AgentResQ

AgentResQ is an AI-powered assistant that automates crisis response—engaging victims, predicting aid demand, creating records, assigning volunteers, and slack updates—all in real time.

Comment 8

🌍 Inspiration

In times of crisis, victims need immediate support — but most aid systems still rely on manual processes, disconnected teams, and delayed responses. We set out to change that.

AgentResQ was born from the idea that AI can do more than just automate — it can respond, predict, and coordinate relief operations with empathy and precision.

🤖 What We Built

AgentResQ is a real-time, AI-powered crisis response assistant, built entirely on Salesforce’s Agentforce platform. Through a single conversation, it transforms victim outreach into a fully automated, intelligent aid pipeline.

Here’s what AgentResQ does:

  • 🌐 Detects the victim’s location via a custom dummy IP-based API
  • 💬 Engages victims conversationally to collect personal and family details
  • ⚙️ Triggers Salesforce Flows to create Contacts, Cases, Aid Requests, and Line Items
  • ✍️ Uses Einstein Generative Prompt to generate empathetic case subjects and fundraising content
  • 📍 Assigns volunteers based on regional routing logic
  • 📈 Predicts future aid demand using a Einstein Predictive GLM model trained on 100K+ historical records[Agent SafeHeaven]
  • 🔔 Sends real-time fundraising and coordination alerts to Slack

Built With

  • agent
  • and-agent-actions-?-ensuring-every-aid-request-is-routed
  • and-severity.-output-is-displayed-in-real-time-through-a-lightning-web-component.-**automation-&-engagement**-behind-every-chat-is-a-powerful-orchestration-layer-of-flows
  • apex
  • api
  • built-with-einstein-studio
  • chat
  • combining-**generative**-and-**predictive**-intelligence-with-salesforce-automation.-**generative-ai**-using-einstein-prompt-builder
  • einstein
  • einsteinpredective
  • flows
  • forecasts-aid-volume-based-on-crisis-type
  • glm
  • gpt
  • lwc
  • model
  • pomptbuilder
  • recorded
  • region
  • slack
  • we-crafted-dynamic-prompts-to-generate-human-like-case-descriptions-and-fundraising-posts-that-reflect-urgency-and-compassion.-**predictive-ai**-our-custom-model
+ 16 more
Share this project:

Updates

posted an update —

Participating in the Salesforce hackathon allowed us to learn about Agentforce and its AI capabilities and explore concepts like agent chain reactions, out-of-the-box topics/action sets, and possibilities to create custom ones. We also delved into the potential of custom knowledge sets and libraries. During the event, we discovered the decisive actions performed by agents, the generation of contextual prompts, and the seamless integration of data from diverse sources in the cloud. We unlocked new possibilities through data consolidation, summarization, analysis, and predictive modeling.

In addition to using Agentforce in a standalone environment, we found it beneficial to learn how to implement it in a multi-channel integration, allowing for effective communication across social channels, Slack, and multiple organizations.

The project AgentResQ stands as an inspiring testament to what can be achieved when we prioritize end-users' needs. It showcases a cutting-edge automated crisis response system that harnesses the power of Salesforce CRM, innovative supply chain management, data cloud predictive analysis, and interconnected social and communication channels.

Log in or sign up for Devpost to join the conversation.