Every online shop has the same leak. People find a product, put it in the cart, get as far as checkout, and then stop. The shop paid to bring them there, through ads, search and email, and in the last few seconds they walk away. Most shops send an email and hope.
Callsy picks up the phone.
What they built
Callsy builds AI agents that call, text and message a shop's customers. The company is based in Vilnius, Lithuania. Its customers are mostly in the Americas, split roughly evenly between North and South America, including shops in Chile and across Latin America.
It began with a problem founder Arunas Vismantas lived through. Arunas spent years building and growing online brands, and noticed that most of the people who came to buy never finished. So the team tried something simple: call them, and ask what was wrong. Often it was a small thing. A missing detail, a question about delivery, a doubt nobody had answered. Just by having those conversations, they turned roughly 20 to 30% of those lost shoppers back into paying customers.
The catch was that it could not be done at scale by people. Hiring a call team is slow, messy and expensive. Callsy's own figures put a human call at €4 to €12, against about €0.18 for one of its AI calls. The team started testing voice AI in 2024. Arunas says it was not very good back then, but it still worked, which told them the idea was real.
Today more than 300 businesses use Callsy, Arunas says. It plugs into more than 50 shop systems, including Shopify, WooCommerce and Magento, so a new customer is set up in a few clicks and the agent already knows the products, prices and delivery terms. The agent says it is an AI on every call.
From one call to a whole team
Callsy no longer thinks of these as simple voice bots. It calls them digital employees. Each one has a memory of every customer and works across phone, SMS, iMessage, WhatsApp and email. One recovers abandoned carts. Another answers incoming questions. Another follows up with a shopper who is not ready yet, by message rather than by phone, and keeps the relationship warm until they are. Some businesses use Callsy as an AI receptionist.
Arunas says the team can move very fast. On some calls a customer asks for a feature, and it is built before the call ends.
Where it works best
Arunas shared the part most founders keep to themselves: which shops it works for.
The strongest fit is anything expensive, especially furniture. A sofa or a dining table raises a hundred questions. Will it fit? When will it arrive? Will I be home? And it is not something you buy hoping to send back, because you almost never do. So people need one final check before they click buy. Callsy's agents give it, then help again after the delivery, when new questions come up. Pergolas and other big purchases work the same way.
The second fit is subscriptions. People who are about to cancel are at a crossroads, and a good conversation at that moment can keep them, often with an offer the brand is glad to make to keep a customer for longer. In health products, the same kind of conversation can stop someone buying something that does not suit them, which saves a refund later.
It does not work on marketplaces such as Amazon, where the shop never gets the customer's details. Callsy works with brands that sell directly to their own customers.
The number that explains all of it: if you have ten real conversations with people, Arunas says, 30 to 40% of them will likely become customers. Leave them on the website alone, and that falls to single digits.
Who else is in the space
Voice AI is one of the busiest corners of tech. Retell AI, Bland, Vapi and Synthflow all sell the building blocks for AI phone agents, mostly to developers and call centres. On the shop side, Postscript and Attentive help brands win customers back by text and email. The voice platforms do not start from the shop, and the messaging tools do not pick up the phone. None of them began with the one call Callsy began with: the shopper who was a click away from buying and walked off.
Why build it from the Baltics
North American investors may not know the region well, so Arunas made the case. The Baltics have deep talent in engineering and marketing, everyone speaks English and many people speak three or four languages, and government policy is friendly to startups and venture capital. Lithuania produces some of the most unicorns in the world for its size, from Vinted to the group behind NordVPN, and Estonia gave the world Skype. Valuations are still low compared with the United States. Arunas also points to a new wave of hacker houses, where young founders move in together and build for 16 to 18 hours a day.
Callsy is Arunas's third startup, after a couple of small exits, and the two co-founders have known each other for ten years.
What they are looking for
Callsy is raising a round of up to $1 million, somewhere between pre-seed and seed, and the round is open. Angels and investors who want to talk are welcome. It also wants more North American shops to work with, especially brands selling expensive products or subscriptions. Find the team at callsy.ai or on LinkedIn.
Why it is bigger than it sounds
This is a smart strategy. Call it narrow, wide, whole.
Narrow: Callsy took voice AI, one of the most crowded and general spaces in tech, and pointed it at one small problem everyone ignored, the shopper who fills a cart and walks away. It calls them and wins back 20 to 30% of them. That is a very attractive pitch, and it gets Callsy into the store.
Wide: once inside, it spreads. From calls out to calls in, and from voice to text, WhatsApp and email.
Whole: by then it covers so much that it becomes a new breed of CRM built for e-commerce, standing on a very strong base in voice calls.
So Callsy is not only a tech company. It is an e-commerce intelligence company. Winning back a customer who has already walked away is the hardest call there is. Callsy has made those calls thousands of times, and it has the numbers on which ones work. That knowledge is what will make its calls different from everyone else's.
What to watch
The first thing is the move from narrow to wide. Every shop that starts with cart recovery and then switches on inbound calls and messaging is proof the strategy works.
The second is the data. The more conversations Callsy has, the better it knows what wins a shopper back, and that is the part a new competitor cannot copy overnight.
The third is the round. An open raise of up to $1 million from a team with this much traction, at Baltic prices, is the kind of deal North American angels rarely see.