-
-
Digital Agent can mimic human employee and work 24/7 using AI
-
Shows Error to configure AI for your ORG
-
Select Brain of your CRM
-
-
Dynamic UI in Studio
-
Insights from Vector embedings
-
Traditional CRM data imported from salesforce
-
its a canvas where dynamic UI items fly in when asked by UI (in the roadmap)
-
Create a Agent (in the roadmap)
-
Target Architecture and roadmap
-
Asbaat.com Landing page
Inspiration
We kept asking a simple question: what if the CRM built itself around the user instead of forcing the user to build around it? That's the spark behind CXOS, an Agentic Customer Experience Operating System that takes everything a Salesforce org (a starting point, taken as best practice) already knows (its metadata, objects, and relationships) and turns it into an AI-native workspace anyone can actually use.
What it does
AI-native workspace that sits on top of your CRM data. AI assembles personalized, task-focused workspaces tailored to each user. Natural language search, so you can ask questions in plain English and get answers grounded in your actual records. A personalized daily dashboard that surfaces the records and next actions that matter today. Multi-tenant, securely isolated workspaces with enterprise-grade access controls, built so each organization's data stays its own.
How we built it
The system flows through a clear pipeline — Signup for Vercel → start writing prompts→ AI agents → dynamic UI generator → manual testing
Frontend and app layer: a Next.js application deployed on Vercel, designed for both light and dark modes and a fast, modern feel.
Challenges we ran into
One of the biggest challenges was generating a complex business application using GenAI. Current UI generation tools like V0 are still not mature enough to build enterprise-grade applications end-to-end. Claude proved to be a valuable coding assistant, but significant manual engineering and architectural decisions were still required to deliver a production-ready solution. Metadata is messy. Every Salesforce org is a snowflake of custom objects, fields, and relationships. Building a retrieval and modeling layer that could make sense of arbitrary org structures, without hardcoding assumptions, was the hardest and most important problem.
Accomplishments that we're proud of
We shipped a working, multi-tenant SaaS that authenticates with Salesforce and retrieves real org metadata, not a mockup. one can migrate their salesforce
What we learned
Vercel and Amazon is probably the best tech stack out there for individuals to deliver production grade highly scaleable apps
What's next for CXOS – Agentic Customer Experience Operating System
Longer term (planned): autonomous AI agents that take action on the user's behalf, workflow automation, predictive insights, and natural language analytics.
Built With
- amazon-web-services
- anthropic
- nextjs
- openai
- postgresql
Log in or sign up for Devpost to join the conversation.