How much work do you do that requires you not to think deeply? Checking numbers, footing tables, copying and pasting information, opening and closing files, reviewing contracts, or doing Internet searches. Have you ever found a glaring conceptual miss in a perfectly accurate spreadsheet? Have you found yourself thinking about an idea and realizing that your repetitive tasks have switched to “autopilot”?
As humans, when we stop thinking or feeling we disengage. Command-driven, emotion-free machines are perfect at executing repetitive, rule-based tasks that require no judgment. Software solutions have been around for decades. What is more recent is the ability to quickly create unique series of complex computer commands that are understood by different systems, are stable, and are flexible under changing conditions. RPA is such software programs that non-IT professionals can create and run.
RPA is the least complex and most foundational subset of AI-related digital transformation. It supports adoption of other technologies, such as complex data analytics and machine-learning, ultimately leading to intelligent process automation (“IPA”). Unlike RPA, IPA can handle fuzzy data and tasks, process exceptions, problem solving and learning over time. In other words, IPA can “think”. Examples of other AI-based tools are image recognition, natural language processing, predictive analytics, and deep learning. Using these tools in your business reduces fixed costs and business risk, and improves effectiveness. In the end, you will have a constantly thinking, innovative, strategic and creative human team, rather than one that is focused on following rules, research, reviewing data or validating facts.
Did you pause when reading the last sentence? If so, RPA is likely a cultural change for you and your organization, as it is for most businesses. What do you feel about it? Do you feel anxious, excited, skeptical or indifferent? When implementing RPA solutions, it is important to create a culturally receptive environment, including the ability of humans to guide and trust machines in ways that are not apparent or possible in most legacy large-scale business software solutions. Change and human impact management is an essential part of digital transformation.
I will provide two simple examples of “attended” softbots (i.e., those that require human intervention to begin and end). Note the lack of need to train, schedule or check the work of softbots, limit their IT or data access privileges, or segregate their duties.
- Company routinely enters into many derivative contracts that require validation and valuation.
| Human | Softbot |
| · Enter into deals and save / forward trade confirmation to a folder
· Set up / document process and rules · Resolve follow-up items identified by the system · Approve system report (summary-level only) · Approve and/or book the journal entry
|
· Obtain Bloomberg strip and save to a folder
· Pull counterparty statements from the web and save to a folder · Extract structured data from PDF trade confirmations (e.g., trade ID, volume, price, dates, counterparty, etc.) · Save structured data to an Excel workbook · Save Bloomberg strip data and counterparty statements to an Excel workbook · Reconcile counterparty statements and trade confirmations · Perform all calculations (fair value, unusual items, etc.) · Produce a report with items requiring approval or follow up (based on pre-defined criteria) · Produce the journal entry |
- Company prepares external reporting.
| Human | Softbot |
| · Set up / document process and rules
· Write disclosure text · Populate and review dashboards (can be automated outside of RPA) · Resolve errors · Run program
|
· Populate all numbers in disclosure from dashboards
· Produce error warnings (based on pre-defined criteria) · Roll-forward disclosures to new periods · Execute version control · Complete formatting (based on pre-defined criteria) · Create new codes/links for new disclosure items |
The first example generally requires an RPA platform (many such platforms are now available). The second example can be entirely programmed within Microsoft Office with no ongoing costs (e.g., OTB Advisory Reporting Automation Tool). It provides an effective automation alternative at a fraction of the fees charged by large reporting platforms.
You can probably think of a very large number of similar examples in finance: account reconciliations, invoicing, matching of orders to cash, customer or supplier communications, production accounting reconciliations, cheque runs, audits of large data files, investment reconciliations, tax return preparation, or even accounting / reporting research. You are looking for a series of relatively simple steps that are rules-based, can be executed by using technology, and are frequent/repetitive enough to justify investing in automation.
Where are companies now on the RPA journey? In April 2019, Grand View Research forecast a USD$4 billion RPA market by 2025. While large, it is only a fraction of the overall market size for AI. (https://www.grandviewresearch.com/press-release/global-robotic-process-automation-rpa-market)
RPA is a mature technology that provides amazing business value. However, adoption rates are still very low and estimated by some at about 1% for large organizations. Current adoption barriers are human, not technological. RPA can be easily implemented on a proof of concept or pilot basis to automate many tasks and processes. It is a learning journey that does not require large investments upfront. This creates an opportunity to start small with solid, low-risk business cases (e.g., legacy system integration in the above examples) and move toward business model transformations as learning increases. As you start planning large-scale implementation, you should develop a strategy, team and workplans, potentially as a center of excellence.
RPA technology will make your business life better if you make it work for you, not wait until it disrupts your business or career. Why manage change when you can lead it?