In the coming year we’re going to see a lot of businesses turn their minds to how they can use generative AI tools to improve their decision making and profitability.
The general public became aware of generative AI with the release of ChatGPT, Midjourney, and other tools about a year ago and we’ve been playing around with it ever since, asking it to write funny poems or mash different images together.
Even with all the experimentation, a lot of people are still a little unclear on exactly what ChatGPT is. In short, it’s generative AI in the form of a chatbot and is based on a large-language-model. It enables the user to generate ‘original’ content, such as text or programming code. It relies on a large language model, which is a deep learning algorithm that can perform a variety of language processing tasks, and so is very useful for analysing text.
Natural Language Processing has been around for several years, but the breakthrough with ChatGPT is the democratisation of AI – for the first time putting AI tools in the hands of everyone with access to the internet.
Instead of needing a background in data analytics and a custom-built language processing model, the technology is available off the shelf and can be used by anyone.
This is what has got businesses so excited, and business leaders are considering the possibilities that arise through extracting and analysing data from vast volumes of text and how they can look for trends and perform predictive searches.
Analysing text for brand sentiment
One example is sentiment analysis – analysing text in online reviews, social media, news stories and so on to track how sentiment towards a brand or product is changing. Sentiment analysis has been done for a few years, but off-the-shelf large language models allow the easy analysis of huge amounts of text at a low cost.
And sentiment analysis is only the first step. The resulting insights can be used for predictions about what’s going to happen next with a brand or product and this can help business leaders determine what they need to do to achieve a desired outcome.
Another example is to enhance risk assessment in the insurance industry. To help better predict the risk of a company being hit with a shareholder class action, an insurer can more easily assess news and announcements that might increase the risk of a class action for a policy holder.
Absorbing all the news articles and stock market announcement about the company into a large language model and identifying ‘red flag’ events can be an indication the company is in trouble. For example, a sudden departure of a key executive could indicate they’ve done something wrong, or that they know something has gone wrong and want to exit before it becomes public. This would be a risk trigger that could be flagged in real-time to the insurance provider.
AI can search through huge amounts of text for other red flags as well, such as declining sales, higher borrowing costs, environmental or safety events, cyber breaches, or litigation claims, among others.
Easy-to-use software applications
The gamechanger for all these applications is taking the advanced analytics models and putting them in a software environment where they can be used by anyone. It’s enabling non-data scientists such as executives to interact with the data in a much deeper and more meaningful way and to use it to its full potential.
This is what we’ve done with a geospatial market identification tool we built recently for a client.
The tool determines the likely market composition within certain geographies for the client’s products and helps them make better decisions about where to build outlets. It absorbs relevant demographic and market data such as age, gender, income and investments, house prices and other factors for people living in a particular area to determine what the likely demand for the product will be now and in five or ten years.
We put the analysis tool into an easy-to-use web interface, so now it’s the business executives who are using this tool and interacting with it to explore different areas around the target locations and their market attractiveness.
Putting the power of data analytics and artificial intelligence into the hands of the business leader is and important step in the adoption of analytics and we’ll see a lot more of it in 2024.