Corporate lodging and meetings platform HRS has launched an AI tool called Copilot to help guide companies with their strategy and decisions around their hotel programs, the company announced.
The Copilot was developed as a "digital-virtual-assistant-managed program" combining AI specialist Anthropic's large language model with HRS' own specialized language model and "extreme wealth of data," HRS chief product officer Martin Biermann said. That data includes millions of traveler profiles, billions of rate impressions, millions of invoice line items and tax information, millions of corporate rates negotiated over the years and market forecasts.
"All of this sits in our data lake, and we use this to train a specialized model, and we train that virtual assistant capacity for lodging and meetings as a specialized capability into that ChatGPT[-style] experience," Biermann said. "We built out an entire library of skills into that virtual assistant."
Copilot can use those skills to analyze companies' both managed and unmanaged data, pulling in data from the HRS platform as well as external data via APIs from sources including travel management companies, expense and card providers and direct from hotel chains, he said. With that data and its specialized skill set, Copilot can highlight program performance and identify actions hotel program managers can take to improve or meet goals. Program managers can interact and ask questions, written or by voice, such as "What is my biggest savings opportunity in the U.S.?"
"It can come back with actual strategies, not just insight, and predictive intelligence on what can be the impact," Biermann said.
Copilot then can help implement those actions with a "one-click" approach, once program managers have looked at the analysis and potential benefits—such as savings or increased compliance—and risks. For example, the one click could initiate communication with hotel partners to address problems causing low compliance, or it could replace highly used hotels in a program with larger carbon footprints with more eco-friendly hotels to help meet sustainability goals, he said.
"It's not only renegotiating or negotiating different suppliers into your preferred program, but it's also adjusting the configurations of policy controls as well as the recommendations and selection of content in the points of sale, both online in offline," Biermann said. "It's also able to configure differently in centralized payment in order to give more headroom to buy ancillaries if we see the satisfaction is low. Across the entire end-to-end procure-to-reclaim process that we built, Copilot is able to make adjustments in all these engines of our platform."
Those strategies and actions can occur across both the transient program and other lodging program components, including meetings and long-term stays that cumulatively make up about half of an entire lodging program, according to Biermann. That part of the program is "extremely hard to be managed," as they are often "based on content and tools not in the regular corporate travel ecosystem," he said.
HRS already has rolled out Copilot with select leading accounts that helped with its development, and ] now is expanding that to all customers that do procurement with HRS, since it already has their targets and external data. It will gradually be rolled out to all customers, Biermann said.
Deloitte was among the companies using Copilot as part of that initial group, and the company's senior travel manager and procurement and strategic sourcing manager for travel technology Leonard Hornsby said the tool has helped achieve an estimated 18 percent in savings above typical lodging program optimization.
"The speed of the query and answer functionality, knowing that the answers are driven from our comprehensive data, has me rethinking our approach to both short-term and long-term management of transient and meetings segments," Hornsby said in a statement. "In our initial trials, we see insights on issues faster than ever and are making precise program changes as the scenario calls for."
HRS also is continuing to add new capabilities to the Copilot tool, Biermann added. That includes analytics around persona insights—recommendations and analysis based on types of travelers, such as frequent travelers versus infrequent travelers—and hyper-local regionalization, which should be released next week, he said.
"We are developing extremely fast right now," Biermann said. "We also use AI to develop the product, so we are very fast on expanding the capabilities."