Can AI quench our appetite for automation?
Posted by Steven Kellar
Earlier this year I attended a CFO symposium where I chaired a round table on the topic:
- Forecasting with Finesse – Overcoming the limitations of dated reporting models to enable predictive, analytical support for business planning
- Appetite for Automation – Exploring the world of robotics, AI and technology systems in business process improvement
The round table focused on the way budget and reporting processes are run today and how Artificial Intelligence (AI) and predictive analytics will change the way companies run in the future. There was particular interest in how this relates to Budgeting and Forecasting and the processes of data collection and reporting.
Very few of the delegates in attendance had implemented new technologies. Those using AI and Robotics painted a positive picture of time savings and improved automation. They found that their operations staff had more time to spend on activities that added business value and gave them more job satisfaction.
AI and predictive analytics in the planning process
As discussions turned to Budgeting and Planning, it was agreed that AI and predictive analytics has the potential to align the budgeting process with external factors and continually changing business directions. As changes occur, AI-based systems can detect and advise on changes keeping systems aligned with reality and therefore improving forecasting accuracy. All those with experience in AI felt certain that it would become an important part of their planning process.
I remember in the past when statistical analysis of data was required to help with forecasting future outcomes. The data had to be manually extracted from various data sources and passed to the Statistical gurus. These data scientists would use statistical modelling tools and techniques to produce future forecasts based on historical data. The results would then be loaded into reporting tools to produce visual representations that the business could understand and interpret. Thankfully, this is a thing of the past with current planning tools like Prevero from Unit4. Today, the best performance management solutions have AI-Based predictive analytics and graphical visualisation built in.
Another area of interest is using AI in budgeting processes to help overcome or limit natural bias. It’s very easy where people are concerned to influence budgeting processes (even subconsciously) by gut feel or other motivating factors not substantiated by facts or historical data. For those interested in this topic I can recommend the following blog post.
How much bias is in your forecast? by Richard E. Reinderhoff (11 Sep 2018)
Natural Language Query
The use of Natural Language Query (NQL) interfaces has been around for some time with the adoption of Siri, Alexa, Cortana, Google Assistant etc. There have always been mixed feelings about whether these interfaces would be of use in a business context, for example by making budget outcomes more accessible to a wider group within the Company. An example could be:-
“send me the budget figures for the APAC region for Q2, for all products by month”
To satisfy such requests today, most of us use software with a selection of reporting options with on-screen menus and drop-down filters. This is not easy to do with mobile devices whereas an NQL based interface would work well.
We are seeing the emergence of Digital Assistants, such as “Wanda – a digital assistant” available with Unit4 Business World ERP, which can respond to a number of natural language requests including scenarios like: -
- Submit absence requests for sickness and vacation
- Submit travel requests for single destination trips
- Enter your daily, weekly or monthly timesheets
- Submit expense claims for one-off expenses and expenses related to single destination trips
- See HR-related info such as payslips, and vacation and overtime balances
- Raise simple purchase requests for internal items such as computer and office supplies
Challenges in relation to Staff
With the introduction of AI and the automation of tasks within the office of finance, there is always the fear of staff redundancies and a downturn in morale. The biggest area of concern in terms of roles becoming redundant is among operational staff where they are focused solely on day to day transactions. Much of the work done by these peoplecan be automated. While redundancies or redeployment would be required at the operational level, those staff that are involved in implementing, configuring and maintaining new AI based systems will learn new skills and become more valuable to the company and the employment market.
From discussions I’ve had it’s clear that finance sees the benefit of new digital technologies in delivering better value to the organization. There’s still a level of fear, however, associated with getting this right, and I think CIOs and CFOs need to work together closely. AI can be very powerful and new and better solutions are coming to market every day. They provide outputs that can be extremely accurate, far beyond what humans can do. But, they cannot replace people and our intelligence. They bring strengths but also limits that we must recognise so that we can find the most effective ways of working together.
Today, software vendors are making tools much simpler to implement and use. Users are becoming more knowledgeable and expect to be able to configure and maintain systems themselves without the need for external consultants. This simplicity and the promise for departments to become self-sufficient at relatively low cost will drive adoption of these new technologies in the finance department and we’ll see a radically different finance team infrastructure emerge over the next five years.