Between bookings, receipts and rejected bills, the “plan → book → spend” cycle still requires too much effort from employees. AI is changing the game by making the OBT (Online Booking Tool) and EMT (Expense Management Tool) more assisted, clearer and more predictable. Here’s a look, from the user’s point of view, at what can be achieved on a day-to-day basis when AI is pushed to the limit and properly managed – with concrete examples to back it up.

Less data entry, more validation

The first tangible effect is the gradual disappearance of unnecessary typing. Thanks toOCR, a hotel receipt taken from a photo is automatically read: amount, dates, currency, VAT; the AI can even suggest itemisation (room, breakfast, taxes). The user validates instead of re-entering everything. The benefits are amplified by automatic integration: corporate card feeds create the expense lines to be reconciled, airline or hotel e-receipts fill in the key fields without scanning, and travel data switches from OBT to Expense (itineraries, dates, cities, cost centres) to generate an almost ready report. In concrete terms, after a booking in OBT, the ticket and overnight stay are already visible in Expense, the e-receipt completes the VAT, and the card line is automatically matched: zero recopying.

Proposals that save time… and avoid hidden costs

The AI doesn’t just display results: it suggests contextualised itineraries. Options are sorted according to policy (ceilings, authorised classes), preferences (time slots, favourite airlines/channels) and objective criteria (door-to-door, CO₂ footprint). Above all, it presents truly comparable fares by including branded fares and ancillaries (baggage, seat), to avoid the fake “good price” that ends up more expensive after adding options. Typical example: “Marseille-Paris day return”. OBT displays several compliant options, explains the difference between “Basic” and “Standard” (baggage included), and highlights the hotel negotiated five minutes from the customer’s site. The employee makes an informed choice, with no surprises when it comes to the bill.

Readable messages and timely alerts

Where policies were sometimes cryptic, AI rephrases them in plain language and suggests an alternative. An out-of-policy flight? The message specifies “+18% of the benchmark fare” and suggests the compliant option at 07:45. The alerts, meanwhile, are actionable: a conflict with the diary (“overlap with a meeting: suggest a second choice?”), or a documentary reminder (“meal >60 € in Paris: nominative invoice required”). The result is faster decisions and less back and forth with the agency or the accounts department.

Estimate the cost before you spend: faster approvals

Another classic irritant: requests that are refused or notes that are rejected for exceeding the budget. AI deals with this issue upstream by building a complete budget estimate for the trip (transport, hotel, extras) based on history, seasonality and policy. This estimate is incorporated into the request for approval, with a contextualised tolerance where appropriate. Example: for London during a trade fair week, OBT proposes a hotel waiver (+10% compared with the local ceiling) with standard justification. Approval is given quickly and, more importantly, the expense claim is not rejected after the event for the same reason.

Compliance at source: less waste, less friction

AI shifts compliance to the moment when it is most useful: before booking and before submitting. On the Travel side, it checks the authorised class, anticipation, and explains the LLF without jargon, proposing equivalent itineraries (for example, same cabin on a night flight if the policy allows) to avoid unnecessary justifications. On the Expense side, it detects per diem vs meals, duplicates (receipt/card line), VAT inconsistencies or missing documents, and alerts the user at the right place. Two concrete examples: a meal is entered when a per diem already covers the day; the alert is displayed before submission and the user corrects it. A duplicate taxi (card line + receipt already used) is detected; the excess line is deleted and the bill is passed on the first try. If an exception is legitimate (delay, overcrowding), the assistant provides a clear, factual , pre-written justification.

What’s in it for the user, day in, day out?

As users work through these mechanisms, they experience three immediate benefits. Time: OCR, e-receipts and card reconciliation reduce data entry to a strict minimum, while suggestions reduce the number of unnecessary clicks. Peace of mind: understandable messages, relevant alerts, budget visibility before spending – fewer unpleasant surprises. Confidence: compliance is controlled from the moment the booking is made, and when an exception needs to be made, justification is assisted.

Simply put, AI puts OBT and EMT at the service of the user: less data entry, optimised choices, an anticipated budget, proactive compliance. The journey becomes clear and predictable – and for travellers and finance alike, this is the best way to save time without losing control.