The image features a stat on the left that reads, “50% customer inquiries resolved by Agentforce.” On the right is a chat interaction of a customer asking Agentforce to cancel a car rental for an upcoming trip to which Agentforce identifies the trip and then cancels the car rental.

Engine resolves 50% of chat requests with agentic service

From cancellations to group bookings, Agentforce takes action instantly and routes complex cases with full context.

The results

50 %

chat inquiries resolved by Agentforce

15 %

decrease in customer support handle time

16 %

increase in CSAT with on-demand service

Increasing requests strained supported

Engine is an online travel platform where businesses and groups can plan anything from a quick trip for one, to travel for teams of 10,000 — supported by expert, high-touch service. With sustained growth, the company is focused on scaling operations without sacrificing service quality or driving up costs.

Engine’s client services team now handles more than 800,000 requests each year. While Engine had a chatbot that could answer basic questions, it couldn’t take action — so every request, for example “cancel my reservation,” required a rep to review details, search across multiple systems, and complete the update manually. This put pressure on the team as volume grew.

Agentic service in seconds

To support their growing customer base and maintain service quality, Engine introduced agentic self-service powered by Agentforce. Their AI agent Eva — the Engine Virtual Assistant — now delivers end-to-end support for hotel, flight, and car rental bookings across chat and voice and gives reps real-time support, unlocking personalized service that scales.

Whether a traveler needs to reschedule a flight or an admin needs to cancel a few rooms from a larger reservation, Eva takes action — retrieving booking details, confirming authorization, and executing the changes automatically.

“We have an extremely high bar for service, which includes the speed and quality of our customer interactions,” said Demetri Salvaggio, Vice President of Customer Experience and Operations. “Agentforce provides a conversational experience and level of personalization no chatbot can.”

When plans get more complicated — like rebooking hotels for a large group after a sudden change — Eva brings in a service rep and creates a case in Agentforce Service with full context, including the conversation history, customer profile, prior bookings, and recommended next steps.

Real-time context for every agent

Data 360 brings Engine’s data together into a single system of context used across the business, connecting customer profiles, bookings, interactions, and knowledge in one place. Engine’s agentic service runs on this foundation, and enables Agentforce to retrieve information, make decisions, and take action in real time.

Customer data from Agentforce Sales and interaction history and knowledge from Agentforce Service connect directly to Data 360, while more than 10 million rows of hotel data from Snowflake and Amazon S3 flow in via zero copy integration. Live engagement signals from Google Analytics and brand guidelines from Confluence are also layered in via a prebuilt connector, so every Agentforce interaction reflects current intent and company standards — without custom model training.

When a customer starts a chat or call with Eva — or a rep engages in CRM — Agentforce uses retrieval-augmented generation to pull the most relevant data.

The image is an architecture diagram of how Engine’s AI agent “Eva” autonomously cancels reservations. It starts with a customer contacting Agentforce, Eva then taking action based on the topic, and then working within Data 360, Service, and Engine’s booking Platform to send a cancellation confirmation.

First Data 360 retrieves the right context, then Amazon Bedrock powers response generation with LLMs, improving accuracy and reducing hallucinations. Agentforce can determine who the customer is, what they’ve booked, and what they’re authorized to do before responding.

With that customer-specific understanding, Agentforce moves from insight to action. Eva identifies intent, calls APIs to Engine’s booking platform, and completes the requested action. Changes are synced across systems and logged automatically, creating a clear record of every interaction.

That visibility feeds directly into Agentforce Observability, where Engine tracks Eva’s performance through built-in dashboards that measure deflection, topic success rates, case creation by topic, user behavior patterns, and CSAT to continuously improve outcomes.

Faster service, higher CSAT

Eva now resolves 50% of incoming chat cases automatically, and with better quality. Chat CSAT increased by 16% — reflecting the helpful, quick, and consistent support Engine strives to provide. “We’re in a period of hypergrowth,” said Salvaggio. “Agentforce helps us scale efficiently while improving the customer experience.”

Eva also reduced the average handle time of customer support cases by 15%. Requests that once required multiple steps and handoffs now happen in a single interaction, allowing Engine’s team to keep pace with growth and maintain 24/7, high-touch support.

With routine requests handled agentically, reps can focus on higher-value cases. Agentforce not only contributes to an increase in employee productivity, but also gives reps more time for nuanced conversations when human support is needed most.

Why Engine chose Salesforce

First-to-market ISV partners

Engine partnered with NeuraFlash, part of Accenture, and Salesforce to launch Agentforce Voice in under three months. NeuraFlash brought deep voice AI expertise and, as a Summit-level partner with early access to new agentic capabilities, helped Engine move faster than would have been possible alone.

Salesforce Professional Services

Salesforce Professional Services worked side by side with Engine from initial deployment through ongoing optimization, creating an agent that can take real action. They connected AI to booking systems, enabling Eva to securely complete tasks like cancellations instead of just responding.

Built-in observability

Engine uses Agentforce Observability to monitor usage, performance, and feedback across all agents — including Eva. When performance dips, teams can quickly update agent logic, test changes, and deploy fixes the same day without waiting for a sprint cycle.