AI Experiments: Top or Bottom line?

In the last two articles, we looked at the CAIO role and how to build AI experiments. This is the third and final article in the series. If you have not read the others first, you might want to start there – especially the article on running structured experiments.

Now that we’ve looked at how to run and structure AI experiments, let’s add a bit more structure. What part of our business are we looking to impact? Are we looking at our top line? Our bottom line? Cost efficiencies? Let’s break that down a little.

Business Impact by AI - Image by Dall.E

Yes, AI is a revolutionary technology. It has the potential to upset industries and make many of today’s organizations obsolete. As is often the case with new technology, those who best make strategic use of it have the best chance of winning in the market. On the other hand, one needs to carefully balance risk with rewards.

Despite its potential, most organizations that I see experiment with AI follow the same approach:

1. Look at increasing efficiency, reducing costs, thereby growing the bottom line;

2. Using standard off-the-shelf SaaS tools, existing API’s (OpenAI, Anthropic, Google) or integrating existing Large Language Models (LLMs).

It’s a very low-risk approach. What could really go wrong? On the other hand, the rewards are also low. The best case scenario is a 5-10% increase in bottom line, whilst creating vendor lock-in, and creating no long-term value.

Top or Bottom line?

Before we start, a quick recap on top vs. bottom line. Top line refers to an increase in a company’s gross revenue or sales. It is called the “top” line because it appears at the top of the financial statement. Enhancing top-line growth typically involves increasing sales volumes, introducing new products, or expanding into new markets. Bottom line on the other hand refers to an increase in net income, which is the profit a company retains after expenses, taxes, and costs have been subtracted from total revenue. It is called the “bottom” line because it appears at the bottom of the financial statement. Bottom-line growth is achieved by reducing costs, optimizing processes, or improving overall efficiency, impacting what the company ultimately earns in profit.

What are the opportunities we could create using AI?

1. Efficiency and Cost Reduction (Bottom line Growth):

   • Approach: Implement AI to automate repetitive tasks, augment staff to optimize efficiency, optimize supply chains, or improve customer service.

   • Examples: Using AI for predictive maintenance in manufacturing can significantly reduce downtime and repair costs. Deploying Microsoft’s Copilot or Google’s Gemini to augment staff. Klarna’s approach to replace its customer service agents with AI chatbots is another good example.

   • Reward/Risk: Low risk, low reward.

2. Growing Business (Top line Growth):

   • Approach: We can increase our upside by using AI to enhance our current offerings, improve our marketing, help us find new customers, to upsell to existing clients, and reduce churn.

   • Example: Implementing AI-driven recommendations in an e-commerce platform can increase average order value and customer retention. There are also numerous AI marketing tools in the market to help us find new customers.

   • Reward/Risk: Medium reward, medium risk. It leverages your existing customer base and product line for incremental revenue growth.

3. Introducing New Business Models (Top line Growth):

   • Approach: Innovate using AI to create entirely new products or services, or by entering new markets.

   • Example: Imagine a healthcare provider using AI to develop a telemedicine platform that offers personalized treatment plans based on patient data and AI analysis. Or, consider how Tesla is using its existing customer base to collect data to build a fully autonomous driving AI.

   • Reward/Risk: High reward, high risk. This strategy can redefine your business but requires significant investment and carries very high uncertainty.

Photo by Yeshi Kangrang on Unsplash

With high risk also comes (potentially) a high reward. In all cases, for us to create lasting value, we need to recognize the value of our data, and of our own models. Each of these create a moat making it harder for future competitors to conquer our market position.

Let’s look at an example, a fintech company with access to vast amounts of transactional data. By analyzing this data, they could develop AI algorithms to predict customer behavior, identify new product opportunities, or even create dynamic pricing models that maximize revenue. They could use existing AI models, or tools such as Prophet to build those predictions. They could also build their own custom model.

For a competitor to do the same, they’d need the same amount of data, AND they would need to build a custom model of their own. This last example also highlights why it’s key to start now. It’s highly likely that at least one of your competitors is already working on tools that will upset your market. You do NOT want to be caught by surprise when that happens.

What’s Next?

Is AI too advanced for your business? Not sure you can benefit? Current solutions and applications have made it simple for any business to start using AI, even if at a small scale, optimizing your bottom line.

However… you can assume someone (startup or corporate) will enter your market soon using highly advanced AI models. In order to prepare now, start analyzing your current market. What is happening across the globe? What tools have others already made available? How could others use their data and custom AI models to upset your market? Then – how will you prepare for that situation?

Considering how AI will impact your market and how to prepare for it – even if you don’t start building your own models – is a strategic action you should undertake today. If you need support, reach out for a strategy session, and let’s chart your course to innovation and success together!