01 / 04
Embedded conversations
A website widget connects a visitor to a knowledge-based chatbot, with a defined handoff to a human when the conversation needs one. The same widget also runs on the Armela site.
AI & Automation
FlowMaticX
A SaaS platform for lead workflows, customer conversations and AI-assisted automation.
- Year
- 2025
- Category
- AI & Automation
- Role
- Founder project
- Live site
- flowmaticx.com
The situation
Lead workflows and AI conversations in one product.
Small sales teams often lose leads in the gaps between their tools. A chat on the website never reaches the CRM, an enriched lead sits in a spreadsheet, and a follow-up depends on somebody remembering. FlowMaticX keeps the whole sequence in one product.
What we built
The parts that do the work
02 / 04
Connected lead work
Research, enrichment and CRM stages sit on the same records, so a lead keeps its context as it moves between stages instead of being re-keyed.
03 / 04
Visual automation builder
Operators build workflows in a visual editor instead of code, so they can change a sequence without a deployment.
04 / 04
Agency-ready operation
Multi-tenant accounts, white-label branding and Stripe billing let FlowMaticX run as a subscription product on one shared install.
Next.jsTypeScriptPrismaMySQLStripen8nNext.jsTypeScriptPrismaMySQLStripen8nNext.jsTypeScriptPrismaMySQLStripen8n
How it came together
From the brief to what changed after
- 01
Approach
The platform combines an embedded knowledge-based chatbot, lead enrichment, CRM workflows and a visual automation builder. It also has agency accounts and billing.
- 01Website visitor
- 02Knowledge-based chat
- 03Human handoff
- 04CRM workflow
- 02
The stack, and why
The application runs on Next.js and TypeScript. Prisma and MySQL hold the lead and workflow records, Stripe handles subscription billing, and n8n runs the automation layer. A relational store fits because lead, account and workflow-run records are heavily related, and a document store would have pushed that joining work into application code.
- 03
After launch
Because the automation layer is visual, day-to-day changes are an operator task rather than a development ticket. The multi-tenant model means onboarding another agency account does not require new infrastructure.
A working SaaS product owned by the founder. Its embedded chatbot is also used in the Armela project.
Tell us what you're building, get a fixed scope
A free 15-minute call, then a written scope and estimate. No obligation, no sales script.