Think about a sales manager at a phone company who sells call routes to other carriers around the world. Every day the work comes in from everywhere: an email from one partner, a WhatsApp message from another, a Telegram chat, a Teams thread, and then the numbers have to be copied into the billing system by hand. Oleksii Vinogradov, who goes by Alex, says a manager working like that tops out at a handful of accounts.
Alex has been building software for that world since 1999. The newest company, DealoAgent.ai, lets one manager handle many times more.
Eleven companies in thirty years
Alex is a serial founder and a full-stack developer, working from just outside Kyiv, Ukraine. Over 30 years Alex has started 11 companies, in telecommunications, mobile advertising, digital health and, for the last five years, AI. In that time Alex also earned a series of certifications as a data scientist.
The first company came in 1999, in telecom. IXC is still running. It makes softswitch and billing software, the systems that route and charge for phone calls, for wholesale and retail voice companies, and its website lists 272 clients in the wholesale business around the world.
Then there is dexStudios, a studio for Web3 and crypto games. It has been mostly on hold for about two years, since that market went quiet. But it taught Alex something. When the studio started, development cost a couple of tens of thousands of dollars a month in salaries. Over about a year and a half, Alex moved the work to AI step by step, and the cost fell to a couple of thousand dollars a month.
A CRM built on AI from the start
DealoAgent.ai is where Alex's attention is now. It is a CRM, the software sales teams use to track customers and deals, built around AI from the beginning. It pulls email, WhatsApp, Telegram, Microsoft Teams and other messengers into one place, alongside the usual CRM records, and lets AI agents work across all of it.
That matters because business no longer happens in one channel. A supplier emails, then follows up on WhatsApp, then says something important on Telegram, and the gaps between them are where deals get lost. DealoAgent reads every conversation, so it can tell you that an invoice mentioned on WhatsApp was never sent, or prepare a contract from what was agreed across three channels.
Alex's favourite example is a Telegram folder. Many people sort their Telegram chats into folders, such as one called Clients. Alex can tell DealoAgent, in plain words: read every chat in the Clients folder, look at the last ten messages in each, and draft a follow-up for all of them. The agent did 75 chats. Each draft reads like a person wrote it: "Do you remember we talked about this two weeks ago? Let's move forward. Are you expecting an email from me?" If the client answers yes, Alex tells it to prepare the email, and it finds the address, writes the text and waits for a confirmation before sending.
You can work entirely through agents, sending them tasks in parallel, or switch to a normal CRM screen and work with leads and customers by hand. The agents can also research leads deeply on the web, dozens at a time, so each message is as personal as if someone had done the homework.
Alex's case for it over the big names is the architecture. DealoAgent is not an AI gateway bolted onto an existing CRM. It was designed around AI, so the system learns the business, remembers what it collects, and stays cheap to run. "Keep it cheap enough," Alex says, "which is also very important."
The market is crowded. Salesforce and HubSpot are adding AI to the CRMs most companies already use. Attio, folk and Day AI are building newer CRMs with AI at the centre. Kommo is built for selling through messengers, and Clay is a favourite for researching leads. What DealoAgent adds is depth in one industry: skills already tested in telecom, and a direct link into the billing and switching systems that phone companies run on.
Telecom first
Telecom is the first market because Alex knows it from the inside. In wholesale telecom, managers are expensive, and companies are under pressure to cut them. When Alex shows them an AI system that lets one manager serve ten times more partners, even a hundred times more, the answer is often: this is what we have been waiting for.
With IXC, the AI can act on the switch itself. A manager can ask, in their own language, to add destinations for certain carriers or change how calls are routed to one country, and the changes go through in a batch, without clicking through screen after screen.
Deals in telecom are slow, because many departments must approve them. DealoAgent is now in integration with mobile carriers in Ukraine, including one of the country's largest, and in the Caribbean, Africa and South Asia. The names are under NDA. Alex expects to sign a big agreement next quarter, or in the second quarter of 2027. Most of these customers came through relationships built since 1999: Alex calls friends, and friends of friends say it sounds interesting. But Alex says cold outreach works too, "if you know what to say and how."
The second market is recruiting, and the website also shows uses for sales teams, venture capital research and call centres.
Four people and their agents
The whole team is four people. The software itself is written largely by agents Alex built. "I compare with like hundreds of software developers behind me," Alex says, "but it's machines." So DealoAgent is not only AI-native in what it does for customers. It is AI-native in how it is made.
Alex funds the company and is not raising money. Investors who can talk business at Alex's level are welcome, but the plan does not depend on them.
What they are looking for
Customers and partners. Telecom carriers and voice and SMS companies whose managers drown in messages and spreadsheets. Sales and recruiting teams who live in email, WhatsApp and Telegram. And partners who want to bring DealoAgent to their own clients through its partner programme. Try it, or get in touch, at dealoagent.ai.
Why it is bigger than it sounds
AI did not just change what Alex builds. It changed how Alex builds: four people and their own AI agents now do work that once took a big team.
This breed of founder has done it the old way many times, and is now doing it the new way. Alex has started 11 companies over 30 years, with the payrolls that came with them. At dexStudios, moving development to AI cut the cost from tens of thousands of dollars a month to a few thousand. DealoAgent is built by four people and agents Alex wrote, and it sells the same idea to its customers: one telecom manager who used to handle a handful of accounts can handle ten or a hundred times more. When building costs that little, a founder does not need to raise a big round to start, and can sell to an industry where Alex has had relationships since 1999.
The formula: an experienced founder plus their own agents can now do what used to take a funded team, and the best first market is the one the founder already knows by name. The same play fits a veteran freight broker building AI for logistics, a former hospital manager building it for clinics, an accountant building it for small firms, or a hotel operator building camera checks for housekeeping and security.
The risk in one line: the big CRMs are adding AI fast, and a four-person team has to keep its industry depth ahead of their reach. If you have spent decades in one industry, build the AI for it yourself before someone who has not does.
What to watch
The first thing is the telecom agreement Alex expects in the next two quarters. One large carrier running on DealoAgent would prove the model in the industry that is hardest to sell to.
The second is recruiting. If the same agents work for hiring teams, DealoAgent stops being a telecom tool and becomes a platform.
The third is the team. If four people can keep shipping at this pace, it will be a working example of how small a software company can now be.