The travel and transport industry is rapidly adopting artificial intelligence (AI) to process information and manage routine interactions. However, travel disruption presents a unique challenge. While AI automation offers incredible speed and scalability, unpredictable events like severe weather or operational incidents require a level of human problem-solving and reassurance that technology alone can’t provide.
This presents a critical support gap: according to our Aviation Disruption white paper, 74% of passengers expect support within an hour of disruption, yet only 42% receive it within that timeframe. When hundreds of people require assistance simultaneously, the scale of the problem calls for a technological solution.
In this guide, we'll explore AI vs human support in travel operations – not to declare one methodology a clear winner, but to examine which elements of aviation disruption management are best suited to automation, and where human intervention remains a vital asset.
Why travel disruption is becoming a test for AI
Disruption management differs vastly from general, everyday travel operations. While AI is highly effective at managing predictable events, disruptions are rarely linear, and what starts as a single initial issue can rapidly snowball into a complex web of interconnected problems.
Data from the EUROCONTROL Delays to Air Transport in Europe report revealed that reactionary or knock-on disruption contributed 8.6 minutes to the average delay per flight in September 2025 – equivalent to eight minutes per flight. A delayed departure impacts connecting journeys, accommodation requirements, onward transport, and crew scheduling.
Whether you're dealing with flight cancellations, extreme weather, or operational incidents, circumstances can change rapidly. The true test isn't whether AI can execute individual travel tasks, but whether automated systems can respond effectively when multiple, compounding problems happen simultaneously and don't follow a predictable workflow.
Where can AI make the biggest impact during travel disruption?
To understand the true value of AI in disruption management, we need to look at how well it can handle high volumes of clearly defined, repeatable operational tasks. By absorbing this administrative burden, automation lets human teams concentrate on more complex passenger and operational needs, such as managing accessibility requirements for alternative transport.
Our own operational data offers a real-world benchmark for this capability. Using figures from 6th-12th July 2026, we found:
CMAC's AI voice agent made over 2,000 calls across 200+ active transport suppliers in a single week, averaging more than 300 calls per day.
Over 70% of answered calls reached a successful outcome.
Crucially, the data demonstrates the scalability of these systems. As AI call volume increased by 14% week-on-week , the successful outcome rate remained relatively stable, falling by just 0.6 percentage points. At the same time, transfers to the Operations team actually fell from 7.5% to 6.7%, meaning that in most cases, the AI was able to handle the query itself. This shows that automation can successfully absorb increased operational load without a drop in outcomes.
Matt Ratcliffe, Chief Operating Officer at CMAC, says: "At CMAC, our international operational hubs combine automated tools with real expertise, empowering our operational teams to work faster and smarter. One example of this is using AI call agents to track and monitor suppliers, ensuring accurate delivery and customer communication."
Not every travel task is equally suited to automation
Discussions around technology often fall into two overly-simplified "AI works" or "AI doesn't work" camps. In reality, there is a clear spectrum of suitability, and our data shows that AI performs much better at certain tasks than others.
According to our internal call handling data, the success rate of the AI voice agent varied depending on the specific operational interaction:
Journey progress – 85% successful outcome
Passenger on board – 82.2% successful outcome
ETA/on-time confirmation – 75% successful outcome
Driver on-site confirmation – 73.7% successful outcome
The highest-volume task (confirming whether a driver was on-site) was the lowest-performing metric, highlighting inconsistencies in outcomes for routine tasks. Our data also revealed that many unsuccessful AI interactions stemmed from mundane friction points, like voicemail dead ends or unclear answers from suppliers. This proves that AI performance can't be treated as a single, universal measure. Instead, travel providers should assess suitability on a task-by-task basis, monitoring performance by call type and outcome data.
When should AI hand over to a human?
Human escalation shouldn't be seen as a failure of AI assistance. Instead, effective automated systems should be designed to recognise when an interaction requires nuance beyond their programming, providing a seamless route to human intervention.
Our AI call handling data highlights that 6.7% of all AI calls were transferred to CMAC's Operations team. This demonstrates a functional hybrid model and shows that, even within highly structured, automated processes, a percentage of interactions will still need human oversight.
Human involvement becomes essential during disruptions that feature:
Situations where empathetic judgement is needed, rather than a predefined workflow
Multiple or rapidly changing passenger requirements
Specific accessibility, medical, or welfare considerations
Highly complex onward travel and accommodation arrangements
Cases requiring reassurance, de-escalation, or sensitive communication
"For CMAC, effective disruption management means maintaining clear, immediate escalation pathways between our customer support processes and our 24/7 Operations team. As automation continues to develop, escalation rates should be analysed alongside successful outcomes to identify passenger needs, rather than treated solely as a metric to minimise," says Paul Wardle, Director of Group Operations at CMAC.
At the core of automating disruption management, the travel industry shouldn't be asking how to eliminate human contact. Instead, the focus should be on how to make sure the most complex cases reach the relevant experts.
Do passengers actually want automated disruption support?
There is a lingering industry assumption that passengers reject digital support during stressful travel moments. However, our Aviation Disruption white paper shows that travellers are actually quite open to automation, showing strong passenger adoption when digital tools are accessible, intuitive, and genuinely useful.
The research reveals:
One in three disrupted passengers receives a self-service link
69% of passengers who are offered self-service actively use it
Of those users, 85% find the process easy to navigate
94% receive a choice of hotels and/or transport, resulting in a 63% overall satisfaction rate with the outcome
Passenger satisfaction with digital tools is also improving year on year. Over the course of two years, accommodation rebooking satisfaction increased from 72% to 84%, while flight rebooking satisfaction jumped massively from 57% to 87%.
When it comes to straightforward disruption needs, passengers value the speed, autonomy, and convenience that digital support provides.
Digital-first doesn't have to mean digital-only
While passengers are eager to embrace self-service for routine disruptions, external research strongly supports the need for a hybrid setup. Speed and digital access are vital, but they're not a substitute for transparency and a human approach.
Research from the UK Civil Aviation Authority (CAA) shows a significant communication gap during disruptions. The CAA found that only 10% of passengers feel fully informed about their rights when a flight is disrupted, with 54% reporting they were not informed of their rights by their airport or airline. As a result, passengers clearly look for more frequent, comprehensive updates from airline representatives.
This sentiment is echoed in wider customer experience research. Salesforce's State of the AI Connected Customer report found that, while 46% of customers said they would use an AI agent for faster service, safeguards are still a priority. 72% state it's important to know when they're communicating with an AI, and 46% (rising to 56% among business buyers) are more likely to use an AI agent if there's a clear escalation path to a human when needed.
Crucially, demographic data from our white paper warns against completely abandoning human support for tech-savvy demographics. Despite being more digitally receptive than older generations, 52% of 18-34-year-olds say they're less inclined to rebook with an airline following poor disruption service.
Digital-first certainly doesn't mean digital-only. The strongest disruption model combines the speed and scalability of AI with guaranteed access to human empathy and expertise when a situation exceeds the limits of self-service.
What should the travel industry learn from AI-assisted disruption management?
Successful AI adoption isn't about maximising automation for its own sake; it's about strategically identifying where technology can improve the passenger experience and intentionally designing human intervention into the ecosystem to fill any service gaps.
For aviation organisations looking to refine their approach, there are several core principles to follow:
1. Automate repeatable processes – AI delivers the best ROI when applied to high-volume, clearly defined tasks with predictable outcomes.
2. Build human escalation into automated journeys – Travellers and suppliers should never reach a 'dead end'. Systems need to seamlessly pass conversations on to a human agent when the automated technology can't resolve the issue on its own.
3. Use automation to increase human capacity – By absorbing the bulk of repetitive administrative tasks, technology frees up human teams to dedicate their time to cases that require more nuanced problem-solving.
4. Design technology specifically for disruption – When planning for disruption, remember that a tool that works perfectly during routine, blue-sky operations will likely need entirely different workflows, stress-testing, and escalation routes during a major event.
5. Maintain transparency – Travellers should always know when they are interacting with an automated system and be provided with clear instructions on how to access further human support if required.
6. Measure outcomes rather than automation rates – Success shouldn't be defined simply by the percentage of tasks handled without a human. Resolution speed, accuracy, passenger satisfaction, and appropriate escalation are far more valuable metrics to make sure automated systems are working well.
At CMAC, we apply these learnings by monitoring AI performance by task and outcome, and expanding our automation only where evidence shows it improves speed or capacity. By using expertise across our seven international operational hubs alongside sophisticated technology, we combine passenger and operational feedback to constantly reassess the balance between automated scale and a human-first approach.
AI vs human support: finding the right balance during disruption
The evidence clearly shows that AI and automation have secured a permanent role in disruption management. At CMAC, our operational data proves that AI can process thousands of supplier interactions at scale, while passenger research demonstrates strong adoption and rising satisfaction rates with digital self-service tools.
However, the data also proves why human support can't be phased out. Automated performance fluctuates by task; some interactions will always need expert support, and the interconnected nature of travel disruption inevitably creates complex, knock-on problems. Most importantly, passengers expect access to a real person when they're stressed or need reassurance.
This isn't a competition between AI and humans. The best approach is a combination of the two.
Build a more resilient approach to travel disruption with CMAC
CMAC helps organisations manage complex travel disruption at scale. We combine automation and 24/7/365 human support to ensure passengers feel safe and secure, empowering our clients to offer seamless disruption responses on a global scale.
Through innovative solutions like Smartlink and Drawdown, we help autonomously redirect passengers to suitable transport and accommodation options, presenting airline-approved solutions in minutes. This technology-driven response is backed by our expert team, based across seven international contact centres, meaning human intervention is available 24/7, 365 days a year. Through this hybrid approach, we're proud to have a 99% satisfaction rate across the 5 million journeys we facilitate each year.
Discover how CMAC can support your organisation today. For deeper insights into passenger expectations, download the full Aviation Disruption White paper.