The shimmering promise of artificial intelligence in business travel is, on its surface, a promise of effortlessness. It’s a world where a complex itinerary spanning flights, hotels, ground transport and policy compliance assembles itself from the digital ether, perfectly attuned to both corporate rules and personal preference. As a financial journalist watching this unfold from the Financial District, I’ve seen countless technologies arrive with similar fanfare. The actual story, however, isn’t written in lines of code or model parameters. It’s written in a far more human currency: trust.
I spoke recently with Evan Konwiser, the chief product and strategy officer at American Express Global Business Travel. His perspective, forged at the nexus of consulting, entrepreneurship and travel tech, cuts to the core of the adoption challenge. “If you don’t trust it, you’re not going to use it,” he told me plainly. Without that user faith, the most elegant AI architecture is just an expensive hypothesis. Konwiser’s view is that trust cannot be a marketing afterthought or a compliance checkbox. It must be the foundational engineering principle, baked into the product’s very architecture from the first line of code.
This challenge represents a fundamental inversion of a previous technological revolution. The early internet won trust through radical transparency and user empowerment. It handed the traveler the keys to the inventory database, replacing the opaque recommendations of a phone agent with a self-directed search through every fare and seat. The trust model was built on visibility and control.
AI proposes a different bargain. As Konwiser explained, it operates on a logic of curation, not cataloguing. The system analyzes the overwhelming array of options behind a digital curtain and presents a reasoned recommendation. Its value is profound convenience, but its demand is significant faith. The user is asked to trust the system’s unseen logic and priorities. In a low-stakes scenario – like a streaming service recommending a movie – this is a trivial gamble. In business travel, the stakes are career and personal realities.
A payment system glitch can be corrected. A travel failure can mean missing a billion-dollar merger signing, a final flight home before a holiday or a child’s graduation. The costs are multidimensional: financial, professional, emotional. This is why Konwiser emphasizes “showing your work” as a non-negotiable feature. If a flight is excluded for violating policy, the traveler must see the rule. If an option is recommended for saving $200, the savings must be highlighted. If the recommendation feels wrong, the traveler needs a clear path to challenge the logic and request a recalculation. Trust, in this context, is the earned confidence that the system’s reasoning is sound, aligned and corrigible.
This leads to the second pillar: personalized trust. An effective AI doesn’t just need data; it needs nuanced context. A traveler might prioritize price for a short hop but demand comfort and schedule certainty for a transoceanic red-eye. Konwiser’s team is developing what they term “Traveler DNA,” a dynamic profile that learns from past choices to anticipate future needs. But here, business travel holds a structural advantage for AI adoption. Its environment is more constrained and rule-bound than the infinite variables of a leisure vacation. The platform knows the trip is for work, understands the corporate policy, has booking history and knows the payment method. This narrower context allows an AI agent to become useful faster, operating within clearer guardrails.
- Trust must be foundational
- AI operates on curation, not cataloguing
- Showing your work is essential for trust
- Personalization requires nuanced context
- Data stewardship is critical
- Adoption hinges on context and rules
Yet personalization immediately raises the critical question of data stewardship. Who owns the preference profile – the employee, the employer, the travel management company? In an era of heightened regulatory scrutiny, from GDPR to emerging U.S. state laws, there are no shortcuts. Konwiser is adamant that robust systems for security, governance and data deletion are prerequisites, not features. This mirrors a broader market lesson I’ve observed across sectors. Consumers and corporations will share data conditionally, when the value exchange is transparent, the benefit is clear and the brand acts as a responsible fiduciary, not a data baron.
From a competitive strategy standpoint, this trust-centric view redefines what constitutes a “moat.” As generative AI models become commoditized, true advantage won’t spring from merely having access to GPT-5 or Gemini. Konwiser points to a more durable combination: proprietary context, marketplace depth and operational execution.
| Assets | Description |
|---|---|
| Policy engines | Frameworks guiding operational rules |
| Corporate rates | Negotiated prices with major hotel chains and airlines |
| Global payment rails | Systems facilitating financial transactions |
| Fare classes | Normalized systems managing various pricing tiers |
| Historical data | Insights into business traveler booking behaviors |
| Integrated plumbing | Reliable underlying systems supporting travel operations |
Amex GBT’s real asset isn’t a language model. It’s the decades of unglamorous infrastructure built for managed travel: the policy engines, negotiated corporate rates with major hotel chains and airlines, global payment rails and the complex systems that normalize thousands of disparate fare classes and room types. It’s the historical data on how business travelers actually book. This proprietary context is what allows an AI to move from a convincing conversationalist about travel to a competent agent that can actually book the right trip, at the right rate, within the right rules. The lesson, seen in previous tech cycles from cloud computing to SaaS, holds true. The flashy interface gets the headlines, but the enduring value is often in the trusted, deeply integrated plumbing.
Finally, adoption will be decided not in a standalone AI app but within the existing workflow. Business travel is a connective tissue in corporate operations, linking to calendars, communication tools like Microsoft Teams or Slack, expense platforms like SAP Concur and ERP systems. Konwiser described the “holy grail” as AI that reduces manual steps across these environments. Imagine a system that detects a calendar entry for a meeting in London, proposes a compliant itinerary, checks for conflicts, posts the final details back to the calendar and pre-populates the expense report. The goal is for AI to feel less like a new destination and more like a quiet, capable assistant within tools people already use, eliminating friction while remaining explainable when it matters most.
The macroeconomic takeaway for any business integrating AI is this. Adoption will not be won by asking users to marvel at the technology’s brilliance. It will be won by consistently proving the system understands the context, respects the rules, protects the data and can explain its choices. In finance, we see this with robo-advisors; their growth was contingent on demonstrating reliable, transparent logic behind portfolio allocations.
In business travel – and indeed, across the enterprise – trust is shifting from an emotional brand outcome to a functional product requirement. It is built through relentless transparency, user control, situational relevance and unwavering performance. The best AI, Konwiser believes, may eventually feel invisible, seamlessly removing daily friction. But the trust required to make it disappear must be earned in plain sight, one verified recommendation, one clear explanation and one successful trip at a time. In the high-stakes world of business travel, that trust is the most valuable currency of all.
*Sources: Interview with Evan Konwiser, American Express Global Business Travel; U.S. Federal Trade Commission guidelines on AI and consumer protection; “The Economics of Trust in Digital Systems,” Journal of Economic Perspectives; Market analysis on corporate travel software adoption from Phocuswright.*
