Digital Financial Reporting

The document; Closing the Skills Gap: Specifics as to what professional accountants need to learn to survive and thrive in the digital age of accounting, reporting, and auditing; is my best attempt at summarizing information which I wish I would have known 20 years ago when I started fiddling around with what became XBRL.

I tried to provide the least amount of what I would consider to be a complete set of basic information a professional accountant would need to correctly understand "the artificial intelligence revolution".  Here is the executive summary:

• Changes that will be brought about by what is being called "the artificial intelligence revolution" among other terms will unquestionably be significant.  It is not useful to overstate or understate the impact.
• A gap exists between what professional accountants understand today and what they need to understand to survive and thrive in the digital age of accounting, reporting, and auditing.
• This gap in understanding capabilities of technologies such as artificial intelligence makes it difficult for professional accountants to grasp the changes that are not only inevitable but are imminent.
• Undergraduate and graduate college curriculums need to be updated for these changes.  Continuing professional education offered likewise needs to be updated.
• The "learn to code" hysteria is not the right answer.  Having professional accountants become more proficient with information technology or computer skills is likewise not the right answer for most professional accountants.
• Computers work per the rules of mathematics.  Mathematics works per the rules of logic.  The primary gap that needs to be filled is for professional accountants to gain a sound understanding of how computers actually work and the skills necessary to leverage these very useful tools of the digital age of accounting, reporting, and auditing appropriately. Understanding the basics of logic is the key to prospering in the digital age.
• Business logic can be used by professional accountants and other business professionals who have little or no formal training in logic.  A sound understanding of business logic enables effective communications with information technology professionals.
• The artificial intelligence revolution will not only impact accounting professionals, rather all business professionals will be impacted.  Closing this knowledge gap sooner rather than later can help position professional accountants as leaders for the digital age of accounting, reporting, and auditing.

Essentially, what professional accountants and other business professionals need to learn to do is to "think critically" about data and information. What I mean by that is that they need to be able to be consciously conscious when working with data and information.  They need to be rigorous, deliberate, and methodical.

Any comments or other feedback would be gladly accepted.

This ZIP archive contains a little Microsoft Access 2010 Database application (prototype) that I created.  If you don't have Microsoft Access 2010 or greater, you can download a runtime version of Access here.

This video shows you what the application does.  Basically, the application provides a number of best practice examples of US GAAP disclosures.  I have provided 33 disclosures.

What the application does is let you look at best practice disclosure examples like these for public companies that have provided that disclosure, like these.

Why this is interesting is that you can easily extract specific examples from XBRL-based financial reports of public companies that file with the SEC.  I am showing 33 disclosures, but you can do the same thing for EVERY disclosure.

I have a much, much better application that I created, but it uses XBRL Cloud's web service.  But to use that, you need a username and a password which you have to get from XBRL Cloud.  I removed all the functionality that I had used from XBRL Cloud and therefore all the features provided by XBRL Cloud, such as the ability to click on all elements and get Accounting Standards Codification references, concept definitions, etc.

What I want to do is create a more complete version of this application and make it available to intermediate accounting students.  If anyone is interested in helping out with this little project, let me know.

Brad Smith, CEO Intuit, is warning accountants that they must embrace artificial intelligence (AI) in their professions or risk falling behind.

Blackline refers to itself as a finance controls and automation platform and provides accounting process automation says that accountants should gear up for AI disruption. Blackline says that artificial intelligence is set ot revolutionize accounting and change the role of accountants.

Xero sponsored a survey which says that 80% of accountants understand that they’ll need to undertake training in the next five years to adapt to technology and that AI is the new frontier for accounting technology.

The Journal of Accountancy is talking about rethinking the audit and the new type of auditor.

Steve Guggenheimer's Blog post, Introducing Dynamics 365 AI Solutions, explains how Microsoft is "helping developers build services that amplify human ingenuity."

SAP is employing artificial intelligence, as is ORACLE; basically I could go on and on.

The risk of not getting on board with the changes artificial intelligence and other technologies are enabling is increasing.  Might be ready to get ready for the digital age of accounting, reporting, and auditing.

Here are some other articles:

• Forbes: Robotics: The Next Frontier For Automation In Finance And Risk Management
• Deloitte: Its Digital Crunch Time for Finance
• PWC: Hi, Robot
• The Economist: Why the future of finance is (still) automation
• Kofax Advisor: Is the Modern CFO Headed for Extension?
• Ventana Research: Welcome to the Age of Robotic Finance

I took the metadata and tools that I createdfor checking the fundamental accounting concept relations continuity cross checks of US GAAP based financial reports and I started to update that for IFRS.  If anyone wants to check this out, here you go:

1. Excel extraction prototype updated with some IFRS metadata
2. Several IFRS example financial reports (used to exercise the Excel spreadsheet application)
3. Easy to read version of the IFRS FULL taxonomy
4. Easy to read version of the IFRS for SMEs taxonomy

Ultimately I will update my XBRL-based metadata.  I will probably do that based on 20-F filings to the SEC that contain IFRS based reports that use the IFRS taxonomy.  There really are not very many IFRS reports around in XBRL to test things with yet.

If you like Sudoku puzzles, you will enjoy fiddling around with this.  Someone may even be able to create a useful product.

Today, what portion of the checks or other review steps performed to make sure the quality of a financial report is what it needs to be are performed using automated processes? The answer is ZERO.  Well...I guess you could count using spell check in Word an automated task.

Why is that?  Why can't certain quality control tasks and other review steps be performed using automated machine-based processes.

The reason all review tasks have to be performed using human-based processes is because financial reports of the past have been unstructured documents which are impossible for a machine to read reliably.  Well sure, a word processor can "read" a financial report; but word processors know nothing about financial reports.  Word processors understand documents and they can tell you if you messed up the document structure (i.e. word, paragraph, section, etc.).  If this is unclear to you, consider watching the video How XBRL Works.

XBRL-based financial reports created an opportunity to change the review process used to make sure financial reports are created correctly.  XBRL alone cannot change this.  But, if you put three things together a transformation can occur.  To achieve this change you need three things:

1. Structured financial reports: Individual pieces of information in structured reports can be reliably addressed and worked with by computer-based processes.
2. Machine-readable rules:  The logical, mechanical, structural, mathematical, accounting, and reporting rules of a financial report must be put into a form that a computer can understand and work with.  The machine-readable rules are used to make sure the structured financial reports are created correctly.
3. Software:  Computer software is the machine that reads the rules and otherwise leverages the rules to (a) help professional accountants create the financial reports and (b) help make sure the reports are created correctly.

Because of XBRL-based financial reports, machine-readable rules, and software that can effectively understands the rules; it is no longer the case that ZERO percent of the process of reviewing a financial report to make sure it has been created correctly.

Today, it is possible to automate some percentage of those quality checks and other review tasks.  What percent?  While the actual percentage of what can be reviewed using automated machine-based processes is unclear; that percentage will be significant.  The percentage will be determined by the volume and quality of the machine-readable rules and the problem solving logic available from the software.

When will this software be available?  Well, the software is available today.  Here are two software applications that perform automated review tasks:

• XBRL Cloud Reporting Checklist (commercially available, in BETA)
• Pesseract Reporting Checklist (working proof of concept to test the feasibility of automated reporting checklists)

You can fiddle with the reporting checklist here if you are interested; experience the details for yourself. This ZIP archive contains an Excel based application that lets you extract and see the machine-readable rules used by both of those software applications. There are THREE Excel spreadsheets: one for US GAAP, another for IFRS, and another for what I call the XASB reporting scheme which is an imaginary scheme I created to figure some things out.  This document explains the details. This document explains the techniques used.

Currently, XBRL Cloud and Pesseract are both flawed in my view.  The flaw is that they are both batch processes run AFTER a report is created.  Why is that a flaw?  It is a flaw because the software does not help you actually create the report.

What I am doing is creating software that is used DURING the process of creating a financial report.  Basically, what I refer to as an expert system for creating financial reports.  The idea of expert systems is not new.  Expert systems is one of the most successful areas of artificial intelligence.  When I talk about human and machine collaboration or machines augmenting humans; what I am talking about is an expert system for creating financial reports.

Ten years ago I could not explain the difference between "syntax" and "semantics".  Literally. When I created that video, How XBRL Works, on May 1, 2008; I was still struggling with understanding the difference between those two terms.

A lot has changed since then.  I have accumulated a lot of information and step-by-step figured out how to create the expert system I believed could be created.  I figured it out with the help of a bunch of other people.  There are still a few details but the problem of how to do this has been solved.  That is the purpose of Pesseract, to figure this out.

Financial reporting...get ready to be disrupted.