1
Artificial Intelligence and Automation
1. What is Artificial Intelligence?
Artificial Intelligence (AI) comprises a set of interconnected technologies aimed at
emulating and replicating human cognitive functions, encompassing processes such as
thinking, speaking, feeling, and reasoning. These technologies leverage computers to
address domains that necessitate knowledge, perception, reasoning, comprehension,
and cognitive abilities. To attain these capabilities, computers must possess the capacity
to:
• Grasp common sense.
• Comprehend facts and manipulate qualitative data.
• Handle exceptions and disruptions.
• Discern relationships between facts.
• Engage in human-like interactions.
• Adapt to new situations based on prior learning.
AI involves various associated technologies, including robotics, expert systems, fuzzy
logic systems, intelligent agents, artificial neural networks, genetic algorithms, and
natural-language processing. While these applications and technologies may not
replicate genuine human intelligence, they unquestionably exhibit a higher level of
intelligence compared to conventional information systems.
2. AI Technologies
2.1 Robotics
Robots stand out as one of the highly successful applications of AI, excelling in executing
straightforward and repetitive tasks. Their deployment proves valuable in liberating
workers from monotonous or perilous assignments. Notably, personal robots have gained
significant attention lately, featuring limited mobility, restricted vision, and some speech
capabilities. Primarily utilised as prototypes, these robots undergo testing for services like
assisting the elderly, serving breakfast, cooking, opening doors, and transporting trays
and beverages.
2
The advantages that robots bring to the workplace include:
• Absence of emotional factors such as falling in love with coworkers, feeling
insulted, or calling in sick.
• Consistency in performance.
• Applicability in environments hazardous to humans, like tasks involving
radioactive materials.
• Freedom from engaging in activities like espionage for competitors, seeking
salary raises, or advocating for extended breaks.
2.2 Expert Systems
An expert system is comprised of programs designed to replicate human thought
processes within a particular domain where human experts have effectively solved
problems. The success of expert systems is contingent upon their application to activities
that human experts have previously mastered, spanning fields like medicine, geology,
education, and oil exploration.
In essence, heuristic data, while not constituting formal knowledge, serves the purpose
of aiding in problem-solving without adhering to a strict algorithmic process.
2.2.1 Uses of Expert Systems
Numerous companies are actively involved in the research and development of expert
systems, leading to their utilisation in various domains, including the:
• Airline industry
• Forensics lab work
• Banking and finance
• Education
• Food industry
3
2.2.2 Criteria for Using Expert Systems
An expert system is deemed suitable for use when one or more of the following conditions
are met:
• The problem requires a significant amount of human expertise, but a single expert
is insufficient to address it alone.
• The necessary knowledge can be expressed through rules or heuristics, lacking a
well-defined algorithm.
• Human experts have already successfully resolved the decision or task, enabling
the expert system to emulate their expertise.
• Consistency and standardisation are crucial for the decision or task.
• The subject domain is well-defined and limited.
• The decision or task involves numerous rules (typically ranging between 100 and
10,000) and intricate logic.
• There is a shortage of experts within the organisation, or key experts are
approaching retirement.
2.2.3 Advantages of Expert Systems
An expert system offers several advantages over humans, including:
• Consistency: It never becomes distracted, forgetful, or tired, making it well-suited
for monotonous tasks that human workers might find objectionable.
• Expertise Duplication: It duplicates and preserves the expertise of scarce experts,
allowing the system to incorporate the knowledge of multiple experts.
• Knowledge Preservation: It preserves the expertise of employees who are retiring
or leaving an organisation, ensuring that valuable knowledge is not lost.
• Decision-Making Consistency: It creates consistency in decision-making
processes and enhances the decision-making skills of nonexperts.
3. Intelligent Agents
Intelligent agents, also known as bots (short for robots), are software entities capable of
reasoning and adhering to rule-based processes. They are gaining popularity, particularly
4
in e-commerce. A sophisticated intelligent agent exhibits characteristics such as
adaptability, autonomy, collaborative behavior, a humanlike interface, mobility, and
reactivity.
One notable application of intelligent agents, already in use, is in Web marketing. These
agents can gather customer information, including items purchased, demographic details,
and both expressed and implied preferences. Intelligent agents are also employed in the
creation of smart or interactive catalogs referred to as "virtual catalogs." These catalogs
present product descriptions based on customers' past experiences and preferences.
Currently, available intelligent agents are categorised as follows:
• Shopping and information agents
• Personal agents
• Data-mining agents
• Monitoring and surveillance agents
3.1 Shopping and Information Agents
Shopping and information agents play a crucial role in assisting users in navigating the
extensive resources available on the Web, yielding improved results in information
retrieval. These agents possess the capability to traverse the Web at a faster pace than
humans, collecting more consistent and detailed information. They function as search
engines, site reminders, or personal surfing assistants. Additionally, Usenet and
newsgroup agents incorporate sorting and filtering features to facilitate the process of
locating information.
3.2 Personal Agents
Personal agents are designed to execute specific tasks on behalf of a user, serving
functions such as recalling information to complete web forms or automatically suggesting
email addresses after inputting the initial characters.
5
3.3 Data-Mining Agents
Data-mining agents operate in conjunction with a data warehouse, identifying patterns
and uncovering novel information and relationships among data elements that may not
be immediately apparent.
3.4 Monitoring and Surveillance Agents
Monitoring and surveillance agents typically track and report on computer hardware and
network systems, predicting potential system crashes or failures.
4. Contextual Computing: Making Mobile Devices Smarter
For years, people have utilised GPS applications on their smartphones, particularly
beneficial for navigating unfamiliar cities. These applications hold considerable
commercial value. The concept of contextual computing is expected to take this idea even
further.
Human decision-making relies on knowledge and emotional responses, drawing from
accumulated life experiences. Often termed the sixth, seventh, and eighth senses,
contextual computing envisions a perpetual computing environment that comprehends
people's surroundings and, based on factors such as their identity, location, and
company, provides personalised recommendations. The fundamental idea behind
contextual computing is that computers can not only sense but also respond to their
surroundings in a manner akin to how the human brain comprehends stimuli. Essentially,
contextual computing enables the customisation of actions for a user in a specific situation
and environment, leveraging information known about the user.
5. AI and Automation
The rapid proliferation of machine learning and AI is evident in self-driving cars, customer
service software with responsive capabilities, and robots managing tasks like assembly
lines, burger flipping, pizza cooking, and inventory checking. Robotic process automation
(RPA) proves advantageous when strategically deployed for specific applications,
offering the following benefits:
6
• Cost effectiveness: RPA has the potential to reduce operational costs by 25–50
percent.
• Improved accuracy and quality: By minimising human errors and eliminating
rework, RPA enhances overall accuracy and quality.
• Consistency: RPA ensures consistent performance of activities across various
tasks.
• Improved analytics: Access to precise data from diverse sources enhances the
quality of analytics.
• Increased employee productivity: RPA liberates employees from mundane tasks,
allowing them to concentrate on mission-critical activities, such as customer
service.
• Increased customer satisfaction: RPA contributes to the delivery of high-quality
products and services, thereby enhancing customer satisfaction.
• Speed: RPA accelerates the execution and delivery of products and services.
• Reconciliation from multiple systems: RPA facilitates the reconciliation of data and
information from various systems without requiring human intervention.
• Versatility: Applicable across industries, RPA demonstrates the ability to perform
a diverse range of tasks.
6. Ethical issues of AI
To enhance the likelihood of success and mitigate risks in the adoption of AI,
organisations should establish robust governance and ethical frameworks,
considering the following five questions before deploying the AI system:
7
• Defining the AI goals: Clearly articulate the objectives and purposes of the AI
system.
• Defining the complexity of the problem: Assess and define the complexity of
the problem that the AI system is intended to address.
• Defining the environment as stable or variable: Determine whether the
operational environment for the AI system is stable or subject to variability.
• Defining and guarding against bias: Explicitly identify and address potential
biases in AI algorithms, considering the influence of developers and designers.
• Defining the level of human involvement: Establish the appropriate degree of
human involvement in AI system processes.
The widespread adoption of AI technology raises several ethical issues that demand
careful examination, including:
• AI bias: The possibility of biases in AI algorithms influenced by the perspectives
of predominantly male programmers and software engineers.
• AI mistakes: Occurrences of "artificial stupidity" when deploying incompletely
developed AI programs.
• Wealth inequality: Concerns that AI could exacerbate the gap between the rich
and the poor.
• Humanity: Potential impacts on human behavior and society due to the
widespread deployment of AI technology.
• Unemployment: The anticipated influence of AI on employment, potentially
leading to changes in job landscapes.
7. Impact of AI on Accounting
How can the integration of AI with other technologies enable accountants to concentrate
on tasks that are high in value and impact?
Accountants play a crucial role in various aspects of business operations, encompassing
the recording of transactions, organisation, and analysis of financial data, and presenting
8
information in compliance with regulations and for decision-making purposes. Managing
a vast array of financial and non-financial data, accountants are also tasked with ensuring
timely, accurate, and consistent financial reporting to meet the diverse needs of internal
and external stakeholders.
Given the constant changes in regulations, especially for companies operating across
different areas, this is a challenging endeavour. To tackle this, organisations have started
leveraging the combination of AI with technologies like robotic process automation. This
integration aims to automate routine tasks traditionally performed by accountants, freeing
up their time to focus on high-value, impactful responsibilities.
By incorporating AI, particularly in conjunction with robotic process automation,
accountants can reallocate the time previously spent on mundane tasks to more high-
value activities.
Moreover, the introduction of AI into accounting processes can enhance output quality by
minimising human errors. AI's ability to emulate human interactions, understand inferred
meanings in communication, and adapt to activities using historical data contributes to
real-time updates on financial matters. This is achieved through processing documents
via Natural Language Processing (NLP) and computer vision at an accelerated pace,
facilitating cost-effective daily reporting. This, in turn, facilitates better and more timely
insights crucial for swift decision-making.
In audits, auditors employ machine learning tools to analyse documents, such as sales
and lease contracts, conduct trend analysis, and identify outliers. This minimises
administrative time spent on reviewing audit documents, enabling auditors to dedicate
more time to areas requiring significant estimates and judgment.
As part of the digitalisation of the audit process, accounting firms can implement an AI
proof of concept utilising computer vision to deploy airborne drones for monitoring
inventory during audits. For instance, a drone can accurately count the number of vehicles
9
in a production plant under audit and relay this data directly into the global audit digital
platform. This innovative use of AI in audits is currently being adopted by firms like EY.
8. Benefits of Artificial Intelligence for Accountants and Finance Professionals
Recent technology is revolutionising work processes across industries, and accounting is
no exception. Artificial Intelligence (AI) holds the potential to significantly enhance the
productivity and efficiency of accountants. The integration of AI into accounting operations
can lead to an 80-90% reduction in the time required for tasks, allowing accountants to
shift their focus to providing valuable counsel to clients. Additionally, the implementation
of AI is expected to elevate the quality of accounting services by minimising errors.
When accounting firms embrace AI, they not only boost their attractiveness as employers
but also position themselves as preferred service providers to millennials and Gen Z
professionals. These tech-savvy generations, accustomed to the latest innovations, seek
employers who leverage technology to offer flexible schedules, and remote work options,
and relieve them of mundane tasks better suited for automation. As clients, millennials
and Gen Zers are likely to choose service providers based on their ability to deliver
insights enabled by automation, giving AI-adopting firms a competitive edge.
Robotic Process Automation (RPA) and Intelligent Automation (IA) are key components
of AI implementation in accounting. RPA streamlines repetitive and time-consuming
tasks, freeing up accountants for more strategic work. IA, a more advanced version,
mimics human interaction, understanding implied meanings and adapting to activities
using historical data. These technologies find applications in various accounting tasks,
providing real-time financial status, enabling daily reporting, and supporting strategic
decision-making.
AI-driven technologies also enhance internal accounting processes, including
procurement, invoicing, purchase orders, and accounts payable and receivables. They
contribute to regulatory compliance by monitoring documents against rules and laws, and
flagging issues that require attention. Machine learning algorithms play a crucial role in
10
fraud detection, quickly analysing vast amounts of data to identify potential issues that
might go unnoticed by human auditors.
Despite the evident benefits, the primary challenge to AI adoption in accounting lies in
overcoming resistance and fostering a mindset shift among professionals. Executive
understanding of AI's strategic advantages is high, with 85% acknowledging its potential
to confer or sustain a competitive edge. To facilitate the acceptance of AI among
accountants, it is essential to communicate the benefits, provide comprehensive training,
and offer the necessary support for optimal utilisation.
The trajectory of AI and automation in accounting and finance is still in its initial stages,
but the technology is evolving rapidly. Accountants who embrace these changes stand to
gain significant advantages in terms of time and cost savings, as well as the valuable
insights AI can provide. Those resistant to these advancements may find it challenging to
keep pace with competitors leveraging AI for enhanced efficiency and service quality.
9. Examples of Artificial Intelligence in Accounting
AI offers several advantages in the realm of accounting, including:
• Fraud Detection and Prevention:
Computers equipped with AI can analyse patterns and meticulously monitor every
document within a company. This enables the system to cross-verify transactions for
compliance with accounting rules and laws. AI is adept at flagging potential discrepancies,
allowing human auditors to conduct thorough checks and enhance fraud detection and
prevention measures.
• Hidden Insights Discovery:
AI can unveil hidden patterns, trends, and insights within large datasets. By leveraging
machine learning algorithms, AI systems can identify correlations and provide valuable
business intelligence. These insights not only offer a competitive advantage over rivals
but also empower companies to make informed and timely decisions based on robust
11
data analysis. The speed and accuracy of AI-generated insights contribute to agile
decision-making processes.
10. Summary
Artificial intelligence systems possess significant power and are advancing rapidly,
offering outputs with exceptional accuracy that, in certain instances, surpass human
capabilities. It's crucial, however, to acknowledge that these systems do not replicate
human intelligence. Understanding the strengths and limitations of this distinct form of
intelligence is essential. Developing insights into the most effective collaboration between
humans and computers is key for optimising their collective potential.
