The Importance of Being Predictive instead of Reactive

Disruption is not new to Singapore. Long has the country recognised the necessity to upskill its workforce and keep abreast of the latest technological developments in order to maintain relevance in a highly dynamic global economy. It has witnessed a growth of approximately 92,000 ICT professionals in 1999 to over 210,000 in 2017, with that number only set to grow as many new initiatives have been launched to train a digitally ready workforce.

Now, we once again find ourselves at the precipice with emerging technologies showing their transformative impacts on the way we work, live and play. Just recently, all non- essential businesses had to implement work-from-home arrangements swiftly as part of the nation’s circuit breaker measures. Singapore’s robust IT infrastructure and digitalisation efforts thus far helped most navigate this transition, but the COVID-19 pandemic has shown us that even the highly ready can be deterred. A possible cause is that circumstances have upended ‘business as usual’ so quickly that we lacked historical data, methods and tools to analyse this data along with the current market inputs to guide our business decisions across finance, HR, supply chain, and sales.

One of the most promising tools to address this gap is via scenario modeling. Particularly AI- based scenario modeling is taking on a whole new level of importance amid the COVID-19 pandemic, as business leaders make difficult decisions impacting their whole ecosystem including their employees, customers, suppliers, and partners. Let’s look at how this can be done practically.

Taking the first steps
As a CXO, you’ll want to look at your key stakeholder groups, identify their risks, and create models for the worst-case and the most likely outcomes of the business decisions you could make. You can then weigh up the costs and benefits to make a final decision that will land you near your desired position within a certain confidence interval.

Start by deciding the scope and issues you need to address immediately, while defining your key drivers. For example, in your first quarter, your focus might have been growth, but now it’s continuity. Next, collect, categorise and normalise your quantitative and qualitative data points so that they will form the foundation of your analysis to make your key assumptions.

Once you have your foundation in place, you can start developing the different scenarios. Consider what scenarios are most important or likely for your line of business and start there. Define what the impacts of each will be on sales, cash flow and Capex, then decide what metrics you’ll use to measure each. Finally, monitor the plan constantly and consider if you’ll need more frequent reporting to respond to changing metrics.

However, current disruption can make accurate scenario modeling a tall order. CXOs have a huge number of stakeholders to consider, and the data points needed are often scattered across different environments.

To make the task easier, business leaders need to involve fewer people in the process and limit the number of scenarios they consider. In fact, model no more than four, but be sure to spend equal time on each, even if you think certain scenarios are less likely.

Making the most of AI
Of course, scenario modelling is only one part of the solution. Unprecedented amounts of data can be a blessing and a curse without the right support. CXOs can be overwhelmed by masses of new data alongside multiple data management responsibilities. Data collection, cleansing and security can drag business leaders away from their desired speed of work.

To help carry the load, CXOs should consider what they can streamline and automate with AI. AI solutions can analyse and interpret vast quantities of data in little time, making it invaluable for scenario planning. It can also automate the many repetitive but necessary tasks associated with data management.

However, it’s worth tempering expectations and being realistic with where the technology is deployed. Companies often struggle to deploy the technology at scale and have unrealistic expectations for it. The last thing you want now is to embark on a costly and ambitious moon shot that fails to meet your objectives.

To make the most of your AI investment, you should both buy and build applications. You don’t need to build everything from scratch and doing so could create compatibility issues later on. What you need is a strategic approach that delivers interrelated solutions to maximise AI’s benefits rather than rolling out a series of disparate solutions.

Special attention should also be paid to data quality. It needs to be complete, cleansed and up to date for an AI solution to deliver accurate insights. Fortunately, AI-driven data engines can cleanse and enrich data records before they are served up for analysis.

Another important consideration is tuning. AI ‘maintenance’ is usually expensive, and manual performed by data scientists, but it’s hardly feasible when your organisation has hundreds of AI models to maintain. Applying machine learning algorithms to this process can automate this expensive task, while keeping your costs under control.

There’s no silver bullet for business disruption. However, scenario modeling, powered by AI and machine learning, can help organisations weather this unprecedented storm.