Salesdep.ai

Sales teams lose deals when a lead from a web form waits hours for a call back. This Kyiv/UK startup set out to automate that first contact: an AI agent reaches the lead within seconds by voice or text, holds a natural conversation, qualifies the prospect and books a meeting with a rep. The product had to serve many client companies at once, bill by actual usage and grow from prototype to paying customers.

As CTO I owned the technical strategy and the whole delivery cycle: service architecture, Node.js/TypeScript backends, Vue/Nuxt frontends and integration with OpenAI and other LLM providers. The platform takes leads from web forms and chats, runs voice and text agents, attaches an AI summary to every conversation and passes results to the CRM. Infrastructure lives on AWS — VPC, EKS, RDS, ECR — described in Terraform modules, with GCP and Cloudflare alongside. I also ran hiring, code review and releases.

Voice and text AI agents with configurable qualification scripts

Live voice demo in the landing hero, mobile-first widget, localised versions

Lead inbox: every conversation with status, source, channel and AI summary

Use cases: reactivation campaigns, instant follow-up, qualification, reminders

Pay-as-you-go billing per message, voice minute and video; plans and invoices

Took the platform from prototype to paying customers over 18 months as CTO

Node.js

TypeScript

Vue / Nuxt

PostgreSQL

Redis

OpenAI

AWS (EKS, RDS)

Terraform

Cloudflare