6'5. Blue Eyes. Trust Fund. Finance.

Not that kind of finance, though.

The Evolution of Finance Process Automation

The landscape of finance and accounting operations has undergone a large shift over the past three decades. What began as a wave of business process outsourcing (BPO) has evolved into a sophisticated ecosystem of automation technologies, with Robotic Process Automation (RPA) playing a pivotal role, and now, generative AI promising to redefine the very nature of process automation in the back office finance org.

The Outsourcing Wave

The finance and accounting outsourcing market has been on a steady growth trajectory, with Everest Group projecting 11-13% growth over the next three years. This growth is real:

• 51% of enterprises outsource finance functions (Deloitte)
• Increasing cost pressures, scarcity of talent, and rising labor costs are driving this trend

As we explored in our "Service as Software " post, the outsourcing model laid the foundation for process standardization and optimization. American companies learned long ago that they gained significant cost savings by outsourcing IT and business services teams in India.

The RPA Revolution: Automating the Routine

Enter Robotic Process Automation (RPA). As we discussed in our RPA post, this technology was "sold" as a game-changer, offering a way to automate repetitive, rule-based tasks without the need for complex integrations. To give a small example of scale for one use case, let's just look at invoices as an example:

• Globally, businesses process over 550 billion invoices annually (more now); in 2019, only 55B of these were paperless
• North American companies alone handle around 25 billion invoices per year, of which 75% require manual processing (costs businesses 200B a year)

The finance sector, with its myriad of routine processes (not just invoices), was ripe for RPA adoption and so naturally, the adoption (not impact) was significant, specifically in the CFO org. A leading RPA provider reports that 7,000-8,000 out of their 11,000 customers started their automation journey in the CFO's organization

However, RPA's limitations soon became apparent:

1. Data Variability: While RPA excels with structured data, it struggles with the variability inherent in financial documents. For instance, in invoice processing, companies deal with hundreds or thousands of vendor-specific formats. RPA can handle the top 20% of standardized vendors, but the long tail remains a challenge.
2. Format Inconsistency: APQC reports that companies with minimal automation receive only 27% of their financial documents electronically, compared to 91% for top performers. This disparity highlights the struggle with unstructured data.
3. Complex Data Extraction: Traditional OCR systems, often paired with RPA, show error rates as high as 20-30% for certain fields due to the complexity of financial documents. Extracting relevant information from financial documents often requires contextual understanding (e.g., "INV number" vs. "Invoice number"). Traditional Optical Character Recognition (OCR) systems show error rates as high as 20-30% for certain fields due to this complexity.
4. Multi-System Integration: Financial processes often span multiple systems. If you track down an invoice's path, it is not just SAP, it is also Salesforce, ServiceNow, and may cut across some homegrown inventory management system. Companies with poor accounts payable performance use an average of 4.6 systems for processing, compared to 3.5 for top performers.
5. Exception Handling: Industry benchmarks suggest that 20-30% of financial transactions result in exceptions requiring manual intervention, a significant limitation for RPA.

Beyond Rule-Based Automation

The limitations of RPA have set the stage for the next wave of innovation: AI-powered automation, particularly generative AI. This technology promises to address the very challenges that have constrained RPA:

1. Universal Document Understanding: Generative AI can interpret a wide variety of financial document formats and structures, much like a human would. Early adopters report automation rates of 90% or more across various financial processes, a significant leap from RPA's 50-60%.
2. Intelligent Data Extraction: AI-powered extraction systems are reporting accuracy rates exceeding 95%, a marked improvement over traditional OCR.
3. Adaptive Processing: Generative AI models can learn and adapt to new document formats and variations without explicit programming, crucial for handling the diverse range of financial documents.
4. Enhanced Exception Handling: This capability could potentially reduce the rate of exceptions requiring human intervention from 20-30% to less than 10% across various financial processes.
5. Cross-System Intelligence: Generative AI can more effectively coordinate data and actions across multiple systems, potentially bringing more companies closer to the top-performer benchmark of 3.5 systems for key financial processes.

The Convergence: BPO, RPA, and AI

What we're witnessing now is a fascinating convergence:

1. Some BPO Providers Are Embracing Technology: The very BPO providers that first adopted RPA are now trying to be at the forefront of AI implementation.
2. RPA Vendors Pivoting to AI: Major RPA vendors are rapidly incorporating AI capabilities into their platforms, blurring the lines between traditional RPA and AI-powered automation.
3. Emergence of AI-Native Solutions: New players are entering the market with AI-first approaches to financial process automation, challenging both traditional BPO providers and RPA vendors with end to end automation of processes.

I believe there will be a lot more exciting use cases not limited to:

• Financial Planning and Analysis (FP&A): AI is automating the creation of financial models, generating comprehensive reports, and providing natural language explanations of complex financial data.
• Financial Close and Reporting: The close process is being accelerated through AI-powered account reconciliations, automated regulatory report generation, and even AI-drafted management commentary.
• Audit and Compliance: AI systems are enabling continuous auditing, analyzing 100% of transactions in real-time, identifying anomalies, and ensuring compliance with evolving regulations.

What is next

As we stand at this intersection of outsourcing, RPA, and AI, several questions emerge:

1. How will traditional BPO providers evolve to stay relevant in an AI-driven world?
2. How will the role of finance professionals change as AI takes on more complex, judgment-based tasks?
3. What will the new characteristics for the infamous TikTok woman be if the ideal man is now automated away? (Just kidding!) Taking new plays on 6'5, blue eyes in the comments.

The finance function is on the cusp of another transformation. Just as outsourcing and RPA reshaped operations over the past two decades, AI promises to redefine the very nature of financial work in the coming years.

For those building tools for back office finance and accounting automation, the opportunities are immense and if you are building in the space, reach out to [email protected]