Over the last 14 or so months since I started writing this blog, I've written about technology, automation, the metaverse, and the unique differences between hospitality and service. Welcome to the conversation ChatGPT.

What is ChatGPT?

ChatGPT, launched in November 2022 by the startup OpenAI, is a Generative Pre-trained Transformer 3 (GPT3) artificial intelligence language processing model. While the model can be trained, is not connected to the internet, and has limited knowledge of the world and events after 2021. In summary, an advanced chatbot that can be trained to answer questions and generate detailed responses in a conversational manner.

Automated Customer Service​​​​​​​

This is probably the easiest leap to make as this is already occurring. Chatbots using pre-scripted responses to common guest requests have been around for at least the last 5 years or so.

Personalized Recommendations

I am a bit skeptical of this one. The tool would need to have a history of your preferences and aversions. Given concerns around privacy and data sharing across apps and platforms, this may have limited value.

In-room Automation

Marriott and Wynn started testing this type of technology with Amazon's Alexa, and Apple's Siri products in 2018. There were, and continue to be concerns around privacy, and how data will be stored and used. It doesn't seem like this ever really caught on with guests, or hoteliers. Amazon only lists one case study and a handful of articles dating back to 2021 that feature their "Alexa Smart Properties for hospitality" product.

Staff Training

I could see this working more like a real-time FAQ or how-to tool. The tool could be trained on all existing documentation for the software and technology used by an organization. A user could then have a single source of support instead of login into each application and using the native help feature. I could also see the tool being trained to behave like a variety of different guest types for frontline associate training purposes. This could be used to practice difficult or uncomfortable conversations.

Data Analysis

Given that the tool is already processing natural language, using it for sentiment analysis is a logical extension of the capability. There are a variety of dedicated sentiment analysis programs on the market today, the majority of which employ AI and natural language processing. Using the ChatGPT as the basis of an in-house sentiment analysis tool could produce a more specific and actionable analysis of guest feedback for hospitality operators.

I've only focused on the possible, and only on how ChatGPT thinks it can be useful to the hospitality industry. ChatGPT has many known faults and shortcomings, for example, it doesn't handle math word problems or math at all particularly well. This technology has also been known to develop biases based if the data it is trained on is biased. There is much refinement to be done, yes. While I don't foresee chatbot front desk agents running the registration desk anytime soon, there are many opportunities to deploy this type of technology against rote task automation.

About BirchStreet Systems

Founded in 2002, BirchStreet is the leading global provider of enterprise procure-to-pay solutions for the hospitality sector. BirchStreet's software platform provides leading global hospitality brands with a full suite of procurement applications, including eProcurement software, accounts payable automation, invoice management, inventory management, recipe management, supplier payments, and reporting and analytics. BirchStreet's customers leverage the company's solutions to access a broad network of suppliers and manage billions of dollars of spend annually across a wide variety of categories, including food and beverage, operating supplies, maintenance services and capital projects. For more information, please visit www.birchstreetsystems.com.