Bombardier gives every seller an on-demand engineer with Agentforce

Complex questions, precise answers

Bombardier builds sophisticated business aircraft for discerning buyers who expect premium sales experiences. During the sales process, those buyers want to know exactly what an aircraft can do — from how many passengers it can carry to how far it can fly under specific conditions.

Answering these questions requires complex calculations that account for weather, airport-specific runway conditions, and wind patterns. Getting those answers could take days of back-and-forth between sales directors and a small, highly specialized sales engineering team. Global time differences added to the wait, slowing deal momentum and making real-time technical guidance difficult.

Bombardier knew that making these complex questions faster and easier to answer was key to meeting buyer expectations and staying competitive.

A virtual engineer for every seller

To bring that speed and precision directly into the sales conversation, Bombardier built Clara, an agentic sales engineer powered by Agentforce. Sellers can ask Clara technical questions and run aircraft performance analyses from their phone, tablet, or computer.

For example, imagine a prospect wants to know whether a Global 7500 can fly nonstop from San Francisco to Tokyo tonight with five passengers. The seller asks Clara, which validates the request and calculates range based on factors like passenger load, aircraft weight, runway conditions, seasonal winds, and atmospheric conditions. Clara documents its assumptions and generates an interactive 360-degree range map showing how far the aircraft can fly.

If the prospect wants to increase the number of passengers or leave from a different airport, Clara recalculates each scenario and updates the visualization.

“When it comes to the computations we've built into this agent, sellers can now generate results on the fly, within seconds,” said Fauve Castagna, Manager, Sales AI & Enablement. “They get rich visualizations that they can share live with customers directly from their screen. If the customer wants to explore different scenarios, they can do it in real time.”

Clara can also answer aircraft specification questions and compare Bombardier and competitor models, with citations to approved technical sources. Sellers can then ask Clara to turn the analysis into a customer-ready email personalized with CRM data.

From question to calculation

To interact with Clara, sellers start a chat in Sales Cloud, Bombardier’s CRM. Clara then determines what the seller needs and routes the request down the right technical path. 

For specification questions and competitor comparisons, more than 500 technical documents from Salesforce and SharePoint are indexed through the Agentforce Data Library and Data 360. Files are chunked and vectorized, stored in a hybrid search index, and queried using semantic and keyword search. Clara uses retrieval-augmented generation to find relevant passages across multiple sources and synthesize them into a cited response.

Databricks adds operational, IoT, and historical aircraft data through a zero-copy connection with Data 360, giving Clara richer context without duplicating the underlying data.

Range calculations trigger a more complex workflow. Clara turns the seller’s request into structured inputs, then uses aircraft, airport, and performance data from within Salesforce to validate speed, payload, and weight. An Apex orchestrator calls Aircraft Performance Group’s API for runway restrictions and maximum takeoff weight, then Bombardier’s API models seasonal winds and reliability to calculate a 360-degree range polygon.

The results return to Salesforce, where a Platform Event triggers a Lightning Web Component to render the analysis on Azure Maps. Each calculation, assumption, and iteration is saved, so sellers can refine scenarios without losing the history behind the answer.

Keeping deals moving

Clara cuts a range-map calculation from at least four hours of engineering work to 10–20 seconds. International sellers who previously had to wait up to four days for technical answers and iterations across time zones can now generate contextualized results and calculations in a fraction of the time.

That speed lets sellers test scenarios with customers in real time instead of pausing the conversation for engineering follow-up. With a 97% accuracy rate, Clara also gives sellers validated technical information that helps them respond confidently without overcommitting on aircraft capabilities.

Bombardier deployed Clara in 15 weeks, and the AI agent earned a 5/5 user satisfaction score at launch. They are now scaling Clara to more than 300 sales team members worldwide.

“At this point, fast response times are table stakes,” said Castagna. “We need to meet the standard customers have come to expect from all of their other digital experiences. Right now, these capabilities help differentiate us but over time, I think the real risk is not adopting them. This is the kind of advancement that may set you apart today, but if you're not doing it tomorrow, you're likely to fall behind.”