Customer support · Production

eSky scales AI customer service across 3 brands and 3 channels with Ada, achieving a 17-point automated resolution increase and 200% ROI

The problem

eSky's flow-based chatbots failed to satisfy customers, who sought human agents rather than trusting automation to resolve their issues, while contact volumes spiked dramatically during irregular travel operations and the company lacked an operating model to scale service quality across multiple brands, channels, and markets.

First attempt

eSky's prior flow-based chatbot approach managed inquiry volume by deflecting tickets rather than resolving them, leaving customers frustrated and seeking human agents instead of trusting the chatbot.

Workflow diagram · grounded in source
1
Customer contacts via preferred channel
Trigger
Customers initiate contact via the communication tool they normally use.
▾ source quote
“they're usually most comfortable with picking up the tool that they normally use for communication”
2
AI agent clarifies and structures request
Ai action
The AI agent works in real time to clarify intent, gather the right details, and structure the request using the right knowledge and Playbooks.
▾ source quote
“the AI agent works in real time to clarify intent, gather the right details, and structure the request using the right knowledge and Playbooks”
3
Automated resolution attempt
Ai action
The AI agent automatically recognizes and resolves customer inquiries when possible.
▾ source quote
“using an AI agent to automatically recognize and resolve customer inquiries”
4
Route to support team if needed
Routing
Inquiries the AI cannot resolve are triaged to the support team.
▾ source quote
“triage them right to the support team if necessary”
5
Human agent handles structured handoff
Human review
Human agents receive a complete, structured request and can immediately focus on resolution.
▾ source quote
“Instead of starting from scratch, human agents receive a complete, structured request and can immediately focus on resolution. That shift not only reduces average handle time, but improves the overall experience whether the AI agent resolves the issue end-to-end or …”
6
Operational gap analytics
Feedback loop
AI reveals knowledge gaps and failure points by exposing where conversations break down, enabling the team to identify improvement opportunities.
▾ source quote
“AI will stick to knowledge resources fully and reveal any gaps instantaneously. We can see everything. We can see where the points of failure are because they always are manifested in the communication channel”
Reported outcome

eSky achieved a 17-point increase in automated resolution rate in four months, 200% ROI, and went live with AI customer service across three channels and three distinct brands managed by one team.

Reported metrics
Average handle timereduces average handle time
Reported stack
AdaAda's MCP ServerWhatsAppMessenger
◆ Does this fit your context?

Compare to your context

Tell us your scale, team, and constraints. We'll show what changes at your size, what fails at your scale, and whether this case is a fit, needs adaptation, or won't scale to you. Free demo, no signup.

Compare to your context →
~30 seconds · free
Source
https://www.ada.cx/case-study/esky
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

eSky achieved a 17-point increase in automated resolution rate in four months, 200% ROI, and went live with AI customer service across three channels and three distinct brands managed by one team.

What tools did this team use?

Ada, Ada's MCP Server, WhatsApp, Messenger.

What results were reported?

Average handle time: reduces average handle time (source-reported, not independently verified).

What failed first in this deployment?

eSky's prior flow-based chatbot approach managed inquiry volume by deflecting tickets rather than resolving them, leaving customers frustrated and seeking human agents instead of trusting the chatbot.

How is this customer support AI workflow structured?

Customer contacts via preferred channel → AI agent clarifies and structures request → Automated resolution attempt → Route to support team if needed → Human agent handles structured handoff → Operational gap analytics.

Related customer support cases

WHAT TO DO WITH THIS

Now compare it to your context

This case is one data point. Whether its pattern fits you depends on your volumes, your stack, and your exception load — that comparison is the step no case study can do for you.

Browse the Customer support cases →