AI in finance

Artificial intelligence is rapidly transforming the finance landscape, moving beyond accounting and making significant inroads into financial reporting, management, planning and analytics. A report from KPMG International reveals that nearly three-quarters of finance teams across diverse industries and company sizes are already using AI to some degree to enhance their financial reporting processes, implementing AI across wider areas of finance, including financial planning, treasury management, risk management, and tax operations.

Fewer than one in five Canadian organizations have adopted AI in the tax function, and many have yet to make the leap from experimenting to actively using AI for day-to-day tax processes. Complex tax regulations, a lack of up-to-date data, and onerous legacy systems for many tax-related decisions are still barriers to implementing AI in the tax function for many organizations.

At this point, many Canadian organizations have focused on ‘classical’ AI, which is founded on traditional data analytics but augmented with machine learning. Only 19% indicated they were selectively or widely using generative AI within the finance function over the previous six months, while 32% said they were piloting it and 42% were in the planning phase. Globally, 28% of organizations are selectively or widely using AI in financial reporting — and that’s expected to jump to 83% over the next three years.

Some 47% of Canadian respondents identified generative AI as a top investment priority over the next three years — the highest of any technology category. While there’s huge interest in generative AI, the average Canadian organization isn’t there yet. The challenge they face is translating it into real use cases in finance. Many Canadian organizations are at a standstill because they don’t know where to start, or there’s resistance to taking the next step because of potential risks.

Globally, business leaders are using generative AI for dynamic reporting and narrative generation, forecasting models and scenario generation, document management, compliance monitoring and reporting, and automated tax preparation, among other applications.

In the finance and risk function, the top AI use cases being piloted or implemented by Canadian organizations include:

• Fraud detection and prevention (54%)
  
• Research and data analysis (50%)
  
• Predictive analysis and planning (50%).

The key to moving forward is being very deliberate about how organizations choose their use cases so they link back to enterprise value.

However, many organizations start by looking for a problem to solve with generative AI versus looking for an outcome they’re trying to achieve — and evaluating whether AI should or shouldn’t be a solution for it. Rather than looking at what to ‘fix’ with AI, it’s about looking at pain points and finding solutions to help organizations become more efficient and productive. For example, one Canadian bank surveyed by KPMG is merging AI with blockchain to ensure secure and transparent financial transactions.

Generative AI in and of itself isn’t going to solve all these pain points, but it’s one component of a broader transformation journey. So, while use cases are important, it’s also critical to come up with a broader vision and strategy for the finance function that is tied to value.

According to KPMG International, the top barrier to AI adoption globally is the potential for data security vulnerabilities. In Canada, the top barrier was tied between data security vulnerabilities and limited AI skills and knowledge (both cited by 56% of Canadian respondents). Following closely behind was the challenge of gathering consistent data (cited by 51% of Canadian respondents).

Overcoming these challenges is key to unlocking the benefits of AI. To launch innovation, organizations need data management and data governance in place. At the same time, their people need the necessary skills to sustain AI initiatives over the long term. It’s critical that humans are at the centre of this transformation with human-centric application design.

Many employees in the finance function are accountants by trade, so they don’t necessarily have the level of digital acumen or data literacy to be able to articulate the use cases that would be most beneficial to them. That means business leaders need to build their own internal resources (either a central team within finance or separate groups within each department) and/or draw more on resources from outside of finance. For more than two-thirds of organizations globally, nearly half are making greater use of external AI resources, such as technology outsourcing companies or consultants. To cultivate a finance workforce that is well-equipped for the workforce of the future, a hybrid strategy should be considered. This approach encompasses three key components:

1. Build – Invest in reskilling your workforce. By providing opportunities for continuous learning and development, organizations can empower their employees to adapt to the evolving AI landscape of finance.
2. Borrow – By integrating technology leaders from your organization into your finance function, organizations can leverage diverse expertise and perspectives. This cross-functional collaboration can lead to innovative solutions and strategic decision-making.
3. Buy – In some cases, it may be necessary to seek external talent to fill specific skill gaps. Outsourcing can provide access to specialized knowledge and experience that may not be available internally.

Beyond this, the finance function can also help the rest of the organization make informed investment decisions and effectively measure the return on investment of AI solutions. That means a chief financial officer becomes a chief value officer, establishing a monitoring framework that enables the organization to reallocate capital faster to the AI projects and solutions that have the highest return on investment for the enterprise. In other words, the finance function isn’t just using AI to automate reconciliations; they effectively become a catalyst for enterprise AI adoption.

AI isn’t a silver bullet. It needs to be combined with existing systems, people, and service delivery models in order to derive a solution that provides support to a particular outcome.

Rather than looking for a problem to solve with generative AI, focus on what you want to deliver and work backwards from there. Then define a portfolio solution that includes AI features within it, which solves for a ‘horizontal’ problem.

Most software vendors that develop tools for the finance function (such as cloud ERP) are introducing AI features within their existing platforms. Users can start to explore those features in their everyday workflows to get comfortable with AI. Start small, learn fast, then scale fast to keep up the momentum. Always connect outcomes to your people, your data, and your processes.

Showcasing enterprise value can help the finance function get onboard with generative AI and overcome fear, caution, or hesitation. It can demonstrate how their role in the organization will become even more valuable to internal customers and stakeholders, helping to drive adoption.

At KPMG, we have made significant investments in AI technology through our partnerships with leading technology providers, positioning KPMG as a trusted leader in AI deployment. We assist clients in understanding the potential of AI for their operations, supported by KPMG’s Trusted AI framework to help design, build, deploy, and use AI solutions in a responsible and ethical manner. This approach accelerates value creation and positively impacts our clients, employees, and communities.
 

Get more insights on AI in Finance in Canada and around the world, read the full global report:

Discover how we’re making a difference for Canadian organizations