Artificial Intelligence For Finance 101
Anders Liu-Lindberg
Leading advisor to senior Finance and FP&A leaders on creating impact through business partnering | Interim | VP Finance | Business Finance
The story continues to develop on Instagram so if you want a look behind-the-scenes of BPI and entrepreneur life you can check it out by clicking here.
This is the channel "Trends in Finance and Accounting". 45,000+ people have subscribed! Click "Subscribe" to receive a notification and e-mail when I publish new articles on this channel every Thursday.
Much has been said about the advances in artificial intelligence or AI and how it’s going to impact businesses, professions and most relevant us in Finance. However, much more has been said than has been done and AI is still more a perception than it is reality. In fact, I’d question if we truly understand what AI is and what its uses could be for the finance function.
That’s why I thought it helpful to strip it down completely to an AI 101 for finance professionals. I’ll follow this up with a few articles around the potential of AI and how to best get started. The goal is to completely demystify what AI is and how it’s relevant for Finance.
Let’s first start with a definition
“a system's ability to correctly interpret external data, to learn from such data, and to use those learnings to achieve specific goals and tasks through flexible adaptation, Wikipedia”
Let’s break that down into a process flow to create a full align around what that means.
Still a bit abstract so let’s try and overlay this process onto something simple we know from our daily work in Finance.
- System not human could, in this case, be an invoice processing solution using optical character recognition.
- Correctly interprets external data which would be data on the invoice such as vendor name, amounts, dates, payment information, product lines etc.
- Learns from the data by looking at what the company has done historically and in principle just copy that. If the system encounters issues that it cannot solve using the historic experience, then humans can intervene and teach it how to solve the issue. If we are to fully remove the human guidance it could be learning from the data by looking up information in other databases such as a contract archive to see if the information on the invoice matches the contractually agreed terms.
- Makes flexible adaptations to for instance make the information on the invoice ready to be recorded in the company’s systems.
- Achieves specific goals or tasks which would be the processing of the invoice to have it registered in the appropriate systems including a journal entry in the accounting system and ultimately paying the invoice.
From receipt to paying your vendor invoices are not processed without human intervention by a system. Many companies are already using solutions like this today so we’re not talking science fiction here. This is just one of the tasks of the finance function that AI can solve for us now. Here are some more ideas to where you could use AI.
- Your billing could be done in a similar fashion using many of the same processes albeit in somewhat reverse order.
- Your forecasting could be fully automated taking input from external databases and internal company information.
- Your intercompany transactions could be fully reconciled without you having to touch them.
The list goes on your mindset should be to look at every single process in your finance function and ask, “could AI solve this process too?”.
AI demystified
I’ll get back to how you could best get started using AI later. What’s important for now is that you understand the basics of what AI is and especially in relation to finance and accounting. This will help you build a solid foundation to act rather than being action-paralyzed from the endless possibilities that AI offers.
Understanding how AI can solve for a single process and the mechanics of each step in the solution should demystify what AI really is. If you still think something is unclear, then state your question in the comments and I’ll be sure to answer them (even if I can’t myself, I’ll find some experts that can).
Next, why not share examples of where you are already using AI? I published an article not long ago with just such an example. You can read it here. What are the next steps you’re planning in AI (and if no steps why not)? It’s time we start fully utilizing the opportunities of artificial intelligence to power the digital transformation in Finance. Are you ready to run with them?
Continue reading below for more articles about how digital is impacting Finance.
The Alternative Excel Methods To VLOOKUP
Why The Digital Revolution Hasn’t Caught Onto Finance Yet
Tech vs. People. Where Should Finance Invest?
A Digital Reality Check Of The Finance Function
How To Make Robots A Part Of The Finance Family?
Why You Should Only Robotize Standard Processes
Robots and Humans. A Marriage Made In Heaven Or Hell?
A Tale Of Robots: From Assembly Lines To Knowledge Workers
Robots Must Solve Business Pains To Be Successful
What AI Competencies Do Your Finance Team Really Need?
Here's How To Test If Your AI Solution Will Be A Success
You're The User Of AI. Yes You, So Take Charge!
Blip. Blop. Accounting Robot. Are You Ready?
Are You Ready For Robotics Process Automation?
Have You Met Your Robot Accountant Yet?
Robots Are The Future Of Analytics
Your Robot Accountant Has A Name, It's Dixie
Anders Liu-Lindberg is the co-founder, COO (Chief Operating Officer), and CMO (Chief Marketing Officer) at the Business Partnering Institute and owner of the largest group dedicated to Finance Business Partnering on LinkedIn with more 8,500 members. I have ten years of experience as a business partner at the global transport and logistics company Maersk. I am the co-author of the book “Create Value as a Finance Business Partner” and a long-time Finance Blogger with 40.000+ followers.
Built to Optimize: Business performance, Strategy, Data platform, and Decision support
4yAnders I participated in a webinar today with the FP&A Trends Group hosted by Larysa Melnychuk, Tina Vermeyen and Stéphanie DIDIER - LE RETIF. They have done a study on AI which has promising results.
This is a fantastic primer on a real world use case of what’s possible. Thanks so much for sharing and teaching!
Sr Manager na Accenture Brasil | Gestão de Projetos Globais de F&A
4yHi Anders Liu-Lindberg Appreciate the content shared and agree with the biggest part. There is only a point I would raise here are the mistaken interpretations about AI and for that I would use an example from Invoice Processing process. Most OCR solutions are sold as machine learning systems, but they are simply templates that collect most of the information in a PDF or other file and there is no learning from the system with the historical data. It is very important that, based on the definition of artificial intelligence, systems can cross data and learn from it. Here is where I think most companies are deceived buying a “Bike for the price of a Tesla” (expression we normally have in Brazil). Quoting your last test about AI: “It’s no longer enough simply knowing how to translate the outcomes of the model to the business stakeholders. No, FP&A professionals must become multilinguistic translators that can speak both business and data.” In my honest opinion finance cannot live from now on without data analytics, machine learning and AI not only in invoice processing, but Forecasting Scenarios, preventing errors that might appear in Closures and all the other activities related. It is a must. Sometimes I wonder how many AI-based systems have predicted a scenario like the one we have now for COVID-19, and even though I imagine that few have, it comes to mind those who still don't use AI have predicted even less.
Vice President | Delivery Lead | Technology | JIRA | PMO | Project Planning & Controls
4yHi Anders Most of my recent projects include OCR, machine learning, and optimizing AI capabilities within the Finance & Accounting functions. So your article truly hits home. I’d to add a note of caution regarding Optical Character Recognition, or “OCR.” Example: If vendor invoices contain logos, graphics and acronyms the OCR technology won’t match that vendor invoice to an existing profile within NetSuite or SAP ERP, for example. This is because OCR is unable to return a match to the vendor profile because the information used to create the profile is a W9 (alpha numeric characters). Therefore, when considering OCR it’s key to ask during the discovery period about the rate of fail when vendor invoices contain logos, acronyms or DBAs. The task to make the edits to true up the invoice to the right vendor profile will fall on the end user. Great topic, thank you!
INDEPENDENT CONTRACTOR
4ylove this..Thanks Anders