About

Mikhail Semenov

Product manager in B2B SaaS. Launched the voice AI agent line at Zvonobot (Prof-IT Group) from zero: 80 paying customers and 500,000+ minutes of live conversations in production in under a year. Pricing, unit economics and go-to-market are mine; I prototype with AI coding agents so hypotheses get tested in days. Remote, GMT+5, open to relocation.

01 · Now

Voice AI agents for B2B

I have been at Prof-IT since February 2022: I came in as a PM on its voice products, zvonobot and effebot, where robots call clients with pre-recorded scripts. Once LLMs learned to hold a real conversation, it became clear that recorded scripts would only get us so far. That is how the group’s new product came about in the fall of 2025, launched and now scaled by me: voice AI agents for B2B that talk to people live instead of playing a recording.

I own the product end to end, from the client console to billing. In under a year: 80 paying B2B customers, 500,000+ minutes of live conversations in production, and a sales team that grew from 4 to 15 to sell the line.

02 · Where I add the most value

AI embedded into real workflows

Products where AI does the actual business job, not just lives as a demo. Sweet spots:

AI that solves a real business job, not AI that lives as a demo
Products where AI removes busywork and speeds teams up
Voice and conversational AI scenarios (one of the domains)

03 · How I validate ideas

A hypothesis gets a working release in days

I build side projects with AI coding agents (Claude Code and Codex) so I can validate a hypothesis in days and speak the same language as engineers. In the day job, product decisions, pricing and go-to-market are mine, and the engineering team ships. Each side project on this site went from idea to a working release in 1-7 days.

04 · Team and people

No formal authority, one shared goal

I launched the line with a team of four engineers, none of whom report to me. What makes it work is a clear spec, an honest priority call and a review where I read every diff myself: engineers move faster when the why is visible.

The sales team had to be sold on the product before any customer was. We went to market on a white-label platform in weeks, proved demand with live customers, and on those numbers the sales team grew from 4 to 15 to carry the line.

The case for building our own platform instead of staying on white-label was made with per-call-type margins and repeat payments, not with a deck. The release cycle runs jointly with sales and support, so customer feedback reaches the backlog in days, not a quarter.

05 · What matters in a role

Ownership and a short idea-to-validation loop

Autonomy and ownership on ambiguous problems
A short loop from idea to validation
Product logic on top of AI, not AI for its own sake
Teams where I can carry a problem end-to-end, not one step in a long approval chain

06 · Open to

Product manager in B2B SaaS, payments and AI products

Open to product manager roles where I own the product end to end: B2B SaaS, payments and subscriptions, AI products. Voice is just one domain where I have already done this; I am interested in any product where AI solves a real business job rather than being a checkbox. Format: remote (I am in the GMT+5 zone), open to relocation. Russian native, English at working proficiency, comfortable in written and async communication.

07 · Where I’m a weaker fit

An honest anti-fit

So we don’t waste each other’s time, here is where I’m not the best pick:

Pure research ML / data science with no product layer on top
A single step in a long approval chain with no ownership over the problem
Products that add AI as a checkbox rather than for a real business job
Roles needing C-level live negotiation in English (written and async are comfortable)

Skills

Core skills

Grouped around what matters in a product manager role: product, AI/LLM, delivery, and a shared language with engineers.

Product

Product discovery
Product strategy
Roadmap
Metrics · unit economics
CustDev
B2B SaaS
Go-to-market

AI / LLM

Voice AI agents
Conversational AI
LLM orchestration
Prompt engineering
RAG
AI in the workflow (not demos)

Delivery

Spec → review → ship
Agent-native (Claude Code, Codex)
Lean Startup · 0→1
Fast hypothesis validation

Shared language with engineers

TypeScript
Python
SQL
REST
Telephony

Get in touch

Open to Product Manager roles in B2B SaaS, payments and AI products.

Tell me about the product and the problem. I will tell you honestly whether I am the right person for it, and where AI would actually move the needle.

Message me on Telegram

GitHub · last year 0 contributions github.com/RebSem