Direct answer: Yes, ecom operators should care, but not yet with their whole budget. Birdsview, a Leipzig startup, just raised EUR 2.5M to scale Avys, an AI agent that writes and sends retention emails without a human strategist, copywriter, or agency account manager in the loop. Birdsview says brands see 8x more revenue per email and 3x more clicks than standard campaigns, built on a product-intelligence layer trained on 1.2 million labelled products and 52 million email interactions a month. The pitch: stop paying for a CRM tool and an agency to run it. Pay for the outcome instead. I ran a CRM agency's economics in my head before I ran the numbers on paper. Here is the verdict.

What Avys Actually Does

Strip away the funding headline and Avys does one specific job: it runs e-commerce retention messaging end to end, the way a marketer would, minus the marketer.

Most email tools, Klaviyo included, give you the pipes. You still segment the list. You still decide what to say. You still write the copy, set the send logic, and QA the flow before it goes live. That is the work retailers hire agencies to do, typically at three to five times the price of the software license itself. You pay twice. Once for the tool. Once for the humans who make the tool useful.

Avys collapses both costs into one system. It reads a retailer's catalog the way an experienced marketer would: what does this product connect to, what need does it satisfy, what is the best angle to sell it, and when is this specific customer likely to need it again. That is the "product-intelligence layer" Birdsview's co-founders built alongside their first 100 customers. It now spans 1.2 million labelled products across 60 industries, and it learns from 52 million email interactions every month.

The output is not a segment blast. It is an individual email, written in your brand voice, to one customer, based on that customer's specific purchase history and repurchase timing. No agency account manager, however good, is writing 50,000 individual emails a month. Avys is built to do exactly that.

The Numbers That Matter

Three numbers carry this story. I want you to sit with each one before you decide what it means for your P&L.

  • 8x revenue per email. Birdsview reports brands using Avys generate eight times the revenue per send compared to traditional campaign tools. That is the headline number, and it is the one that will get quoted in every deck this quarter. Treat it as a ceiling from early adopters, not a guaranteed floor for your store.
  • 3x more clicks. Individualized subject lines and product recommendations beat segment-level blasts on click-through. This tracks with basic personalization research going back a decade. The delta here is doing it at the individual-customer level without adding headcount.
  • 1.2 million labelled products, 52 million interactions a month. This is the moat, or at least the thing Birdsview is betting is the moat. Data like this cannot be bought off a shelf. It compounds with every new customer the system onboards. Whether that compounds fast enough to stay ahead of Klaviyo, Attentive, or a well-funded competitor doing the same thing is the open question.

Layer on the market backdrop and the timing case gets stronger. Customer acquisition costs have risen 222% over eight years, and multiple 2026 industry reports peg average D2C brands at losing roughly $29 on every new customer acquired. Retention is not a nice-to-have line item anymore. It is the only lever left that does not require you to outbid Temu and Amazon on ad auctions. Fortino Ventures, the lead investor in Birdsview's round, frames the target market at $56B across retention SaaS and services combined. That is the investor's number, not an independent audit, and I flag that distinction on purpose. Independent market-research firms size the narrower "customer retention software" category closer to $5-9B today. The $56B figure only holds if you fold in every agency dollar spent running that software, which is exactly the spend Avys is targeting.

The Agency-Replacement Math

Here is where I want you to open a calculator, not a deck.

If you run a mid-size ecom brand at $5M-$20M in annual revenue, your retention stack probably looks like this: a CRM/ESP license (Klaviyo, Attentive, or similar) running $1,000-$5,000 a month depending on list size, plus an agency or in-house retention manager running $4,000-$15,000 a month for strategy, copy, and flow management. Call it $60,000-$240,000 a year, fully loaded.

Avys is pitched as a replacement for both halves of that stack, not an add-on to it. If the pricing lands anywhere near what early Seed-stage agentic tools typically charge, you are looking at a fraction of that fully-loaded cost, with the agent doing individual-level personalization that no agency team, however large, has the hours to do manually. Birdsview's own math backs this: they estimate agencies represent roughly 40% of the incumbent tooling vendors' revenue, which is precisely why those vendors have little incentive to automate their own distribution channel away. Avys has no such conflict. It was built by people who ran a CRM agency themselves and, in their words, decided to make that business model obsolete.

The math only works, though, if the output quality holds without a human checking every send. That is the caveat that matters more than the price tag.

What Could Go Wrong

I do not fall for the demo. I fall for the numbers after six months of live traffic. Here is what I would want proven before I moved my whole retention budget onto this.

  • Seed stage, not proven at scale. EUR 2.5M is a Seed round. Birdsview was founded in 2023 and has built its product-intelligence layer alongside roughly its first 100 customers. That is real traction, but it is not the same as thousands of brands across every vertical running Avys for two years. Ask for reference customers in your specific category before you sign anything.
  • Brand voice fidelity at scale. Writing in "the brand's voice" for one customer is a solved problem. Writing in that voice consistently across tens of thousands of individualized sends, every day, without drift, is a much harder problem. Ask to see 50 real sample emails from an existing customer in your category, not the three cherry-picked ones in the sales deck.
  • GDPR and data handling. Birdsview is a German company, and GDPR compliance is table stakes there. But you are handing over your full customer purchase history and email interaction data to a third-party model. Get the data processing agreement reviewed before onboarding, same as you would with any ESP migration.
  • Vendor lock-in on the wrong thing. If Avys becomes both your CRM and your agency, you have no fallback layer if the company stumbles, gets acquired, or pivots. Keep your customer data exportable and your list ownership clean, the same discipline I'd apply to any single vendor doing this much of your revenue engine.
  • The 8x number is an average, not your number. Averages hide dispersion. A DTC apparel brand with a tight SKU count and a supplements brand with a 40-day repurchase cycle will not see identical lift. Pilot on one segment of your list before committing the whole file.

The Sovereignty Stack Verdict

I evaluate every AI vendor pitch against what I call the Sovereignty Stack: does this tool make you more independent, or does it make you more dependent on someone else's black box? Three layers matter. First, data sovereignty: do you own the customer data and the model's outputs, or is it locked inside a platform you cannot leave without losing your retention history? Second, process sovereignty: can you still explain why a message went to a customer, or is the logic entirely opaque? Third, economic sovereignty: does this tool lower your fixed cost structure, or does it just move the cost from a line item you control (an agency contract you can renegotiate or fire) to one you cannot (a black-box subscription with usage-based pricing that scales with your success)?

Avys scores well on economic sovereignty if the pricing is transparent and usage-based rather than a revenue-share trap. It is unproven on process sovereignty until independent operators publish their own send logs and results. Data sovereignty depends entirely on the contract terms you negotiate, not the marketing page.

At Hartford-Munich Re, I was one of 15 innovation scouts in a 55,000-person organization. My job was to find the technology that would make existing processes obsolete before someone else did it to us first. This is that kind of technology for ecom retention. The CRM agency model, three to five times the software cost, paid for humans to do work a well-trained model can now do faster and, per Birdsview's numbers, more profitably. I have run ad agencies. I know exactly which line items get automated first. Individualized, high-volume, pattern-based copy and send-timing decisions are first on that list every time.

The Verdict

I would run a pilot, not a migration. Take one segment of your list, maybe your win-back flow or your post-purchase sequence, and run Avys against your current agency's output for 90 days on the same segment. Measure revenue per email, click rate, and unsubscribe rate side by side. Do not touch your full retention budget until you have your own numbers, not Birdsview's.

The direction is not in question. Retention is now the cheaper, more defensible growth lever compared to acquisition, and CAC increases of 222% over eight years are not reversing. The agency-plus-tool model was built for a world where personalization meant "segment of one thousand." Avys, and whoever else builds a credible competitor in the next 18 months, is built for a world where personalization means "segment of one." Systems beat slogans. An 8x number in a press release is a slogan until it is running against your list. Pilot it, measure it, and let your own P&L tell you whether to fire your agency.

If you are still running retention on rule-based flows with no AI layer at all, read my breakdown of a 9-flow retention engine that runs without you before you evaluate any agentic tool. And if Klaviyo is your current system of record, my owner-operator review of Klaviyo's CDP and AI layer will tell you what the incumbent is already doing to defend against exactly this kind of challenger.

FAQ

Q: Is Avys a replacement for Klaviyo or Attentive?

It is pitched as a replacement for both the ESP license and the agency or in-house team that operates it, not as an add-on integration. Whether it fully replaces your ESP depends on whether you need Klaviyo's broader data warehouse and SMS features for functions outside retention email. Some operators will run Avys alongside a lighter ESP plan. Others will migrate fully.

Q: How much does Avys cost compared to a CRM agency?

Birdsview has not published public pricing as of this writing. The company's stated goal is to undercut the combined cost of a software license plus an agency retainer, which typically runs three to five times the license price. Get a direct quote and compare it against your fully-loaded 12-month agency and tooling spend before deciding.

Q: What is the product-intelligence layer, exactly?

It is a model trained on 1.2 million labelled products across 60 industries plus 52 million monthly email interactions. It maps how products relate to each other, what customer need each product satisfies, and when a given customer is statistically likely to need a repurchase. That timing and relevance layer is what Birdsview says produces the 8x revenue-per-email lift.

Q: Is this safe for GDPR compliance if I run a European store?

Birdsview is a German company headquartered in Leipzig, so GDPR compliance is built into their operating environment by default. That does not remove your obligation to review the data processing agreement and confirm where customer data is stored and how long it is retained before you connect your customer list.

Q: Should I fire my CRM agency now?

No. Pilot Avys, or a comparable agentic retention tool, against one list segment for 90 days first. Compare revenue per email, click rate, and unsubscribe rate against your current agency's output on the same segment. Make the call with your own data, not a press release.

Jeff Barnes is the founder of demg.ai and Angel Investors Network. He has no personal position in any company, fund, or platform named in this article unless explicitly stated. demg.ai provides marketing education and systems for owner-operators, not investment advice. All investments and business decisions involve risk, including loss of capital.