AutoSalesPilot.ai — AI Sales CRM and Co-Pilot
An AI-powered sales CRM with a daily AI Sales Brief, 0–100 lead scoring, deal health, forecasting and human-approved follow-ups, and lead capture from web forms, Meta Lead Ads and email.
Designed and developed by Canny Technologies AI Labs for our client.
Client and industry
- Project type
- Client Project
- Service
- AI CRM Development & Business Automation
- Industry
- Sales SaaS / CRM
- Client
- AutoSalesPilot.ai
- Region
- Global
- Developer
- Canny Technologies AI Labs
The problem
Most CRMs are filing cabinets: they store contacts, deals and activity, but leave the rep to work out each morning which lead to call, which deal is slipping and what to say in the follow-up. Leads arrive from website forms, Meta Lead Ads, inbound email and spreadsheets, so they get duplicated or lost and nobody knows which source actually produces revenue. Forecasts are guesses typed into a spreadsheet, follow-ups depend on memory, and the AI tools that promise to help either sit outside the CRM or send messages to customers without anyone checking them first.
What the client needed
- A CRM that tells reps what to do next, not one that only stores records
- Leads from web forms, Meta Lead Ads, email and spreadsheets in one place, de-duplicated and attributed to their source
- A forecast that separates committed revenue from hope
- AI help inside the CRM, with nothing sent to a customer without human approval
- Multi-tenant SaaS with a self-serve trial and subscription billing
What we built
We built AutoSalesPilot.ai as a CRM that helps close deals, not just store them. Alongside contacts, companies, deals, quotes, tasks and calendar, an AI layer produces a daily Sales Brief ranking actions by revenue, urgency, deal risk and buying intent; scores every lead from 0 to 100 on stage, recency, engagement and intent; puts a health score and reason on every deal in the Kanban pipeline; and splits the monthly forecast into committed, best case, low confidence and at risk. An AI assistant answers questions about the workspace and links straight to the records involved. Follow-ups are drafted with awareness of the customer's last reply, and WHEN → IF → THEN automations run the routine work — but every external message is drafted for human approval by default. Leads flow in from embeddable website forms with spam protection, Meta Lead Ads, inbound email and CSV import with de-duplication and source attribution, and teams get line-item quotes with a customer portal and view tracking, reports, an executive dashboard, and a versioned REST API with webhooks.
Features
- Daily AI Sales Brief ranking actions by revenue, urgency, deal risk and intent
- 0–100 AI lead scoring with the reasons behind each score
- Kanban pipeline with an AI health score on every deal
- Four-bucket forecast: committed, best case, low confidence, at risk
- AI workspace assistant that links to the records involved
- Reply-aware follow-up drafts, approved by a person before sending
- WHEN → IF → THEN workflow automations
- Line-item quotes with a customer portal and view tracking
- Reports and an executive dashboard
- Versioned /v1 REST API with webhooks
Architecture
- 1Vite + React single-page application for the CRM workspace and customer-facing quote portal
- 2Node.js + Express API with Sequelize over PostgreSQL — 66 tables across workspaces, contacts, companies, deals, quotes, tasks, automations and audit logs
- 3AI service built on Anthropic Claude for the Sales Brief, lead scoring reasons, deal health, workspace assistant and reply-aware follow-up drafts
- 4pg-boss background job queue for automations, lead ingestion and scheduled scoring
- 5Socket.IO real-time updates so pipeline and task changes appear instantly for the whole team
- 6Lead ingestion from website forms, Meta Lead Ads, inbound email and CSV/Excel import with de-duplication and source attribution
- 7Gmail, Google Calendar and Microsoft Outlook integrations with credentials encrypted per workspace
- 8Workspace isolation, owner/admin/rep roles, audit logging and TOTP two-factor authentication
- 9Versioned /v1 REST API with per-workspace keys and outbound webhooks; Razorpay subscription billing
Integrations
- Anthropic Claude
- Meta Lead Ads
- Gmail
- Google Calendar
- Microsoft Outlook
- Razorpay
- Website lead forms
- Inbound email
- CSV / Excel import
- Outbound webhooks
Screenshots
Tech stack
- React
- Vite
- Node.js
- Express
- PostgreSQL
- Sequelize
- Anthropic Claude
- pg-boss
- Socket.IO
- Meta Lead Ads
- Gmail / Outlook
- Razorpay
Timeline
Duration / status: Live at autosalespilot.ai
Phase 1
Standalone CRM core
- Standalone backend, web app and database built for multi-tenant SaaS
- Contacts, companies, deals, Kanban pipeline, tasks and calendar
- Workspaces, roles, JWT auth and two-factor authentication
Phase 2
AI co-pilot
- Daily AI Sales Brief, 0–100 lead scoring and deal health scores
- Probability-weighted forecast and in-app AI assistant
- Reply-aware follow-up drafts with human approval by default
Phase 3
Integrations and launch
- Website forms, Meta Lead Ads, inbound email and CSV lead ingestion
- Gmail, Outlook, quotes portal, /v1 REST API and webhooks
- Reports, executive dashboard, pricing and self-serve trial
Results
0–100
AI lead score on every contact
4
Forecast buckets (committed → at risk)
12
Industries with ready-made workflows
66
Database tables on one schema
0
External messages sent without human approval (default)
14 days
Free trial with Professional access
AutoSalesPilot.ai is live as a self-serve SaaS with published per-user pricing — Starter at $15 and Professional at $39 per user per month — and a 14-day free trial that needs no card. Reps start the day from a ranked list of what to do next instead of an inbox, managers see a forecast that separates committed revenue from hope, and every lead can be traced back to the form, ad or email that produced it.
Live link
Related Services
Discuss a Similar Project
Planning something like this? Talk to the team that built it about scope, timeline and a fixed-price estimate.
More Client Projects
All case studies- Client Project
AuraOne — One Platform for Work and Life
A work-and-life SaaS platform on web, iOS and Android, built on one API, with two-way Google and Outlook calendar sync, team workspaces and an AI auto-scheduler.
- Client Project
AI Passport Photo Processing Platform
A computer-vision pipeline that checks and corrects passport photos against the standards of 50+ countries and produces print-ready output.
- Client Project
Property OS — Real Estate Asset, Lease & Compliance Platform
One audited register for a commercial property portfolio: assets and ownership, leases and rent, statutory compliance, documents and deals.
