Let me cut to the chase: anyone can build an AI agent. In fact, you can get an AI agent to build an AI agent!
Right now in corporate travel, not a week goes by without the launch of an AI agent, with another fancy name. But if you look more closely, these are just widgets that sit on top of a booking tool. They are essentially screen scraping, and that’s not a new innovation, nor a great use of AI. These AI agents might fill in some gaps but they are not bringing a huge amount of value. It’s AI for AI’s sake.

I’ll be honest – if SAP Concur were a standalone travel company, we’d be in the same situation because we couldn’t afford the investment required for tokenization – only the tokens themselves.

But SAP Concur is part of SAP SE, Europe’s largest software company, and it’s investing significantly in AI.

Furthermore, SAP is a multitude of companies offering software, cloud applications and data management platforms for business operations – SAP Fieldglass, SAP ERP, SAP SuccessFactors and SAP Ariba, to name just a few. Concur is just one of them.

What we’re working on now is for Joule, SAP’s AI assistant, to work across all of these platforms. It’s about fully embedding AI into an organisation’s work structure, its entire ecosystem, rather than having a standalone tool that just sits on the side or only sits on top of a portion.

Having AI agents that work in isolation is meaningless and that’s why you’re now seeing people talking about MCPs, Model Context Protocol, which allow AI agents to talk to each other. But agents shouldn’t be standalone separate ones you use for travel, and another for expense, another for your TMC, and another for your finance.

We don’t want to ‘lock out’ external AI agents,  as we do connect to partner and supplier agents, with the proper data sharing approvals. What we want to avoid is an AI agent that just works in a silo, as this will bring no tangible benefits in the long term.

“These AI agents might fill in some gaps but they are not bringing a huge amount of value. It’s AI for AI’s sake”

Having fully-embedded AI brings a number of key benefits. Firstly, it means you don’t necessarily need to be in Concur any more to make a travel booking. Let’s use an example that could happen in the future:

A Senior VP is leaving and this requires an exit interview before they hand over their laptop. The AI agent will inform the HR manager of this through SAP SuccessFactors. But it won’t just flag that an exit interview is required, it will go into Concur Travel and find three flight options that will allow the HR manager to be there for the interview at 3pm on the employee’s last day. The HR manager won’t need to go into Concur to book the flight, as it will sit within SuccessFactors, so productivity is already improved. Better still, the AI will also be embedded in expense, and also in finance, so the travel options being presented will automatically be in budget, in policy and payments will be automatically processed.

Another key benefit of moving to an autonomous enterprise approach relates to trust and governance.
77% of the world’s transaction revenue touches an SAP system – ERP (Enterprise Resource Planning) data, expense and travel data, CRM, finance data and so on.

We’re a large European company that’s heavily regulated, so trust and governance is built throughout all of our products – all of SAP’s solutions adhere to UNESCO’s ethical AI standard. With AI, we are able to bring everything together and use it in an intuitive way, but also avoid the risk of employees sharing corporate information where they shouldn’t.

In our recent survey of 3,300 travellers and 700 Chief Financial Officers, we found 75% of travellers are already using AI tools and these are often public tools. They’re asking these public tools to help draft their work emails, summarise a contract, or organise their holiday. They’re sharing private information on a public AI tool, even though they wouldn’t dream of sharing on Instagram that they’re going on holiday for two weeks.

Many corporates are still slightly reluctant to share data with a third-party AI tool, and that’s because of concerns about trust and governance but also because the legal teams are all navigating guardrails to put in place in this rapidly expansion of AI. They need to know if it’s secure, which LLM does it connect to, is it shared amongst large language models that other people may have access to, etc. More traditional companies are holding back, concerned about where their data is going, who is going to be handling it, and what it’s being used for. 

“With this fully-embedded agentic AI, I believe travel managers have the potential to play an even bigger role in their organisations”

We’re now bringing that into the AI world, which is giving these companies more confidence and they are happier to share their data with us. And once that data is shared, it’s amazing what you might be able do with it in the near future.

In travel and expenses, for example, trending analysis brings tremendous opportunities. We can review expenses to see if your people are suddenly travelling more frequently to a destination where you don’t have a corporate deal in place. In the future you will be able to go through SAP Ariba and get a negotiated rate put in front of the travel manager before the travel manager even knows that they need to do so.

The fully-embedded agentic suite knows your policies, it knows your programmes, it knows your finance systems, it knows your HR, and so on, so it can now come up with true recommendations as part of the flow of data and analytics.

Of course, you’ll still need actual intelligence on top of it, and real people to approve the recommendations, but the job of a travel manager, a finance manager, an HR manager, will be easier and more productive.

With this fully-embedded agentic AI, I believe travel managers have the potential to play an even bigger role in their organisations. It can open up opportunities because you’re talking about managing a process that no longer has limits. T&E will no longer be a standalone division, it will be fully embedded in the organisation along with finance and HR, where everything is controlled centrally. Therefore, travel managers should be embracing AI and securing themselves a place on their organisation’s AI taskforce.

Travel managers who step forward now and position themselves as AI champions within their organisations will be the ones defining what comes next. This is not a threat to the profession, it is arguably the biggest opportunity in a generation to expand influence, drive real business value, and lead a transformation that reaches far beyond booking flights and processing expenses. The tools are ready. The only question is whether travel managers choose to own this moment.

concur.co.uk