaining that performance.
• Prescriptive and predictive analytics are future-focused. Predictive analytics use algorithms to identify patterns in historical data and predict what will happen. For instance, they could calculate expected future sales using past historical data. Prescriptive analytics are concerned with the actions needed to achieve an objective. For instance, a firm may wish to optimise inventory levels, minimise costs and ensure product availability. Prescriptive analytics would consider the demand forecast, supplier lead times, bulk order discounts and inventory holding capacity to understand how many units to order and when to order.

The key decision-maker needs to understand insights gleaned from data analytics. Complex data modelling may be too technical for some users to understand, requiring the accountant to use different forms of data visualisation to share information. For instance, instead of providing the decision-maker with a complex model or spreadsheet with thousands of data points, it could be summarised in a line graph or bar chart. The added benefit of data visualisation is that there is a pictorial superiority effect, which simply means that it is more memorable. However, accountants must also be wary of developing cluttered charts their audience does not understand. Moreover, information is likely more persuasive if told as a story and not simply presented as facts. The type of visualisation used should be tailored to the audience, particularly to the person who ultimately needs to make a decision.

Machine learning and AI

Many analytical techniques discussed in the previous section utilise AI, like machine learning, to find patterns in data and to make predictions. Machine learning is a subset of AI and includes supervised and unsupervised learning. Unsupervised machine learning employs techniques such as clustering, which finds patterns in data without the need to train the model with labelled data. Patterns are identified based on the features of the data.

In contrast, supervised learning requires both training and testing a model. The ‘machine’ learns to predict or classify data using labelled data to train the algorithm. For example, banks constantly decide which customers to extend credit to. The algorithm would initially be trained with data that includes labels like approved and rejected customers for loan applications. Once trained, the algorithm will learn to recognise risky customers when new applications are evaluated.

Generative AI and large language models using machine learning, such as ChatGPT and Google’s Gemini, are increasingly changing workflows. For instance, McKinsey (2024) reported that 65% of respondents to their recent survey use AI at work in at least one business function. A Gartner (2024) survey of CEOs found that they believe AI is the technology most likely to impact their industry. Generative AI uses algorithms that draw on existing artefacts to create new content or artefacts like videos, images and audio. Large language models focus on textual data. Custom generative AI and large language models can be developed and enhanced using proprietary data if organisational resources permit.

AI may disrupt accounting work in various ways, including helping to identify and leverage new revenue channels by analysing market data, optimising costs, generating reports, improving risk assessment, and modelling scenarios. Accounting software vendors like Sage increasingly incorporate copilots (AI assistants) using large language models to automate workflows aimed at continuous accounting, assurance and insights.

It is important to acknowledge that AI consumes significant energy and resources and that this may have an impact on the drive for organisations to become more sustainable. Read the following articles to learn more about AI’s energy consumption and progress towards improving efficiency and use of renewable energy:

• Will AI accelerate or delay the race to net-zero emissions?
• AI is an energy hog. This is what it means for climate change.

A concern when using AI is that it might be trained using biased datasets. This happens when the data used to train the algorithm does not represent the population. Other problems include cost overruns and vulnerability to cyberattacks. There are many examples in which AI implementations have failed, such as Amazon’s recruitment solution, which was gender-biased. Human accountants thus have a key role to play in evaluating and monitoring AI use in organisations.

5.4 Implications and challenges for accounting and the profession

Pouring money into new technologies, including accounting information systems, does not necessarily translate into better organisational processes and outcomes (recall the Revlon example discussed in Unit 5.2). New systems and technologies may not succeed if organisations are unprepared, such as when they have not identified technology skills gaps and fail to manage change effectively. Accountants can play a key role in appraising new technologies like RPA and AI to ensure that the suggested solutions are a good fit for the organisation and that the risks, including cost overruns, are mitigated by breaking down large projects into smaller ones. Accountants also need to consider how existing organisational controls are affected by new technologies like automation and AI.

Organisations may use AI to support decision-making. Some algorithms that AI uses for decision-making, like neural networks, can be hard to understand and explain because of the many hidden layers in processing the data to make a recommendation. Managers may refuse to consider AI recommendations if they do not understand the underlying assumptions or if the recommendation is counterintuitive. For instance, explaining to a customer why a credit sale was declined may be challenging if the underlying model is opaque.

Coupled with automation, AI-enabled bots are expected to work autonomously to execute and manage business processes. Organisations might employ several employees to oversee accounts payable and accounts receivable. This is often viewed as important as it ensures separation of duties for placing orders with vendors, receiving goods, approving invoices and approving payments to minimise fraud. In very small organisations, however, separation of duties is not always possible as it is unaffordable, and the control function in these circumstances is normally undertaken by the owner. Some questions that need to be considered include the following:

• Is a similar separation required if these functions are delegated to intelligent bots? What if a fraudulent employee takes control of a bot and, thereby, an entire business process?
• More generally, who should be accountable if decisions are based on incorrect recommendations by AI and consequently lead to sub-optimum decisions and strategies? The vendor of the bot, those in charge of configuring the bot or the manager of the business (for example, the accounts payable manager)?

These questions must be answered before AI can be successfully rolled out in organisations.

In a data-driven economy, organisations are keen to collect vast amounts of data, particularly from customers, to turn data into useful information for decision-making. While more data can be useful and relevant for drawing insights into patterns for prediction, companies often collect it without generating value from it. Before investing in data analytics, it is important to know the key business questions you want to answer so you collect the right data. This is critical because the storage and governance of data can be costly, risky and subject to laws.

For instance, in Europe, data laws like the General Data Protection Regulation (GDPR) and the AI Act aim to protect people. Cybercrime is widespread and many organisations experience data breaches or other cyberattacks like ransomware, resulting in disruptions to operations, loss of customer confidence, and fraud. Organisations that breach privacy laws also pay hefty fines to regulators. For example, Facebook was fined a record five billion US dollars in relation to the Cambridge Analytica data breach, when Facebook shared the profile data of its users with a third party without their consent.¹⁵

Faulty vendor upgrades may also take systems down, such as the 2024 CrowdStrike meltdown that affected millions of Windows computers running security software. For a time, many industries were disrupted, including transportation, healthcare and financial services, which cost the world economy billions of pounds.¹⁶

Vulnerabilities related to data governance, including privacy, pressure organisations to develop more effective digital risk management. Accountants should play a key role in safeguarding organisational data as they have traditionally done.

Technologies are advancing quickly. To effectively use new technology such as AI or analytics and manage its risks, accountants must keep up with technological developments and upskill. As technology takes over some of the technical and repetitive tasks that accountants have traditionally undertaken, people and critical thinking skills will become increasingly important. Accountants will be freed up to be able to consider holding organisations accountable for their moral and social obligations, and tackling some of the biggest problems facing the world today.

5.5 Summary

• Organisations rely on accounting information systems to capture, process and store accounting transactions.
• Automation and AI free up time, allowing accountants to focus on strategic tasks rather than repetitive ones.
• New technologies create risks for organisations that accountants must recognise and manage.
• Information security, including data privacy, is a key issue that organisations must prioritise to protect sensitive information, comply with regulations, and maintain trust with customers and stakeholders.
• Accountants need to upskill to play a key role in assessing new technologies’ various opportunities and risks.