How We Won "Best Use of Data Cloud" & "Business Impact" at the
Dreamforce Hackathon
From Sales Admin to Revenue Engine: How Team Fleurby used Agentforce to build the "Digital Air Traffic Controller" for Revenue Operations.
Key Outcomes
Best Use of Data Cloud
Award for Business Impact
Automated CRM Data Entry
Sentiment-Based Marketing Suppression
Technology Stack
The Challenge: The "Post-Call Void"
Most AI tools solve the easy part: transcription. But for Revenue Operations, the real pain begins after the call. We identified two "silent killers" of revenue that occur in this void.
Sales Administration
The Time Sink: Reps were wasting hours after every call manually typing notes, logging tasks, and drafting follow-ups. This "admin tax" kept them from doing what they do best: selling.
The Marketing Disconnect
The Churn Risk: Marketing teams were flying blind, sending generic "nurture" emails to customers who were actually frustrated or at-risk. This deaf approach wasted spend and annoyed customers.
The Solution: Fleurby
We built Fleurby, an intelligent agent living entirely within Slack. We described it to the judges as a "Digital Air Traffic Controller"—listening to the radio (transcripts) in real-time and directing the traffic (Salesforce data) to prevent collisions.
The Command Center (Slack + Agentforce)
Instead of forcing reps to toggle screens, Slack acts as the UI for the entire process.
Pre-Call Briefing
When a Salesforce Opportunity opens, Slack automatically spins up a dedicated deal channel. An Agentforce agent generates a briefing using chained prompt templates—pulling deal context, past sentiment, and open tasks instantly.
Post-Call Ingestion
The meeting transcript is ingested via a custom Apex webhook, creating a "Meeting Insight" record in Salesforce automatically. This turns the unstructured conversation into actionable data instantly.
The Brain: Data Cloud
This specific architecture secured the technical award. We used Data Cloud to bridge the gap between "Hearing" and "Acting."
Sentiment Extraction
Our agent analyzes the transcript for customer sentiment (Positive, Neutral, Negative).
Unified Profiles
This sentiment data is immediately mapped to the Contact’s Unified Profile in Data Cloud via Data Graphs.
Real-Time Trigger
If a client mentions a security concern, the system flags the sentiment as "Negative." Data Cloud instantly moves them into a suppression segment in Marketing Cloud Growth—stopping generic emails and triggering a high-priority alert to Sales.
The Action: Closing the Loop
Fleurby closes the loop without manual effort, directly impacting the bottom line.
Task Automation
Using a "next action suggestion" prompt, the agent identifies promises made during the call (e.g., "I'll send you the whitepaper"). Once the rep gives the "green light," the agent instantly creates correctly assigned Salesforce Tasks.
Drafts on Demand
A "Generate Draft Email" prompt reads the meeting insight and drafts a context-aware follow-up email, saved directly to a custom Salesforce object for final review.
The Team Behind Fleurby
Winning a hackathon requires more than just code; it requires a cross-functional blend of strategy, architecture, and creativity. We brought together experts across the Salesforce ecosystem.
Lead Architect & Developer
Focused on the heavy lifting in Apex and the custom integrations between Slack and Salesforce. Built the webhook architecture that allowed real-time transcript ingestion.
Data Cloud Specialist
The architect behind 'The Brain.' Designed the Data Graphs and Unified Profiles that allowed us to translate raw sentiment into marketing segmentation triggers.
Prompt Engineer & Strategy
Crafted the Agentforce prompt templates to ensure the AI didn't just summarize, but actually extracted actionable business insights and tasks.
Why These Awards Matter
Winning Best Use of Data Cloud proved that robust architecture can turn unstructured conversational data into structured business signals in milliseconds.
Winning Business Impact validated that we weren't just building a toy; we were moving the organization from Reactive (fixing a relationship after churn) to Proactive (adjusting strategy the moment sentiment shifts).