Why corporate travel programs still operate in the dark, and how AI is changing that

HRS’ Copilot Enables Actions On Insights — Another Swipe At Traditional Sourcing

_{by Jay Campbell of The Company Dime / 20 June 2025}

Corporate buyers have been hacking away at the loathed hotel RFP process for years by securing longer terms, reducing portfolios, incorporating dynamic pricing with caps and making off-cycle tweaks. But complexity festers. Data sources seem to multiply rather than simplify. Meanwhile, travel buyers are expected to have a better handle on analytics.

In HRS’ view, this should be a job for AI.

Deloitte and about 15 other customers helped the Germany-based firm build a hotel program optimization assistant that takes in data, compares it against benchmarks and best practices, produces insights and makes market-specific recommendations. Users can chat and respond by clicking a button to change settings in downstream systems, including third-party online booking tools and the HRS sourcing, re-shopping and payment solutions. “I suggest altering that rate cap. Accept with a click.”

Announced last month, the Copilot solution is available free of charge to HRS “Lodging as a Service” clients. The company did not disclose how many of those it has.

“We are mirroring the skillsets of travel managers and procurement managers, and we’re training the model to act and think and work like that,” said Martin Biermann, HRS Group Chief Product Officer, during a May 28 interview. “We’re taking the nitty-gritty data specialization out of it, and we want to empower them to focus on their programs and travelers’ needs.”

The Claude large language model from Anthropic and a specialized HRS industry model leverage anonymized rate and profile data. HRS hosts customer data in a “secured, certified cloud environment,” as it traditionally has. According to HRS, on average, more than 15 system vendors hold relevant hotel program data for a given client, ranging from traditional booking, expense and payment details to sustainability and risk information.

In a BCD Travel April survey of 197 travel buyers, 89 percent said their firms used travel management company data for their travel programs. About two-thirds each used payment, expense and online booking tool data. Fifty-eight percent used data from travel suppliers, 31 percent from traveler surveys and 21 percent from HR systems.

HRS can take all that in by API or batch file, apply persona-based behavioral segmentation and understand corporate goals to make program recommendations. It provides those tips along with their pros and cons, as well as the reasoning behind the conclusions.

To ensure trusted AI output, HRS built the solution with anti-hallucination safeguards, “human-in-the-loop” validation, role-based training and insights, and “chain of thought” explanations.

Martin Biermann, HRS Group Chief Product Officer

“You want to make sure that when you ask it a specific question about your program or a particular market in terms of performance, it responds in the right language and with certain kinds of metrics and insights that are relevant for you and your role, whether you’re a travel manager or procurement manager,” said Biermann.

“We trained these profiles based on the location because if you’re a U.S.-based travel manager, you ask different questions than a Europe- or Asia-based travel manager. And the same for the procurement managers. And we trained the model on answers experts would provide.”

The operational execution of these insights can be applied to RFP management and supplier agreements (e.g., automatic contract terminations for non-performing hotels), rate loading and auditing, search and booking configuration, or policy and compliance management. It’s multilingual and available globally.

Who Needs Data Wranglers?

While almost everyone dislikes the traditional process for obtaining them, few buyers question the value of negotiated hotel rates.

Responding to an April-May GBTA survey sponsored by Radisson Hotels, 37 percent of 171 travel buyers said they had seen a significant or moderate increase in the number of room nights booked using negotiated rates versus a year earlier. Seven percent saw a decrease. At the same time, 20 percent reported a decline and 16 percent noted an increase in the use of non-negotiated rates. “More or a lot more” of the negotiated rates were dynamic for half of the buyers polled, while 31 percent reported increased use of chainwide rates and 21 percent observed increases in fixed rates from individual properties.

According to HRS, new demands over roughly the past decade related to areas including traveler satisfaction, duty of care, personalization, regulatory compliance and sustainability caused the number of criteria in RFPs to skyrocket. At the same time, the company finds that many travel departments have not kept pace in terms of resources and skills.

According to the BCD Travel survey, 74 percent of buyers believed they possessed the necessary skills and experience to interact with data. Fifty-six percent said they “prefer to leave working with data to the professionals,” but 55 percent of companies did not have an analyst working with travel data. Three-quarters said they did not use AI for collecting, analyzing and/or reporting travel data. “Copilot opens a new door for me each time I engage with its functionality,” according to a press statement attributed to Leonard Hornsby, senior travel manager and procurement and strategic sourcing manager for travel technology at Deloitte GmbH, which uses both BCD Travel and HRS. “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. In our initial trials, we see insights on issues faster than ever and are making precise program changes as the scenario calls for. Our initial estimates point towards 18 percent higher savings with Copilot compared to traditional program optimization.”

HRS lists Allianz, DHL and Siemens on its website as enterprise customers of Copilot.

The initiative helps further animate HRS’ longstanding vision for “continuous” sourcing:

“We’re not looking at replacing the annual RFP,” said Biermann. “But we do recognize that by its nature, in the way that it’s structured — with regard to running data analysis, deriving strategies, updating the demand based on internal knowledge that you have about, I don’t know, M&A, then moving from there into a solicitation list by market, reviewing that, starting the negotiation, sending out notifications and stuff to national account managers and individual hotels, collecting bids, analyzing them, going to rebid, doing rate comparison — that process is very lengthy and not optimized for on-demand optimizations on a micro-destination level.

This is where we had initiated some years ago a new way of running continuous procurement, and that is something that we’re going to innovate next in order to make that more vivid and meet the expectations of the user interacting with Copilot. Because if you now run your data analysis instantaneously, you get the strategy and recommendation instantaneously, and you can, in one click, activate to say, ‘Yes, I want this change in my program in this market,’ then you need to act and also engage with the suppliers accordingly. And this is where we are building an agentic AI later that basically takes over and automates the negotiation and RFP details for continuous procurement practices.”

HRS is testing the use of AI agents for sourcing, which could be ready for production this year. The company already has voice bots handling customer service interactions on bookings it makes.

Rather than creating a risk of making some roles redundant, HRS sees the solution as addressing a mismatch between required skills and available talent in travel management. According to Biermann:

“AI has that general risk, but it’s not our intention. We’d rather look at this from an empowerment point of view for travel managers and procurement managers. The complexity of specialization they need to have, to understand the deep data details of sustainability information, duty of care information (particularly after the pandemic) on top of all of the savings aspects — understanding how rates work, dynamic rates, static rates, chainwide discounts — it’s just added to the list of competencies that these profiles need to bring. The outcome is that it takes more time to hire these people, and we see this in feedback from our clients. They say that even if they wanted to basically backfill the roles that are retiring now, they don’t find the people. And so we see that the biggest challenge in finding these competencies is dealing with the data, whereas understanding the industry — how hotels work, and how to run a hotel negotiation in an RFP — is where you will find people. And this is what we can compensate for with Copilot.”

There are no plans at this stage to introduce Copilot to procurement for air, car or other non-lodging segments.

Republished with permission.

_{Copyright 2025 B2B Reporter dba The Company Dime}

‍