ChatGPT and scenario planning

How companies could use ChatGPT to create, model, and evaluate scenarios?

Scenario planning in finance is a strategic process that involves analyzing different plausible future scenarios and their potential impact on financial outcomes. It includes identifying key uncertainties, analyzing data, evaluating implications, and developing action plans. By considering multiple scenarios, decision-makers can make more informed and resilient financial strategies.

Scenario planning becomes crucial today. The economy and enterprises must embrace mighty crises such as the C-19 pandemic or war in Ukraine. These crises are called black swans: events of a major and widespread (global) impact. Even though, before they happen, they are commonly assessed as being not probable.

Companies must adapt quickly to the changes caused by the black swans. If they don’t, they may disappear from the market. Just like that. If they do, they may gain a strategic advantage and achieve enormous success.

In scenario planning, factors to consider include macroenvironmental variables, industry-specific factors, geopolitical events, social and demographic trends, technological advancements, risks, opportunities, and stakeholder perspectives. By incorporating these factors, organizations can develop well-crafted scenarios that assess the impact on financial outcomes and support decision-making.

Anyway, that’s a lot of stuff to remember, let alone to do that…

Scenarios must assess both financial and non-financial measures. Financial measures include revenues, expenses, profits, cash flow, and investments. Non-financial measures encompass customer metrics, operational metrics, employee metrics, market dynamics, and reputation/brand.

Considering both types of measures provides a comprehensive evaluation of the potential impact across various aspects of the organization.

Tu put it straightforward: scenario planning is a complex, multidimensional exercise. And there is usually very little time to design them. The question is if we could already leverage technology advancements, such as ChatGPT in that process. Let’s look at the below case study. Case study text was generated using GPT-4 (only small corrections from my side).

Case Study: “GlobalTech Enterprises: Navigating Financial Difficulty and Supply Chain Disruptions”

Company Overview

GlobalTech Enterprises, a leading multinational corporation in the electronics industry, designs, manufactures and sells an array of digital devices, including laptops, tablets, and smartphones. The company, founded in 2005, has expanded to have a presence in over 40 countries, commanding a strong market position and customer loyalty for most of its operational years.

Challenge

In 2023, GlobalTech faced unprecedented financial difficulties driven by a series of events that disrupted its supply chain. It caused significant production delays. A surge in global geopolitical tensions, coupled with the aftereffects of the C-19 pandemic, resulted in logistical nightmares, impeding the smooth movement of raw materials and finished products.

This supply chain disruption had immediate consequences for the company. Firstly, it led to an inability to meet product demand, resulting in substantial revenue loss. This, along with additional costs incurred to attempt to manage the crisis, drove profits to an all-time low.

Production delays further led to declining customer satisfaction. A backlog of unfulfilled orders and longer wait times created a ripple effect of disgruntled customers, damaging the brand’s reputation and straining customer relationships.

The financial health of GlobalTech began to deteriorate rapidly. The lack of cash flow and mounting debt meant the company was unable to invest in R&D, innovation, or essential infrastructure improvements, leading to a lack of competitiveness in the fast-paced electronics market.

Prompt to ChatGPT (GPT-4)

What are possible remedy scenarios for GlobalTech Enterprises? Simulate potential results: from negative to positive, look at profit and cash flow.

ChatGPT response:

Scenario 1: Bleak Outlook

In this scenario, GlobalTech’s attempts to diversify the supply chain, reduce costs and improve customer retention are ineffective. The company’s new supply chain partners fail to deliver due to continued global unrest and further unforeseen issues. Cost reduction initiatives result in the layoff of key personnel, leading to poor morale and reduced productivity. Customer communication and retention efforts don’t sufficiently restore faith, and the company continues to lose market share.

Financially, these factors would result in declining revenue and increasing operational costs. With an ever-growing backlog of orders, customer returns and refunds could further drain the already limited resources. A steady decline in profits would continue, and cash flow would remain negative.

Scenario 2: Slow Recovery

In this scenario, the company manages to make some headway with its strategy. The new supply chain partners can fulfill orders, but not as efficiently as hoped. The cost reduction measures lead to some improvements in operational efficiencies, but not without initial hiccups. Customer retention efforts prevent further customer loss but fail to attract new ones.

As a result, the company sees a slow uptick in revenue but not enough to offset all losses. Profits remain low, but they show signs of improvement quarter over quarter. The cash flow situation improves slightly but remains tight, limiting investment potential.

Scenario 3: Successful Turnaround

In this scenario, GlobalTech’s strategy proves highly effective. Supply chain diversification effectively mitigates risks, ensuring a constant flow of materials. Cost reduction and efficiency measures significantly improve operational performance. The customer retention strategies not only help regain the lost customers but also attract new ones, restoring the company’s market share.

This scenario would result in a healthy increase in revenue, marking a return to profitability. As a result, there is a significant improvement in cash flow, which the company can start reinvesting into R&D, product innovation, and infrastructure, helping regain a competitive edge in the market.

Scenario 4: Unprecedented Growth

In this most optimistic scenario, GlobalTech’s strategy exceeds all expectations. The new supply chain partners prove to be even more reliable and cost-effective than previous ones. The cost-cutting measures, in addition to improving operational efficiency, also stimulate innovation. The customer retention efforts not only restore faith among existing customers but lead to a surge in new customers.

Financially, this translates to a rapid increase in revenue and profits. The cash flow situation significantly improves, allowing for aggressive investment in R&D and infrastructural improvements. The company’s market share grows beyond pre-crisis levels, positioning GlobalTech as a more resilient and customer-focused organization than ever before.

As one can see, ChatGPT produced 4 different scenarios: from negative to positive one. It also provided some descriptive characteristics, that can now be converted to more solid, numeric assumptions.

Let’s try to compare these scenarios in a less descriptive form.

ChatGPT (GPT-4) prompt

Can you provide a simulation of the results of these scenarios in tables? What are the threats and opportunities associated with each scenario?

ChatGPT response:

Certainly, below are the tables simulating the financial results under each scenario, as well as the associated threats and opportunities.

Each scenario presents a different trajectory for GlobalTech, with its own set of potential threats and opportunities. The key is for the company to adapt its strategy according to the unfolding circumstances, mitigating threats and capitalizing on opportunities.

Let’s ask ChatGPT to write a model in Python, where we could test these four scenarios. It took me three promptings rounds to get a result that I liked.

Prompt 1

Can you write a code for a model, where I could put some base numbers and then different assumptions for selected ratios?

Prompt 2

Can you alter the model so that it presents results over the next 3 years?

Prompt 3

Can you modify the code so that it produces results with no decimal places?

Below is the generated code.

class GlobalTechModel:
    def __init__(self, base_revenue, base_profit, base_cash_flow):
        self.base_revenue = base_revenue
        self.base_profit = base_profit
        self.base_cash_flow = base_cash_flow

    def bleak_outlook(self):
        results = []
        for i in range(1, 4):
            revenue = round(self.base_revenue * (1 - 0.30 * i))  
            profit = round(self.base_profit * (1 - 0.50 * i))  
            cash_flow = round(self.base_cash_flow * (1 - 0.50 * i))  
            results.append((revenue, profit, cash_flow))
        return results

    def slow_recovery(self):
        results = []
        for i in range(1, 4):
            revenue = round(self.base_revenue * (1 - 0.10 * i))  
            profit = round(self.base_profit * (1 - 0.15 * i))  
            cash_flow = round(self.base_cash_flow * (1 - 0.15 * i))  
            results.append((revenue, profit, cash_flow))
        return results

    def successful_turnaround(self):
        results = []
        for i in range(1, 4):
            revenue = round(self.base_revenue * (1 + 0.20 * i))  
            profit = round(self.base_profit * (1 + 0.30 * i))  
            cash_flow = round(self.base_cash_flow * (1 + 0.40 * i))  
            results.append((revenue, profit, cash_flow))
        return results

    def unprecedented_growth(self):
        results = []
        for i in range(1, 4):
            revenue = round(self.base_revenue * (1 + 0.50 * i))  
            profit = round(self.base_profit * (1 + 1.00 * i))  
            cash_flow = round(self.base_cash_flow * (1 + 1.00 * i))  
            results.append((revenue, profit, cash_flow))
        return results

# Create an instance of the GlobalTechModel with some base numbers
model = GlobalTechModel(base_revenue=1000000, base_profit=100000, base_cash_flow=50000)

# Use the model to simulate each scenario
print("Bleak Outlook:", model.bleak_outlook())
print("Slow Recovery:", model.slow_recovery())
print("Successful Turnaround:", model.successful_turnaround())
print("Unprecedented Growth:", model.unprecedented_growth())

Now let’s run the code in Google Collab. That’s the result we get: estimations of revenue, profit, and cash flow for three consecutive years. One quick code update before I show you the result and…

We can now test various assumptions to see an optimal set of results. The model requires further work, yet you could see how quickly it works, and how quick was whole conceptualization and deployment of such a model was.

Scenario planning is crucial during unexpected crises, or “black swan” events. These situations demand quick adaptation and informed decision-making.

ChatGPT can help businesses navigate such scenarios, as demonstrated in the case of GlobalTech Enterprises. In response to a prompt, ChatGPT generated four potential financial outcomes, providing a detailed and strategic approach to scenario planning. Further, it was able to craft a Python model to simulate these scenarios over three years, illustrating its potential as a dynamic tool for financial forecasting during crises.

Did you like that article? If so, consider subscribing to my site on medium.com.

Links to my other media: 1. My blog (in Polish): www.michalszudejko.pl

2. My LinkedIn: https://www.linkedin.com/in/michal-szudejko/

3. My YouTube (in Polish): https://www.youtube.com/@YTmichalszudejko