Can Intelligent Planning really be Artificial?

By Craig Marsland
3 mins
Last updated: 10 September 2024

What is Artificial Intelligence?

Artificial intelligence (AI) is a broad term for the ability of a computer to perform various analysis and decision processes associated with the human brain. Machine learning (ML) is a specific type of AI in which a computer analyses data, builds a model to make predictions (or decisions) from this data, evaluates the output, and then learns from its results and tunes the model for optimal predictions.

When I was working for one of the big 4 in the 90’s I visited a client who wanted some benchmarking work done. I asked the client what they were looking to benchmark and what insight they were hoping to gain from it. I was surprised to learn that they were unsure. This scenario is just one example of a recurring trend we have observed over the years as Management Consultants. Encountering clients who latch onto buzzwords without truly comprehending their meaning or the value they can provide.

So, can AI really help improve the accuracy of planning and forecasting, or is it just the latest hype with little or no understanding of what it actually is and how it can be used effectively?

Will AI increase forecast accuracy?

AI and Intelligent Forecasting in Housebuilding and Construction

After the 2008 crash, I joined a FTSE100 housebuilder. The housebuilding industry operates in a cyclical market and is highly influenced by external factors, such as inflation and interest rates.

The company had been badly burned during the crash due to decreased affordability for new houses – despite there still being a huge demand. This caused land values to drop, resulting in write-offs from the balance sheet.

To prevent such incidents in the future, we developed a scoring system to try to anticipate market trends. The concept was simple: acquire cheaper land when the market nears its bottom and allocate resources to other areas when the market approaches its peak, such as distributing dividends to shareholders.

Imagine the possibilities if we had access to an AI tool capable of accurately predicting market conditions based on mortgage data, interest rates, inflation, unemployment, and other macroeconomic factors. By leveraging this technology, we could make even more informed decisions and potentially mitigate risks associated with the volatility of the industry.

So, will AI truly increase forecast accuracy? While there is undeniable potential for machine learning to continually refine its predictions with improved precision, it’s important to consider the limitations of forecasting accuracy. External factors such as weather conditions, political events, and human errors always have a significant impact on the outcome.

When it comes to standard FP&A forecasting, it may be challenging to see the immediate benefits of AI. However, AI proves its value in predicting trends and assessing the influence of external factors in other use cases, such as supply and demand planning.

It is also important to consider the nature of the business before utilising AI technology. For companies with frequent promotions or unpredictable demand, AI can be particularly valuable. Whereas, companies that have a steady or predictable demand, or market makers like Apple, who generate their own demand, may not benefit from AI in the same way.

While forecasting accuracy can never be perfect, leveraging AI in the right context can enhance predictions and provide valuable insights.

AI or Advanced Analytics?

In order to solve complex linear optimisation problems, Anaplan offers the optimiser add-on. This add-on utilises the industry-leading Gurobi optimisation engine. It can be used in various scenarios such as manufacturing planning, workforce optimisation, and even call centre operations, where a desired outcome must be achieved while factoring in complex constraints and variables.

For businesses with more advanced needs, Anaplan’s Plan IQ machine learning tool is available. Plan IQ provides a full suite of advanced AI/ML forecasting models, ensuring access to the best-of-breed AI/ML forecasting methods for better results.

These advanced tools and features solidify Anaplan’s position as a versatile and capable platform, addressing a wide range of needs and complexities for businesses seeking success.

PlanIQ uses machine learning to generate forecasts for your business planning process. Machine learning (ML) improves the accuracy of forecasts. Combined with the use of algorithm training, ML helps PlanIQ make intelligent forecasts, which are based on historical data and any additional datasets. Teams across the supply chain, finance, sales, and human resources can use PlanIQ to make confident decisions and be better prepared for the future. 

Anaplan, 2023.

In certain instances, PlanIQ has the potential to significantly enhance the planning process by incorporating previously overlooked factors and conducting reanalysis of data, to make better decisions. However, it’s important to acknowledge that integrating PlanIQ into Anaplan models and existing apps introduces another level of complexity and administration. In many cases, it may just be another forecasting method to include alongside the other 19 that are already included in the statistical forecasting app.

Data quality plays a pivotal role in all planning activity and is critical in machine learning. If an incorrect dataset is fed into a machine learning model, then the insights derived from the data analysis will be flawed. The accuracy and effectiveness of any ML model will only ever be as good as the quality of the data it is based on. Therefore, it’s essential to ensure that the data input is reliable and precise.

Balancing AI Advancements with human judgement.

Balancing AI Advancements with Human Judgement

Humans are fallible, but I’d still rather be operated on by a surgeon than a computer-controlled robot that could sever an artery if it loses the Wi-Fi connection.

People inherently buy from people and despite our failings, decisions will always require human intervention. A number of factors will make senior executives wary of using AI for decision-making. Including lack of understanding, accountability, available AI skills, perceived lack of control, and the potential poor decisions resulting from poor-quality data. Humans will ultimately decide what data is fed into the AI tool and how it will be used to make predictions.

AI may provide some useful analysis for planners and decision-makers, but it doesn’t replace the need for collaboration and communication–connected planning.

Conclusion

Integrating artificial intelligence into planning processes shows immense promise, but choosing the right tool is vital. While no planning method can achieve perfection due to unforeseen influences and data quality (which will always be a critical factor) AI’s ability to analyse years of historical data and patterns at speed undoubtedly adds value to organisation’s forecasting strategies. However, a cautious evaluation is necessary to make the most of its capabilities.

To find out more about PlanIQ or Advanced Analytics within Anaplan, why not get in touch?