With new technologies such as AI poised to transform the business landscape, a strong data and analytics foundation is a necessary part of the modern finance organization.
Evolving technology, regulatory changes, and shifts in corporate and public thinking have all motivated organizations to pursue greater data and analytics capabilities and strengthen the talent behind it all. With new technologies such as generative AI poised to transform the function, a strong data and analytics foundation is a necessary part of the modern finance organization.
Finance leaders face tremendous pressure to elevate their functions’ technology capabilities and position their organizations for the future. The continuing developments in ESG and data protection regulations have significant impacts on finance innovation, particularly as differing rules and compliance approaches emerge across global business regions. Technology transformation is another essential consideration, as organizations firm up data and analytics capabilities with an eye towards emerging technology such as artificial intelligence (AI). Anticipating and meeting the evolving data needs of an organization involves more than technology—the expectations of talent have changed, so organizations must accommodate different approaches to technology and ways of working as well as plan for necessary skill changes. While these undertakings are daunting, the potential benefits of cutting-edge data analytics and AI make them not merely worthwhile, but necessary.
Given how closely KPMG works with many of the world’s leading organizations, we have distinct insights into how finance leaders are approaching these topics. Below are three areas that corporate controllers and chief accounting officers are focused on as they navigate data, analytics, and AI transformation:
Finance leaders have embraced data and analytics in their operations, particularly in areas such as risk assessment and analysis, and reporting, yet they still trail other business areas in terms of utilizing cutting-edge applications such as AI. Before finance can move deep into AI, however, it must first prioritize enhancing data and analytics. No matter the end goal, any analytics initiative needs to start with clean data—only once data is clean can automation, AI, or significant consolidation of systems move forward.
Efforts to improve data sourcing and quality may include implementing data quality initiatives, setting up data lakes, and standardizing reporting practices. More cutting-edge approaches target data lineage to trace source data and its consumption across downstream systems. Data lineage offers increased visibility into issues like duplication in repositories, allowing organizations to clean up internal data and establish a single source of truth. Some leaders are also interested in another emerging concept: process mining, which compares the expected path of a transaction and what actually runs through the ERP system. This helps identify exceptions and variations that can then be analyzed further.
The amount of data is rarely an issue—while companies have large datasets, many are still working on prioritizing and arranging the data before they can cleanse and mine it. Establishing appropriate controls and access mechanisms is crucial to ensure information can be efficiently retrieved from systems and data lakes. As companies look forward, having high-quality data to feed into AI systems poses a significant challenge, especially in terms of real-time, continuous flow. This must be a major consideration as organizations consider integrating AI into their finance operations.
Consolidation of systems and data is a heavy lift, particularly for companies with more substantial amounts of data and disparate systems. Cleansing data and establishing standardization methodologies beforehand are necessary but can also be expensive first steps. Ensuring individual business units can independently cleanse, consolidate, and standardize their data is often a more reasonable near-term goal than targeting corporate-level consolidation, as it can enable the finance organization to more easily consolidate clean outputs manually. Regardless of scope, consolidation can help organizations unlock cost savings, eliminate redundancy, establish standardized business processes, and more efficiently use data across the enterprise.
Questions around who “owns” the data and how it is governed have also become increasingly crucial as data generated by different parts of the enterprise, or by third-party sources and customers, needs to be in one place in order to make the best use of it. How the data is governed has therefore become more important from the standpoint of ensuring that stakeholders “play well” together, and that companies meet ever-changing regulatory requirements. In response, many companies are creating Centers of Excellence (COEs), which bring together representation from different business areas and create policy, best practices, and governance processes for data use. Some COEs have grown organically around the information technology (IT) or finance areas, but depending upon their needs and structure organizations may also opt to create a centrally located department that acts as a “hub and spoke” model where data can come from various areas of the company but runs through the central COE.
Though generative AI solutions are quickly becoming commonplace, many finance organizations have yet to widely deploy generative AI tools. Adoption can be more prevalent in some shared services and other areas of organizations, and less so in finance despite considerable recognized potential for automation in transactional, corporate, and consolidation contexts. Internal “playgrounds” or sandbox environments allow for safe experimentation with AI technologies, but limited access (many companies have completely blocked access) and funding can restrain development. Per the above section, standardization across different centers is a major obstacle to effective enterprise-wide deployment of AI solutions, and incompatibilities between AI tools in different operational areas can limit the scalability of AI enablement. Specific to accounting, the practice can sometimes outpace technology due to constant change and additions to judgments and principles, casting doubt on how easily it can be automated. For example, new guidance on revenue recognition and lease accounting makes accounting processes more complicated and less easy to automate. That said, most finance leaders are encouraged by the promise of AI and are starting to move forward.
Despite being in the early stages of their AI journeys, leaders have developed numerous use cases demonstrating the value of AI. Many are starting on the customer-facing sides of their businesses with applications like chat bots in customer care centers. Other applications include helping to write rough drafts of contracts, drafting outlines for copy and charts, and putting together processes and workflow. While AI can provide a good starting point or rough draft, the new technology doesn’t remove the need for human interaction; it excels at taking on some of the lower-level, administrative tasks and calculations, but as of yet still requires review and verification. When experimenting with AI, it’s recommended to focus on low-risk, “low-hanging fruit” that can show an impact to the business (e.g., accounting policy data that generative AI can reference for easier use internally).
Several tools have been recently introduced or are forthcoming to the vendor solutions market. Microsoft 365’s CoPilot has helped many organizations automate basic tasks like ingesting email and scheduling meetings in existing software and may be useful for more complex tasks like automated creation of PowerPoint decks, RFP responses, and similar content. ChatGPT has captured the lion’s share of attention in the generative AI space, while OpenAI’s additional industry-standard solutions have enabled those without coding experience to better interface with technology using natural language, for applications like no-code SQL query generation. Anthropic’s Claude is a chat tool for deriving data insights using conversational language. No single vendor has a one-size-fits-all solution, so working with various vendors may be necessary to find tools that fit a company’s needs.
The potential of AI is somewhat offset by its risks. While the focus for many users is on AI’s benefits, cybersecurity and data privacy are major concerns that have prompted some organizations to ban the internal use of generative AI altogether. Other risks include bias in algorithms, “hallucinations” that deliver inaccurate or misleading information, and copyright infringement. Not wanting to stifle innovation, some companies are establishing clear safeguards around AI usage, particularly ChatGPT. Others are looking for ways to create their own private generative AI applications to safeguard their data privacy, intellectual property, and trade secrets from larger platforms and systems.
As enterprise systems and regulatory requirements change, organizations must develop new skill sets to match. Programming languages such as SAS, R, and Python are in high demand as organizations form dedicated teams focused on data analytics, AI, and other emerging technologies. With high demand comes scarcity, and leaders are seeing widening talent gaps around data and analytics, gaps that are only expected to grow as organizations pursue greater adoption of AI. Centralizing upskilling efforts and sharing knowledge across divisions have also been helpful for some. However, skill sets available in the current and forthcoming talent market may not align with current enterprise needs and keeping pace with these changes can be difficult from an infrastructural perspective. This dilemma often leads to organizations either sticking with outdated software or undergoing infrastructure changes when adopting new tools. Relying on established processes and on-the- job development is not ideal, but it can support organizations as they seek individuals with specific skill sets and platform experience that can fit within their existing structures.
While external assistance is available for organizations as they change systems and develop skills, internalizing the capacity to manage solutions is often preferable, especially as more data science roles are added. Hiring individuals with relevant experience from outside the organization and identifying individuals within the organization for upskilling are both key ways to meet organizational needs. New talent can provide value both in their relevant skill sets and in introducing new perspectives to the enterprise.
Adoption of new tech and development of new skills often require a significant change management focus from leaders. Some generations of workers are simply more attracted to and accepting of technology than others, which can lead to issues as organizations only grow more reliant on tech. Resistance to change, often among established workers, can hamper adoption of new tools and bring down morale. On the other hand, the opportunity to work with emerging tech can also boost morale, keep employees engaged, and even help with talent attraction. While the tech is “supposed to be the hard part” of implementation, talent is often what decides the success or failure of a new technology or process.
As leaders manage significant technology upgrades, they are also thinking carefully about how to compose the right teams. Whereas traditionally the upgrades have been the sole purview of IT, now companies are putting together cross-functional teams that have subject matter expertise in different areas. This helps to ensure that different areas of the organization, including finance and compliance, understand the data that is needed for tracking and analysis. Further, by bringing roles that were traditionally spread across multiple divisions under centralized leadership, leaders can leverage transformation to facilitate upskilling in consistent ways.
Finance executives face important decisions for their organizations regarding data, analytics, and tech talent. Consolidation and standardization of data practices enable more effective data utilization and reporting across the enterprise, while solidifying key skill sets and attaining the right talent prepare the organization for future developments. Data quality, sourcing, and change management, while burdensome, are critical foundations of the forthcoming AI revolution. Leaders must take care to establish robust controls and protections to support generative AI implementation but shouldn’t delay experimentation with AI—no company will get things perfect the first time, and those who wait too long risk falling behind.
Corporate Controller & CAO Hot Topics: Accounting for Economic Uncertainty
From SEC filings to impairment, disclosures, and even talent strategies, the current economic climate has created several challenges for financial executives.
The Intersection of Technology and Quality in the Audit
There is little doubt that technology is transforming the audit, with tech solutions giving auditors more insight than they ever thought possible.
Three Key Areas where Technology is Modernizing the Audit
Game-changing technologies are driving some of the most profound changes in the modern audit.