
AI SaaS Developer — Real-Time Sales Performance Platform
- or -
Post a project like this- Posted:
- Proposals: 90
- Remote
- #4501812
- OPPORTUNITY
- Paused







Description
The stack is fixed:
Lovable · Supabase · Next.js 15 · Cursor
What the build requires:
• Multi-tenant architecture with Supabase Row Level Security (RLS)
• Real-time data pipelines and live UI updates via Supabase subscriptions
• Webhook ingestion from external CRMs (HubSpot, Salesforce, Pipedrive etc.)
• AI agent integration using Claude API — conversational agents that interact live with sales data, surface insights, and respond to rep and director queries in real time
• Clean, scalable schema design before any front-end work begins
This is not a design job or a CMS project. We need an AI-native developer who builds SaaS products using modern tooling and understands multi-tenancy at architecture level.
Budget: Fixed price, milestone-based. Serious applicants only.
To apply, answer these two questions:
Daniel A.
100% (61)New Proposal
Login to your account and send a proposal now to get this project.
Log inClarification Board Ask a Question
-

-/ For the AI coaching agents, what depth of conversational interaction is expected? Should they provide prescriptive guidance, alerts on missed opportunities, or contextual explanations based on historical performance?
-/ How granular should the multi-tenant data isolation be? Will there be shared datasets for benchmarking, or strictly siloed data per tenant?
-/ Regarding webhook integration with external CRMs, do you anticipate one-way ingestion only, or should the platform also push insights or updates back into the CRM? -

Hi Daniel, for the initial CRM integrations, should we treat HubSpot, Salesforce and Pipedrive as equal first-release requirements, or is there one CRM you want prioritised first so the webhook model can be built and tested around the most important data path?
-

Have you already defined your tenant structure, user roles, and CRM data mapping requirements, or would you like assistance designing the multi-tenant architecture and Supabase RLS strategy?
-

A few questions:
1. Will each customer have a separate workspace/organization model, or are there more complex hierarchy requirements?
2. Are the AI coaching agents expected to operate entirely through Claude, or is retrieval/context storage also planned?
3. Will CRM integrations be webhook-only, or will periodic synchronization also be required?
4. Has the initial database schema already been designed, or is that part of the first milestone?
-

A few questions:
1. Will commission calculations follow a fixed rule or different rules for each company?
2. Do you already have AI coaching workflows defined, or should we help design them?
3. Which CRM integration would you like to launch with first? HubSpot, Salesforce, or Pipedrive? -

A few questions before finalising scope:
• Have you already defined the commission calculation rules, or will those need to be designed as part of the project?
• Will each tenant connect to a single CRM, or do you expect some organisations to connect multiple CRM sources simultaneously?
• Are AI agents expected to operate purely through chat, or should they proactively generate alerts, coaching recommendations, and performance summaries?
• Do you already have a Supabase schema drafted, or is the database architecture being designed from scratch?
• What is the expected tenant and user volume for the first 12 months?
• Will the AI agents require historical trend analysis and forecasting, or primarily real-time insights from incoming sales activity?
• Is there an existing UI design system, or is the frontend implementation expected to follow designs generated through Lovable?
-

Daniel, what is your expected scale in terms of tenants, daily webhook events, and active users, so the architecture can be designed correctly for performance and real-time updates from day one?
