ACRO-Banner
ACRO-Logo

Inspiration - When Flights Are Disrupted, Customer Recovery Becomes the Real Challenge

A flight cancellation may take only minutes to announce, but recovering the journeys of thousands of affected passengers can take hours or even days.

When disruptions occur, passengers face uncertainty surrounding their replacement flights, baggage, accommodation, entitlements and onward travel. At the same time, airline customer service teams face surging enquiries, fragmented information and increasingly complex recovery decisions.

Major incidents such as Southwest Airlines' December 2022 operational disruption, the July 2024 global IT outage and Heathrow Airport's March 2025 closure illustrate the scale of this challenge. While the underlying causes differ, the consequences for passengers and airline operations are remarkably similar.

The problem is no longer just getting an aircraft back into operation. It is getting every affected passenger's journey back on track.

Traditional recovery processes often operate across disconnected reservation systems, service channels and operational teams. Passengers may be required to repeat their circumstances across multiple interactions, while employees spend valuable time retrieving information and coordinating the next course of action.

We saw an opportunity to reimagine this experience.

What if airlines could move beyond reactive customer service and instead provide an intelligent, coordinated recovery journey - one that understands passenger context, recommends appropriate actions, supports human decisions and communicates recovery outcomes?

That vision became ACRO - Agentic Customer Recovery Orchestrator, developed by HCLTech Airgents.

What We Built - One Recovery Journey, Two Connected Experiences

ACRO is an AI-assisted passenger recovery orchestration platform designed to help airlines respond to disruptions with greater speed, consistency, visibility and accountability.

Rather than treating every passenger interaction as an isolated enquiry, ACRO brings together passenger context, recovery options, operational decisions, human escalation and customer communication in one coordinated workspace.

The solution is designed around two complementary personas.

Persona 1: The Passenger - Clarity, Confidence and Continuity

For a disrupted passenger, the immediate priority is not understanding the complexity of airline operations. It is knowing what happens next.

Through ACRO's conversational passenger experience, travellers can seek assistance with alternative flights, baggage, accommodation and other recovery concerns.

The experience is built around continuity. A passenger who has already been offered a replacement flight should not have to begin their entire support journey again when asking about their baggage or accommodation.

ACRO retains the recovery context so subsequent enquiries can be interpreted in relation to the passenger's existing journey.

For example, a passenger may have been rebooked following a cancellation but still be concerned about checked baggage remaining at the departure airport. ACRO can interpret that new enquiry, identify the relevant recovery information and provide contextual guidance.

The intended passenger benefit is simple: fewer repeated explanations, greater clarity over the recovery process and more confidence that their journey is being managed.

Persona 2: The Airline Operator - Intelligence, Visibility and Control

While passengers need reassurance, airline recovery teams need the ability to coordinate thousands of individual journeys under significant operational pressure.

ACRO provides an operational Command Center that brings together affected passenger cases, recovery progress, outstanding actions and cases requiring human attention.

Airline operators can examine an individual passenger's itinerary, entitlements, conversation history and recovery status within a unified workspace.

Agentic AI assists by interpreting enquiries and identifying relevant recovery options, while human specialists retain oversight of consequential decisions and complex exceptions.

When automated assistance is insufficient, the case can be escalated with its supporting context, reducing the need for specialists to reconstruct the passenger's situation.

The platform also maintains execution histories, allowing operators to understand which recovery tools were invoked and what information was returned.

For the airline, ACRO's value lies in turning fragmented service interactions into a more coordinated, transparent and manageable recovery operation.

How Agentic AI and AWS Communication Developer Services Work Together

ACRO combines two complementary capabilities: intelligent recovery assistance and actionable customer communication.

At the centre of its AI capability is Amazon Bedrock AgentCore, using Amazon Nova to interpret passenger requests and select appropriate recovery tools.

Rather than relying solely on predefined chatbot scripts, the agent can determine which approved tool is relevant to the passenger's enquiry.

Our deployed implementation demonstrates three recovery capabilities: searching alternative flights, checking accommodation entitlements and retrieving baggage status.

This provides a foundation for more contextual assistance, where the response is supported by relevant recovery information rather than generated in isolation.

Equally important, ACRO is designed to preserve human accountability. The AI currently uses controlled, read-only recovery tools; it does not independently commit airline reservations or perform unrestricted transactions.

For the communication component, we integrated Amazon Simple Email Service (Amazon SES) to deliver actual recovery itinerary notifications.

An authorised operator can initiate the email dispatch through a protected backend workflow. We successfully verified genuine SES message acceptance and receipt of the corresponding recovery email.

This distinction is intentional: AI assists with understanding and recommending the next step, while the airline maintains control over decisions and external communications.

Together, AgentCore and SES demonstrate how intelligence, operational oversight and communications can become complementary parts of the customer recovery journey.

What Makes ACRO Different

ACRO's differentiator is not simply its ability to answer a passenger's question.

It is the ability to connect that question to a broader recovery process.

A conventional chatbot may provide an answer and end the interaction. ACRO is designed to preserve the context of the disrupted journey, identify relevant recovery information, support operational decision-making and maintain visibility into what happens next.

The platform brings together capabilities that would otherwise be fragmented across customer service and operational workflows:

1) Context-aware assistance: Passenger enquiries are considered alongside available journey and recovery information.

2) Agent-assisted recovery intelligence: AI selects relevant tools to support recovery recommendations.

3) Human-in-the-loop control: Specialists remain responsible for complex decisions and consequential recovery actions.

4) End-to-end visibility: Recovery cases, conversations, escalations and execution records are brought into one operational experience.

5) Customer communication: Authorised recovery updates can be delivered through a genuine AWS communication service.

The result is a solution designed to address not merely the efficiency of individual customer interactions, but the effectiveness of the overall recovery operation.

Challenges We Encountered

Turning a Complex Airline Recovery Process into a Coordinated Customer Journey

One of our biggest challenges was determining how to represent the full passenger recovery experience without reducing it to a simple chatbot conversation.

Airline recovery involves several interconnected considerations: flight availability, passenger eligibility, baggage handling, accommodation, communication and operational escalation.

We therefore designed ACRO around the complete recovery case rather than a single message. This allowed us to demonstrate the relationship between the passenger experience and the operator's decision-making environment.

Designing for Enterprise Airline Systems Without Direct Production Access

Real airline recovery depends on authoritative information from reservation, inventory and operational platforms.

Systems such as Amadeus Altéa and Navitaire New Skies represent the types of airline passenger service platforms that could provide important booking and flight information in a production implementation.

However, accessing these enterprise systems requires authorised airline relationships, approved interfaces and integration arrangements beyond the scope of a hackathon.

We addressed this by creating representative airline data and controlled recovery tools, allowing us to demonstrate the intended orchestration approach while keeping a clear distinction between the working application and future production integrations.

This also reinforced an important commercial consideration: ACRO should complement an airline's existing systems of record rather than attempt to replace them.

Balancing Agentic Intelligence with Operational Accountability

A second challenge was determining how much autonomy an AI agent should have when assisting with passenger recovery.

Finding alternative flights or checking baggage information is fundamentally different from confirming a reservation, issuing compensation or notifying a customer.

We designed the implementation to demonstrate genuine AI-driven tool selection while maintaining boundaries around consequential actions.

This approach reflects our belief that enterprise agentic AI should enhance employee decision-making without removing the controls required for accountable operations.

Connecting Recovery Decisions with Meaningful Communication

Customer recovery does not end when an appropriate option has been identified.

Passengers must also understand their outcome and receive reliable follow-up communication.

Integrating Amazon SES allowed us to demonstrate a genuine outbound communication capability, connecting ACRO's recovery environment to an AWS Communication Developer Service.

We deliberately retained an authorised operator action for email delivery rather than allowing every conversational request to trigger an external message.

What We Accomplished

We delivered a functioning, AWS-hosted recovery orchestration application that demonstrates both the passenger-facing and airline-operator experiences.

Our implementation brings together unified recovery case management, conversational passenger assistance, real Amazon Bedrock AgentCore execution, human escalation workflows, auditable tool traces and actual Amazon SES email delivery.

Most importantly, we demonstrated how an AI-assisted recovery enquiry can be linked to the passenger's broader operational context.

The application uses representative airline and passenger data, and its conversational interface does not currently connect to a live WhatsApp or RCS service. However, the AgentCore runtime execution and SES email delivery are genuine AWS integrations.

For us, the most significant accomplishment was validating the recovery orchestration concept: AI does not need to replace airline operations to create value. It can help the people responsible for those operations access relevant information, coordinate decisions and communicate more effectively.

The Industry Impact - Moving from Reactive Service to Proactive Recovery

Airlines operate in an environment where disruption cannot always be prevented. Weather, technical failures, airport restrictions and operational constraints will continue to affect passenger journeys.

What can change is how effectively airlines respond.

ACRO offers a potential pathway towards a recovery model where airlines can:

Improve service productivity. Reduce the manual effort involved in locating recovery information, understanding passenger circumstances and coordinating routine enquiries.

Strengthen passenger confidence. Provide more consistent, context-aware assistance and clearer communication throughout a disrupted journey.

Improve disruption resilience. Give operational teams a consolidated view of recovery cases and enable better prioritisation when enquiry volumes increase.

Maintain operational accountability. Combine AI-assisted recommendations with human oversight and traceable execution.

Protect long-term customer relationships. Transform a potentially negative disruption experience into an opportunity to demonstrate care, responsiveness and service reliability.

These are the commercial outcomes ACRO is designed to enable. Quantitative benefits such as reductions in handling time, recovery cost and escalation rates would require validation through an airline pilot.

Beyond improving individual interactions, the broader opportunity is to help airlines manage disruption recovery as a coordinated business capability rather than a collection of disconnected service activities.

What We Learned

We learned that successful agentic AI applications require more than a capable language model.

They need meaningful business context, controlled access to relevant information, clearly defined responsibilities and an understanding of how AI recommendations translate into operational outcomes.

We also learned that passenger experience and operator experience cannot be designed independently.

A passenger's confidence depends on the airline's ability to make and communicate recovery decisions. An operator's productivity depends on having the right context and tools to manage those decisions efficiently.

ACRO was built around connecting those two experiences.

What's Next - From Hackathon Solution to Enterprise Recovery Platform

The next stage for ACRO would be integrating authorised airline systems of record, including passenger service platforms such as Amadeus Altéa or Navitaire New Skies, together with operational baggage, customer profile and disruption management systems.

These connections could allow recovery recommendations to be informed by live flight availability, booking status, passenger eligibility and operational constraints.

We also envision extending ACRO's communication capabilities to approved SMS, RCS and WhatsApp business channels, enabling airlines to reach customers through their preferred communication methods while preserving recovery context.

Future enterprise development would include stronger identity and access management, integration governance, operational monitoring, scalable data services and appropriately controlled transactional workflows.

Although aviation provides a compelling use case, the orchestration pattern is reusable.

The same underlying approach could support service recovery across banking, logistics, utilities, healthcare and other industries where organisations must manage customer-impacting events, coordinate decisions and communicate outcomes.

Our Vision

ACRO represents a shift from reactive customer service to intelligent, coordinated recovery.

For passengers, that means greater clarity and continuity when their journeys are disrupted.

For airlines, it means the potential for more productive recovery operations, better-controlled decisions and more consistent customer communication.

For the wider industry, it demonstrates how agentic AI and AWS Communication Developer Services can work together to turn complex recovery challenges into connected customer experiences.