TextYess Raises €4M to Turn Its WhatsApp AI Agents Into an Autonomous CRM for E-commerce
TextYess, whose WhatsApp AI agents serve 350+ e-commerce brands, is moving to an autonomous CRM where AI agents run the campaigns themselves. How it works, how it stacks up against Klaviyo, and what it means for brands running conversational channels.
Key Takeaways
- Milan-based TextYess closed a €4 million seed extension led by Italian Founders Fund. Having built its business on WhatsApp AI agents for more than 350 e-commerce brands, it is now shifting its focus to an "autonomous CRM" in which AI agents run the marketing work itself
- The role of AI agents is widening from answering and recommending inside a conversation to deciding who gets what message and when, then sending it. Klaviyo uses the same "autonomous CRM" language, and the race among seller-side agents has begun
- For Japanese merchants running conversational channels such as LINE, the groundwork is to unify orders, conversations and browsing history into one customer profile and to design a flow in which people approve what the agent proposes
A WhatsApp AI Agent Company Takes On the Whole CRM
The start-up is developing an automated customer relationship management system for e-commerce platforms based on AI agents
www.startupbusiness.itOn October 5, 2026, Milan startup TextYess announced a €4 million funding round. According to the company's announcement, it is an extension of its seed round, led by Italian Founders Fund with participation from existing investors VC Partners and Entourage. Total funding since founding now stands at more than €7 million.
Leaders League reports that the round was structured as a SAFE (Simple Agreement for Future Equity). The valuation was not disclosed. Dealroom places the round in the top 1% of 255 comparable seed deals raised by marketing startups in Italy.
The company was founded in 2023 by Riccardo Russo (CEO) and Edvaldo Gjonikaj and provides an AI CRM for e-commerce brands. More than 350 brands use it, including Ducati, Veralab and PittaRosso. According to Startupbusiness, the money will go in two directions: hiring to strengthen the CRM and its AI agents, and expansion beyond Italy, starting with the UK, where it already has its first customers.
Lead investor Italian Founders Fund is an Italian venture fund managed by Koinos Capital SGR. Tech.eu reported that it launched a €50 million debut fund in 2024 to invest in founders at the pre-seed and seed stages.
From 160 to 350 Brands: Two Years of WhatsApp Selling
To read this announcement properly, it helps to know where the company came from.
TextYess's product launched in January 2024. After raising €450,000 in May 2023, it closed a €2.4 million seed round in September 2025. At that point more than 160 brands were using it, and the company said it had passed €1 million in annual recurring revenue (ARR) in August 2025. The new figure of 350+ brands means the customer base has more than doubled in about a year. Current ARR has not been disclosed.
The product back then centered on three use cases: broadcasts in which each customer receives a different message based on purchase history, automated flows such as abandoned cart recovery and win-back of inactive customers, and an AI shopping assistant trained on the product catalog that answers questions around the clock. In the 2025 announcement, Russo said a WhatsApp message delivers three times the conversion rate of an email marketing campaign. That is the company's own claim, and the basis of comparison was not given.
Recommending products, answering questions and carrying the shopper through to purchase inside a conversation: up to this point, it is a textbook picture of conversational commerce.
"The Limit Was Never the Channel. It Was Time."
What the company put forward with this round is not about channels at all. It is a plan in which AI agents run the CRM in place of people.
Working with more than 350 brands, we learned that the real limit was never the channel. It was the time to do the work. Brands know what they should send, and to whom. They don't have the people to do all of it.
The announcement frames the problem like this. CRM, covering messaging, campaigns and customer service, is one of the biggest revenue drivers in e-commerce. Yet traditional software only hands over the tools, leaving the team to define the audience, write the content, review it and send it. Every new channel and segment adds work, and companies end up back at the same mass email with the same discount code for everyone. They are pushed into choosing between hiring more people and continuing to send messages that do not work.
TextYess's answer is to build the CRM and the agents that operate it as one system. The CRM merges each customer's orders, conversations and purchasing behavior into a single profile. On top of it, agents find the opportunities worth acting on each week, build segments, write campaigns and automated flows, and run them across email, WhatsApp and the website. Because they connect directly to the store and catalog, the company says they can also check stock, update customers on deliveries and handle return requests.
The company's website makes the workflow more concrete. An agent called "Atlas" proposes a plan, the team approves or edits it, the agent builds and sends it, and the results feed into the next plan. Segments can be described in a sentence, such as "bought the serum, no reorder in 45 days", and nothing is sent until someone approves it.
The key design choice is that autonomy increases in stages. The announcement says agents first propose and the team approves, then take on more autonomy as they prove their results. Rather than going fully automatic on day one, trust is built before more is handed over. That sequence speaks directly to the anxiety of anyone entrusting promotional budgets and customer touchpoints to AI.
Discussion of agentic commerce tends to focus on buyer-side agents such as ChatGPT finding and buying products. TextYess shows what is happening on the other side, the seller side, where agents are expanding from sales associates into the people who run the campaigns. Recommending a product to one customer in a conversation and choosing which of tens of thousands of customers to contact next both rest on the same data. Russo also said the same system and agents could eventually run paid ads, search, social and the website, which positions CRM as the entry point.
Klaviyo Uses the Same Words
The term "autonomous CRM" does not belong to TextYess alone. Klaviyo, known for its email and SMS marketing platform, calls itself the "autonomous B2C CRM."
In its June 30, 2026 announcement, Klaviyo moved its marketing agent, Composer, into public beta. Composer audits live campaigns and flows, ranks revenue opportunities and builds email and SMS campaigns. Its service agent, Customer Agent, responds to customers across web chat, email, SMS and WhatsApp, and can complete returns itself.
| Aspect | TextYess | Klaviyo |
|---|---|---|
| Starting point | AI agents on WhatsApp (launched 2024) | Email and SMS marketing platform |
| How it describes itself | Autonomous CRM for ecommerce | Autonomous B2C CRM |
| Customers | 350+ brands | 196,000+ paying customers |
| What the agents do | Find opportunities, build segments, create and send campaigns and automated flows | Composer surfaces opportunities and builds campaigns; Customer Agent handles service and returns |
| Channels | Email, WhatsApp, website | Email, SMS, web chat, WhatsApp (Customer Agent) |
| Human role | Agents propose and the team approves at first; autonomy grows as results are proven | Nothing goes live without approval |
The two companies make strikingly similar claims. Unify customer data in one place, let agents find opportunities and build campaigns, and have people approve. The difference lies in where they started and how big they are. TextYess came in through AI agents on a conversational channel, while Klaviyo is layering agents onto a massive email customer base. Klaviyo's customer count is more than 500 times larger.
On WhatsApp itself, the platform owner is moving too. Meta's pricing page added a "Meta Business Agent" message category from July 2026. These are responses powered by Meta's own agent, priced at $2 per million tokens, or roughly 4 to 5 cents per message. AI agents on conversational channels have become a space where specialist startups, CRM giants and the platform itself compete on the same ground (we covered Meta Business Agent in more detail in this article).
The Ground Rules of WhatsApp
TextYess's business sits on top of the rules of the WhatsApp Business Platform, and those rules have shifted considerably over the past year or so.
The first change is pricing. Since July 1, 2025, Meta has charged per delivered message. Rates depend on the category, such as marketing, utility or authentication, and on the recipient's country calling code. Within the 24-hour window opened by a customer's message, businesses can reply freely, but to reach out first they must use template messages, and promotional templates incur marketing rates.
The second is the terms of use. According to Business Standard and Gulf News, from January 15, 2026, general-purpose AI chatbots can no longer be offered as the main service through the Business API. AI built into a broader business service, such as customer-support bots, is still permitted. A service like TextYess, focused on a brand's own selling and service, falls on the permitted side.
When every message carries a cost, sending the same promotion to everyone turns directly into wasted spend. In our view, the autonomous CRM idea of having agents narrow down who gets what and when fits well with this pricing model. At the same time, the more sending decisions are handed to agents, the heavier the operational question of how to avoid over-messaging that leads to blocks and opt-outs.
What Japanese Merchants Can Take From This
In Japan, LINE rather than WhatsApp sits at the center of conversational channels, with different mechanics and pricing. TextYess's product does not translate directly. Structurally, though, two lessons carry over.
First, data unification comes before handing operations to agents. Both TextYess and Klaviyo start by merging orders, conversations and browsing history into one profile per customer. If LINE friend data, e-commerce purchase history and support records are managed separately, an agent cannot judge who should receive what.
Second, decide on the approval flow in advance. Both companies start with human approval. If a company agrees internally on which campaigns can be delegated and which metrics will be used to confirm results and widen autonomy, it will be ready to accept what agents propose. For the difference between conversational commerce and agentic commerce, see this explainer.
Conclusion
In absolute terms, TextYess's €4 million is not large. What matters is that a company that began by selling on WhatsApp is moving toward handing the entire customer relationship to AI agents.
Klaviyo is heading the same way, and on WhatsApp, Meta's own agent has started to operate. Can a small specialist show a meaningful difference from the giants in the UK market? And how far will autonomy, starting from propose-and-approve, actually extend? Those are the two things to watch next.
