OpenAI’s large language model (LLM) AI-based chatbot known as ChatGPT shattered multiple records when its monthly active users count hit 100 million just two months after its initial release. ChatGPT’s adoption numbers continue to soar as the chatbot rapidly evolves with a growing list of plugins and a mounting number of third-party distributions. 

Despite its overwhelming popularity and adoption rates, figuring out how to employ tools like ChatGPT for corporate use cases remains both a challenge and an opportunity. 

Vendors have been quick to pounce by incorporating ChatGPT in existing enterprise-level technologies, including a spread of Microsoft embeds ranging from Bing to Dynamics 365, and a recently released enterprise-packaged version in Azure OpenAI Service

Here then are six ways companies are already using LLM models, and how experts believe generative AI such as ChatGPT could fold into enterprise needs moving forward.

1 – Finding insight in seconds

Financial services companies are eagerly tapping into generative AI. American Express is looking to put LLMs to work approving transactions around a customer’s spending habits, and even approving new cards and lines of credits. For example, financial data company Bloomberg recently launched BloombergGPT, a customized AI model trained on financial data. Bloomberg’s customers can gain insights into financial scenarios by simply asking the AI a question in their own language. And Bloomberg is looking to build new products and services on top of Bloomberg GPT to create new revenue streams and other business lines.

“For all the reasons generative LLMs are attractive – few-shot learning, text generation, conversational systems, etc. – we see tremendous value in having developed the first LLM focused on the financial domain,” says Shawn Edwards, Bloomberg’s Chief Technology Officer in a release in March. “BloombergGPT will enable us to tackle many new types of applications, while it delivers much higher performance out-of-the-box than custom models for each application, at a faster time-to-market.”

Even as quickly as Bloomberg is moving forward with its AI model, it came two weeks behind a similar move by financial services company Morgan Stanley, currently OpenAI’s only strategic client in wealth management. Morgan Stanley says its Wealth Management division is developing an internal-facing GPT-based AI service to “deliver relevant content and insights into the hands of financial advisors in seconds.”

“The work we are with [OpenAI] will allow us to bring out expansive intellectual capital into the hands of our financial advisors in seconds – it will be like having our Chief Investment Strategist, Chief Global Economist, and Global Equities Strategist on call for every financial advisor 24/7,” Jeff McMillan, Head of Analytics, Data & Innovation for Morgan Stanley Wealth Management said in a release in March.

Courtesy of Bloomreach and Boden

2 – Tailoring marketing and ads

AI has also had its hands in developing marketing and advertising campaigns for brands — and has for some time. A 60-second TV spot for Lexus, which the company claimed was written entirely by AI, aired all the way back in 2018. Now, advertising giant WPP and Nvidia are building an advertising engine designed to give generative AI tools to creatives, tapping into Getty Images and other resources, the two announced in May.  

AI tools can develop and serve marketing at the micro level too, creating merchandise displays or personalized options honed to a customer’s unique interests and their location. Bloomreach, for example, partnered with British clothing retailer Boden to deliver online clients clothes tailored to their preferences, and in combinations that worked together as an outfit. The AI also tapped into where a buyer lived, serving up clothes that made sense for their climate. 

New launches from Google and Meta will put AI tools into even more hands. With Google’s newly launched AI-based Product Studio, sellers can design entire photoshoots online, changing out backgrounds, creating seasonal layouts and boost resolution. Meta unveiled its own AI-driven advertising tool in May, while Snap is piloting sponsored ads generated by AI based on chatbot conversations with customers. 

“The way we think about e-commerce 10 years from now may be entirely different than how we’ve thought about it these past 10 years,” says Raj De Datta, the co-founder and CEO at Bloomreach.

It is widely predicted ChatGPT will become a personalized shopper, suggesting specific goods to people based on their preferences and needs such as activities, interests, and sizes – rather than customers searching different websites to see what products are available. Already AutoGPT produces AI agents that can seek out and deliver user-specific products and services that best fit specifications from quality to price, size to color. 

“E-commerce is soon going to change in seismic ways, and that is incredibly exciting for businesses and shoppers alike,” says De Datta.

Courtesy of Instacart

3 – Engaging directly with customers

ChatGPT and other AI tools are also working their way into the enterprise space to converse directly with consumers. Instacart is launching an add-on with ChatGPT to help customers dream up meal ideas from their grocery list, and Salesforce’s Einstein GPT is powered by OpenAI and designed to work with the company’s client base.

General Motors, already in partnership with Microsoft to make driverless cars, is exploring several ChatGPT projects centered on making these vehicles more appealing and useful to customers.

“ChatGPT is going to be in everything,” said Scott Miller, General Motors vice president, to Reuters. “The chatbot could be used to access information on how to use vehicle features normally found in an owner’s manual, program functions such as garage door code or integrate schedules from a calendar.”

4 – Implementing real-time error controls

In manufacturing, Retrocausal, a provider of AI-powered augmentation systems, recently released its proprietary LeanGPT series, including its first application called Kaizen Copilot

The product integrates Lean Six Sigma, a method of improving manufacturing processes by gathering information on assembly lines as it’s happening, and offering corrective signals to workers. It also taps into Toyota Production Systems (TPS), an operations system that helps Industrial Engineers (IE) design and improves industrial assembly processes more efficiently. 

Retrocausal has unveiled joint projects, including one with Honda Innovations, that observes workers’ actions on a production line, record errors, and use an audible tone to notify workers if they perform a process step incorrectly. The company also teamed with Siemens Digital Industries to similarly manage live manual assembly processes via AI video-based analytics and real-time task guidance for workers on the line.

Audi is also using AI to find small production line errors, such as welding imperfections or spying cracks in metal that come off the press line. And the carmaker is eager to put AI into practice in other parts of the production process, including helping to “increase the output or performance of processes or machines,” says Audi’s AI expert Rüdiger Eck

5 – Finding and fixing production issues 

Ford Autosan, Ford’s manufacturing company in Turkey, uses AI to find weakness along its line, propose ways to optimize their output, and even run virtual mockups or “digital twins” of the suggested changes. 

Semiconductor maker Analog Devices uses AI to monitor production lines for “wander,” and can also automate repairs to keep manufacturing schedules running. Concerns are sent to engineers that oversee specific products so they can be studied further.

Retrocausal helps companies locate and catalog issues, and share them with stakeholders to “…compress tasks that used to take weeks into minutes and hours,” says Zeeshan Zia, the company’s CEO.

6 – Delivering rapid prototyping.

Generative AI has helped to spur innovation, generate ideas, identify design flaws, calculate production costs, and rapidly produce blueprints and instructions. For example, NASA turned to AI to craft elements for telescopes and scanners designed to travel into space. But prototyping isn’t limited to manufacturing. 

The method can also create consumer products or services, such as specialty AI bots, computer codes, mobile apps, computer gaming, blog and website designs. As with prompt generators already available online, consumers can find online manuals on how to direct AI to craft a customer-facing support bot, for example, or a meal-planning app.

With ChatGPT and other GPT model applications already at work at the enterprise level in marketing, manufacturing, and beyond, companies that haven’t begun to consider how generative AI will augment how they run their business should start now. It’s clear their competitors already have.

“The significance of what we’re on the precipice of with AI really can’t be understated,” says De Datta.