Generative AI and the path to predictive analytics

It appears all but certain that generative AI, or one of its leading products, such as ChatGPT, will become the technological buzzword of the year for 2023. The rapid development and rollout of these advanced artificial intelligence programs have been both astonishing and worrisome for those fearing the dangers of growth that outpaces regulation. While it’s impossible to predict where generative AI will lead us, it already appears to be driving significant change in the realm of analytics. 

At an enterprise level, generative AI possesses the potential to counter significant bottlenecks in what organizations and teams alike can accomplish, even when facing stringent deadlines. 

Artificial intelligence is also, theoretically at least, free of the biases and cognitive difficulties that humans can experience in forming and testing ideas at scale. This notion, however, has been contested due to human bias that could influence the datasets that AI uses.  

Away from this, there’s little contesting the time- and resource-saving qualities of generative AI and the insights that it’s capable of producing. While a major drawback of big data is that humans simply cannot interpret thousands of pages of information at a rapid pace, AI can not only ingest it in an instant but interpret key points and metrics to deliver immersive data insights for users to consume. 

Generative AI’s potential is such that Goldman Sachs estimates that the technology could deliver a 7% boost to global GDP over the course of the next ten years while also lifting productivity growth by 1.5 percentage points.

>>Don’t miss our special issue: The Future of the data center: Handling greater and greater demands.<<

For business leaders, generative AI and predictive analytics are set to become a partnership that’s impossible to ignore. With many firms already actively undergoing digital transformation, the incorporation of artificial intelligence represents a major step towards keeping heads and shoulders above the mire of a hyper-competitive landscape.

The path to predictive analytics

For businesses seeking to optimize their inventory throughout the year, generative AI is an essential component in powering projections concerning vital customer data. This helps to better budget stock and work more efficiently with supply chains. 

As the technology matures, businesses will be able to use the technology to analyze large datasets and spot trends that they can use to predict future customer demand or changing consumer preferences. 

One of the strongest examples of generative AI leveraging predictive analytics today can be found in the events industry. Software firms like Grip and Superlinked have created services that use predictive AI to help event organizers make data-driven decisions about the different aspects of events. 

Here, these firms have used generative AI in analyzing attendee data from past events to gain insights for future events. 

We can liken this process to Google Trends, which can use search data to show when certain terms are being queried more frequently. Generative AI models can take similar indicators of audience sentiment, like which individual areas of events have drawn larger crowds and which individual speakers or performers have generated the most interest online, and consider vast arrays of big data to draw concrete analytics. 

With the arrival of predictive analytics, businesses will have the power to look beyond sentiment and to consider metadata surrounding specific conversions, popular locations, advanced weather forecasts, variations in social media sentiment, and possible confounding external factors to deliver a comprehensive analysis of exactly what, when and where demand is likely to emerge. 

We’ve already seen firms like JetBlue, a U.S. airline, partnering with ASAPP, a technology vendor, in implementing an AI-based customer service solution that can save an average of 280 seconds per chat, paving the way for saving 73,000 hours of agents’ time per quarter. This platform will one day be capable of learning from customer sentiment and the recurrence of queries to make actionable recommendations to decision-makers regarding processes and the acquisition of stock. 

Predictive analytics: The next generation of data analytics

Having the ability to analyze vast quantities of big data isn’t “generative” by definition, but this part comes into play when generative AI models like ChatGPT use data to create software code that can build deep analytic models. 

According to GitHub data, 88% of surveyed respondents believe that they’re more productive using GitHub Copilot, an analytical tool that’s built on OpenAI’s Codex. Furthermore, 96% of respondents believe that the process makes them “faster with repetitive tasks.”

This will invariably be an invaluable tool for business leaders to generate far more focused data analytics through automated coding. For instance, AI programs have the ability to deliver “automated decision support,” which makes recommendations based on masses of big data. 

In the future, programs will monitor the output and possible areas of employee skillsets that may require improvement and autonomously develop bespoke training programs designed to specifically strengthen these areas based on the employees’ most receptive learning styles. 

Programs could also work in tandem with other sprawling analytical platforms, such as Google Analytics (GA) or Finteza, and use their insights to make automatic tweaks and improvements to company websites based on traffic and performance insights, as well as forecast future traffic.

In addition to this, if a generative AI program learns from GA’s or Finteza’s analytical data that visitor figures have fallen at a time when social media sentiment and seasonal trends indicate that increased engagement should occur, the program could study the issue and make corrections accordingly, while notifying relevant parties or web developers of any changes for subsequent review. 

ChatGPT, for instance, is currently being used a lot for content creation. However, it does come with limitations. For example, below is an example of content generated by ChatGPT. 

The first article is titled, “4 Ways To Recycle Your Glasses,” the second, “How To Recycle Your Glasses.” While both pieces have very similar headlines, the approach to writing the article and the points discussed should vary quite a lot (in real life, at least).

Yet, in the case of ChatGPT, both articles are very similar — identical in some instances: 

As you can see, some content is pretty much identical. Hence, once more than one person opts to use ChatGPT for a similar headline, the issue of duplicate content will arise pretty much immediately. 

This is expected simply because no generative AI can live the lives of thousands of people and experience all of the possible scenarios based on very different life events, situations, personal experiences, characters and habits that human beings possess. All of these factors affect how people write content, the language they use, their writing style and the examples they use. 

Based on this, we can expect to see businesses take on a far more assistive role in realizing the potential of a data-driven future for businesses.

Instead of using platforms like ChatGPT to work on our behalf, these programs can support our business decisions — even if those decisions stem from the example above, whereby generative AI can offer comprehensive discussion points to support content plans. 

Prioritizing privacy

Although the regulatory framework surrounding the growth of generative AI and predictive analytics is still subject to development, early signs suggest that the technology can bring key innovations in the age of GDPR. 

This is because generative AI has the ability to anonymize sensitive data before it’s viewed by human eyes. This empowers predictive analytical tools to generate synthetic data that mimics real datasets without containing any identifiable information. 

>>Follow VentureBeat’s ongoing generative AI coverage<<

Likewise, the software could automatically add and remove identifiable parameters within data, which could help in industries like pharmaceuticals, where drug trials operate on a blind and double-blind basis. 

This represents another major opportunity for businesses seeking to tap into generative AI. Through the creation of privacy-oriented algorithms that protect sensitive information while empowering organizations to analyze the available insights, more firms can act decisively in improving the customer experience.

The greatest business opportunity of the 21st century?

While there’s certainly plenty of work still to be done in terms of creating a regulatory framework to ensure that generative AI grows in a sustainable manner, the potential utility of the technology in the field of predictive analytics is certainly a cause for optimism. 

Because of generative AI’s ability to act decisively in using big data to offer actionable insights, it’s imperative that businesses move to access this potential before they lose ground in the battle for breathing room among companies undergoing digital transformation. 

As well as a significant time-saving tool, generative AI-powered predictive analytics can help organizations gain more immersive insights into performance, which can lead to vast operational improvements. 

Although the technology may need more time to mature in the short term, its future utility can bring significant cost and productivity benefits throughout virtually every industry. 

Dmytro Spilka is the head wizard at Solvid.