In today's AI-powered world, no-code tools have made it easier than ever to automate tasks, but most of them still rely on rigid, pre-scripted workflows.
Omnisales, a new startup founded by Sezer Kemer and Michael Nefedov, is working on the next leap forward: autonomous software programs, or intelligent agents, that can understand user goals and work with the right tools to execute them across multiple platforms.
Omnisales is building the infrastructure to make these programs accessible to everyone. Instead of requiring technical expertise or a complex setup, the platform lets users launch AI Agents directly from their browser which can connect to tools like CRMs, email platforms, and databases, then act independently to complete tasks, follow up, and learn from outcomes over time.
The result is a new layer of automation that can think for itself, understand context, and carry out complex tasks just like a skilled co-worker. Kemer and Nefedov see these tools as the missing piece in today's business world.
Here's a closer look.
The tech behind the pitch: turning automation workflows into autonomous agents
Most no-code automation tools rely on static workflows in which users define a series of steps and the system executes them exactly as written, but Omnisales takes a different approach by letting users create intelligent agents.
At the core of Omnisales is model context protocol (MCP). MCPs are shared standards designed to integrate AI models with external systems like CRMs, email platforms, spreadsheets, APIs, and databases, which helps agents access and act on business data across these systems rather than being limited to a single database. When MCP was first introduced in 2024, it was optimized for desktop applications, limiting its use to data on single machines with no memory of past prompts.
With Omnisales, Kemer and Nefedov built a system that can deploy in-browser, helping users launch and manage AI agents that have broader capabilities without needing any installation.
These agents can connect to and act across a wide range of tools. For example, users can prompt an agent to follow up with a sales lead, make outbound phone calls, update records in a CRM, and send personalized emails based on recent interactions. The agent then determines which language model to use, interprets the data it encounters, and takes the necessary steps across the CRM and email marketing system to achieve the goal.
The agent remembers previous prompts, analyses, and workflows to learn and improve over time because the data is not stored locally.
"This basically means that we can give the power of agents that leverage MCP to any consumer or non-technical person," Nefedov explains.
Kemer and Nefedov: a team built on speed and contrast
Kemer and Nefedov didn't know each other a week before they founded Omnisales together, but they were aligned from day one. Kemer brought an instinct for product and go-to-market strategy which perfectly complemented Nefedov's deep technical expertise in systems engineering and AI infrastructure.
Their differences are what makes the partnership work so well. Sezer is self-taught in sales, GTM, and startup ops, while Nefedov has extensive experience in reinforcement learning, NLP, and systems engineering and has worked at Amazon, Salesforce, and Synoptic.
Nefedov is the engineer behind the scenes who builds the backend and model integration, and Kemer is in charge of the user experience and identifying potential applications for the technology. Together, they bring speed and depth to every decision, as well as a track record of building and shipping quickly.
Beyond the pitch deck: the actual value of Omnisales
Omnisales isn't just a wrapper around an AI algorithm, it's a full-fledged platform that lets non-technical users build intelligent agents directly from their browser to automate complex, multi-step tasks across tools like email, CRMs, and spreadsheets, all without writing a single line of code.
Omnisales is designed to let any user build context-aware AI agents that understand the situation in which they operate, not just the task at hand. These agents can deploy branching logic, meaning they can make decisions based on different conditions, and they use persistent memory to retain information across sessions. With LLM routing, they can choose the best language model for each task.
This is what makes Omnisales stand out next to traditional automation tools, which require users to define every step in advance. Omnisales agents can handle ambiguity and evolve over time, executing user natural language requests by interpreting intent, reviewing the execution and success of past workflows, and hooking into all necessary external systems to reach the goal.
While the founders eventually see Omnisales working across industries, their first target is responsible and ethical debt collection. Its agents can already identify unresponsive leads, qualify leads, send personalized follow-ups, escalate high-priority responses, and update CRM records without human intervention. They adapt based on how prospects respond, learning which messages work best and adjusting their behavior accordingly.
Think chatbot combined with task automation run by a human agent providing quality control in the background, all handled directly from your browser.
The bet behind the build: why the timing is right for Omnisales
Although AI-native infrastructure is in high demand, few current tools are built specifically for autonomous agents that can reason, adapt, and operate across systems. That's what sets Omnisales apart.
The rise of standards like MCP has made it possible to connect language models with other business software in a structured way, and Omnisales is already rolling out early access agents designed to handle lead follow-ups, qualification, and escalation. The system's modular backend and inherent learning ability also makes other use cases possible in the future.
For Kemer and Nefedov, Omnisales is more than simply another AI tool, it's a fundamental shift in how automation gets built and used. By making intelligent agents accessible through web browsers, they're turning what used to require engineering teams into something anyone can deploy in pursuit of their ultimate goal: improving how businesses automate and changing who gets to do it.
VentureBeat newsroom and editorial staff were not involved in the creation of this content.