Jesse Friedman: Welcome to Impressive Hosting, a podcast about the role hosting plays in shaping the open web. I’m your host, Jesse Friedman. On this show, we go deeper than uptime and dashboards. We talk about hosting as infrastructure, about ownership, independence, and what it takes to build ethical, high-end WordPress hosting that actually serves creators, businesses, and the internet itself. Before we dive in, head to impressive.host. That’s where you can comment on episodes, ask follow-up questions, and help shape future conversations. You’ll also find links to follow, like, and subscribe wherever you listen. We are sitting down with Vito Peleg, the founder and CEO of Atarim. We just had a great conversation. If you haven’t heard it, go back and listen to episode one. We talked about a lot of different things. We talked about the power AI plays in the agency space. We talked about networking and communicating with fellow like-minded folks in the industry and the benefits of that. We talked about Vito’s past and how he was building GeoCities websites while he was rocking out in London, and how that helped him build a small agency that grew and grew. He eventually built Atarim, which is an agency empowerment tool to help them speed up their work with clients and give them a leg up with AI. Vito, you put AI into WordPress in 2020. That might as well have been 50 years ago with the speed at which AI is moving. But you have something cool coming out, right? You have a new set of AI features inside of Atarim that are going to help agencies. Can we talk about that? Is it too early?

Vito Peleg: Well, it’s coming either now by the time this goes live or in a few days from watching this. So this is definitely a good time to talk about this. We launched AI in Atarim about a year ago, so it’s not something we’re completely introducing from scratch. But the vision of it is now coming to fruition with version five of Atarim, which is what we’re talking about.

Jesse Friedman: That’s great. And so the way you described it, is it correct to say that it’s an agency in your admin?

Vito Peleg: Not so much. Actually, we have a pretty big principle of taking the user out of the admin. Out of the WP admin, specifically. Especially the client. Get them out of there because that’s not where they want to live. The agency has acted as an abstraction layer for the WP admin for the past 15 years in the eyes of clients. So we’re leaning into that reality. We spoke about this in the previous episode. I’m not doing anything new. All I’m doing is digitizing human experiences and observing how people act in the real world. We call it the AI moat for web agencies. The reason we call it that is because there are three pillars that align with the current reality of how clients interact with AI and how agencies should leverage AI for their own abilities. The three pillars are self-serving clients, proactive agency, and automatic upsells.

If we break down each one of them, the reality of three-day SLAs, where clients ask you to change something on the website and if they’re lucky you can do it right away, but maybe not, maybe it’s the weekend, maybe you’re busy with another project, or maybe their email is the tenth one that hasn’t been opened yet. That reality is gone. We all expect things right now, on demand. And it’s not only through AI. People used to watch TV in a certain way, and that’s not reality anymore. Everything we want is at our fingertips. Changing a phone number on a website, publishing a new blog post, adding a new portfolio item, updating a product. We can’t live in this reality where we expect clients to wait on us and still pay premium.

So what we’re trying to do is align the client with their expectation, which is to self-serve on the small bits and pieces, while providing the agency the guardrails to make sure things don’t go sideways, that there are always traces, that there’s always a way to revert back any rogue changes. And then the agency acts as the escalation layer for those requests.

On the reactive side, you might already see some solutions like this being promoted, whether it’s Arrow or Hostinger’s Cody or Elementor’s Angie. These cover just the reactive side. You can go in, make a request, and the AI will do its best to perform that change for you. But all of these lack the escalation layer. I’m sure you have AI support to some degree in WP Cloud or through the Automattic portfolio. And even there, you wouldn’t let it be the only layer. You would have humans tackling maybe 80% of the work with 20% done by AI. Now it’s flipped. 80% is done by AI, 20% by the human. We look at it in that exact same way. That’s number one. Allow clients to self-serve because that’s what they’re after.

But at the same time, 75% of agencies say they’ve already lost work to AI because it’s instant. You can get results instantly if you just go talk to something else. So don’t lose your clients to AI. Give them something before they go looking elsewhere. That’s number one.

Number two is the proactive agency. This is a fundamental mindset shift that I’m sure successful agencies have already figured out. There are two types of agencies in this world if you boil it down to basics. You have the normal agency and the good agency. The normal agency is reactive by nature. Clients might pay you $100 or $250 a month for the right to reach out to you. If they do, it’s 48 to 72-hour SLAs and all of that. And then…

Oh, what’s up, Cliff? Give me one second.

Jesse Friedman: This is great. Don’t worry about it.

Vito Peleg: Sorry, man. There’s a holiday here in the UK.

Jesse Friedman: Don’t worry about it.

Vito Peleg: So yeah, the normal agency versus the good agency. The normal agency is the reactive one. Clients pay them $50, $100, $200 a month and if they’re lucky, someone will be there when they need them. Or they might be busy with other work, or on holiday. That’s one type.

Then you have the good agencies. The proactive ones. The ones that act as advisors, that reach out every month, get on a call, and look for opportunities. They ask: How is business going? What are your goals for the year? Let’s build a strategy so we can do something every month. How about we do this? How about we do that? We found a problem, fixed it, and it’s already done. Nothing for you to do. These are the ones that get referrals. These are the ones that grow organically without resistance from the market.

For the reactive side, the yes-man agency, there is a complete breakdown in the value chain between what the client perceives and what the agency provides. If you talk to these agencies, they call their care plans insurance. The client is paying for insurance. But if you ask the client, they’re paying a tax. There is a complete disconnect between what value means in the eyes of the client compared to the agency.

Some of these folks, including myself back in the day before I realized it, when you have a client paying for two hours of your time per month and they haven’t reached out in two months, it feels amazing. I made money without doing anything.

Jesse Friedman: It’s like the empty gym membership model. The gym wants as many memberships as possible and wants the gym as empty as possible.

Vito Peleg: Right. That’s exactly the same mindset. So the second pillar is making every agency proactive, and this is through autonomous workflows that run in the background and perform work for you. We’ve always had workflows and systems like this inside Atarim. But they were just about moving work from person to person. Now they can actually perform the whole cycle.

We talk a lot about WordPress and the open web. Nobody likes updating plugins, checking if they worked or didn’t work, reverting back, checking security. Nobody wants to deal with a downtime situation, a hacked website, any of that.

At my agency, we had what we called the Friday emails. Every Friday, we would spend four hours where every project manager or person in charge of projects would write to their active clients: here is what we did, here’s what we’re going to do next, and here’s what we need from you. It took four hours every Friday, but it was so worth it. Even if clients didn’t read it, they knew we were on their side. That increased trust and reduced micromanagement. When they know you are on it, you can lead the relationship rather than just react to it.

So that’s the second pillar. Moving people from being the yes-man to being the trusted advisor, from reactive to proactive, without actually doing all the work manually. You just decide which workflows you want to run autonomously for your clients.

Imagine the AI does a 30-day check-in after a client just joined, looking at all the work, the communications, where the project stands, and sends them a report of what was already done and where things are going next, based on project stages and all the features we already have inside Atarim.

The third one is the upsell machine. The concept here is that AI will never be able to do everything. Anything it can’t do, or that’s too risky, or that falls within your area of expertise, gets surfaced up to you. For example, let’s say I’m really good at SEO. I don’t want any AI-generated blog posts on my client’s website. I want to vet all of that and have every piece surfaced to me for approval before it goes live. But I’m not great at design, so go ahead and do whatever you can to make sure the design looks good.

On the flip side, if I’m a designer, there’s no way I’m putting AI-generated designs on my client’s website. I want every design to come back to me. But writing a blog post every week for each of my clients, within their niche and their tone of voice? That sounds great. Do that.

You can set those guardrails in natural language. You just tell it what you enjoy doing so it brings those tasks to you. But not only that. Every time these opportunities are surfaced to you, they get quoted by the AI and come to you as an upsell opportunity. That’s how we built our agency back in the day. Every scope creep is an opportunity for revenue. I’ve been saying on stages for years that scope creep is the healthiest thing for your agency’s growth. Encourage it. It means people are inspired. It means they’re still with you.

Jesse Friedman: Frankly, scope creep comes from the fact that they’re excited enough about the project to say, “I want to add that,” or, “I want it done that way.” And we’re constantly saying, “That’s not within scope. That’s not what you paid for.” Maybe there’s a way to work that in, and maybe AI can help with that.

I really like the analogies you’re using. One thing I keep going back to is the medical industry, at least here in the States. My mom was a nurse growing up, and I spent a lot of time in hospitals because I was born with arthritic knees. I was born with 80-year-old knees.

Vito Peleg: Really?

Jesse Friedman: So I was in hospitals all the time. But what I noticed is that when a patient comes in, the attending physician, the top doctor in the room, doesn’t see the patient right away. What they typically do is have an intern or resident, someone more junior, do an H&P, a history and physical. They ask all the questions, review everything, create a diagnosis, maybe order X-rays, and prepare everything before presenting to the attending. Then the attending goes in and says, “Okay, I’m going to meet with the patient. I’m going to make sure the student didn’t miss anything. I’m going to read the X-ray, and I’m going to say, you are diagnosed with this, and here is your treatment.”

I think this is so similar to how we’re starting to operate with AI. The client comes to the agency and says, “I want a new portfolio item. I shot this video, I created this commercial for my client, and I want it on my portfolio.” Typically that means getting into a queue, paying the agency, waiting until they have time. Or learning it yourself, but you want to go create commercials, not learn how to build websites.

What you’re saying is that your tools make it possible for a client to say what they want, the AI gets you most of the way there, ordering the X-rays, running the tests, making the diagnosis. But there’s still a human applying their talent and expertise to make sure it was done correctly, that it serves the rest of the website, thinking holistically about the project and the vision. Then they sign off and hit go.

In that way, you’re blending the need for AI to make agencies faster while also giving the agency something real to charge for. You’re not just handing it off to AI and saying, “Hey, we’re reusing something you could have done yourself.” You’re actually applying your value.

I actually held a fireside chat at the Automattic Dev AI event in New York last month where I interviewed Hillary Mason, who runs Hidden Door. She’s been at the forefront of AI for over a decade. What she talked about was taste. The idea that your best skill set is knowing what you’re best at. If I’m a web designer, I apply my taste to what an AI creates. The AI isn’t replacing me. It’s operating like a junior designer, and I’m still in charge of approving things and applying my standards. It’s just accelerating the grunt work I don’t care much about anyway. Does that sound right?

Vito Peleg: Yeah, and this is exactly the mindset. I call it the POV. All we have is our point of view. Our POV dictates how we act, how we interact with the world, and how we make things. That’s what art is. It’s the POV.

So I’m totally on the same page. Taste can also be replicated, though. People have been saying taste is the last differentiator for the past three or four years with AI. “It will never have taste.” But now you have design.md files that show pretty good taste. Maybe not yours, but it has taste.

So I think that has also been absorbed. But the decision-making, the point of view, that is where we shine. And it will get to a place very soon where you don’t even have the quirks anymore. Not the black text on black background hallucinations that happen because the design system wasn’t properly set up, or the mismatched border-radius on buttons. That’s being resolved in front of our eyes through agentic capabilities and the loops these agents run to recheck their work at different levels of fidelity.

But the point of view is the thing. And I would add one more: the relationship. We talked about this at the beginning of the first episode. In today’s reality, relationships are where it’s at. While AI does a lot of the work, having someone you trust who will be by your side if things go sideways, someone who is ahead of you, even by a little, to guide you to the next step, that still has a lot of value.

I remember when I started teaching guitar at 15. I had only been playing for one year before I started teaching. Not because I was amazing, but because I was teaching nine-year-olds.

Jesse Friedman: Right, right.

Vito Peleg: It’s always that game. The professional in any field lives and breathes their own space. Wherever they go online, they get served content relevant to their industry. Your clients will not. The point of view and the relationship you form with them is the way to create stickiness in any business.

Jesse Friedman: Yeah. One thing you touched on earlier that I want to circle back to is that agencies don’t want to deal with downtime, security issues, hacks, and things like that. I’d be remiss if I didn’t connect that back to the way people buy hosting.

One of the things I keep talking about with WP Cloud, and the reason I push WP Cloud with hosting companies, especially ones who say, “Our IP is our infrastructure, we’re a host, that’s what we’re selling,” is that the reality is becoming more clear. People aren’t buying hosting as much anymore. They’re buying a promise that they’re going to be successful, that they’ll have great uptime, that they’ll have secure sites. They’re not analyzing the infrastructure itself.

This is where WP Cloud has a unique opportunity to help hosting companies, because we’re setting the gold standard with fully managed WordPress hosting. Our partners like Pressable and Bluehost’s agency program are offering 100% uptime. Not “nines.” 100%. The reason is because you can make that promise on this platform. And the agency cares less about which actual boxes they’re on and more about that promise.

So when clients are interacting with agencies and using AI tools, I’d imagine, and you can correct me if I’m wrong, that a client saying, “I want to put this portfolio item up,” is going to care less about who does the grunt work and more about the outcome. They’re going to care about whether it was done efficiently, quickly, affordably. They care about quality over time, whether it’s up, whether it’s fast, whether they can afford it for their client. Same relational things agencies care about with hosting.

The thing I keep coming back to is faith. The same issue customers have when they set out to build their own website is: will it pay off? All the time spent learning WordPress, learning design, learning web skills they may not even care about. Will there be a finished product at the end? If their faith is low, they’re less likely to succeed.

It goes the same way with agencies. If the agency is building and strengthening that relationship so clients have faith that the work will reflect what they’re paying for and the value they’re getting, they’re going to care far less about who’s doing the intricate pieces and more about who’s signing off at the end.

Vito Peleg: Yeah, I’m with you. The DIYer trying to build a website is a perfect example of this. As an industry, we’ve been promising business owners one-click websites for the past 20 years. I haven’t seen one yet. But people still believe it. When you go into Wix or through a host into a website builder, you believe you can do it. But we also know from surveys, and Wix shares a lot of these stats publicly as a public company, that 82% of people who start building a website with Wix never finish. 82%.

Jesse Friedman: Yeah.

Vito Peleg: And in the hosting space it’s not that far off. We’re talking 50 to 80% with DIY tools. Sometimes we feel threatened by DIY solutions on our professional services side. But there was a stat from WebPros that did a market survey and found that DIY within our market is only 4.8% of the market. More than 95% is professional services, people going in and building. But the barrier to entry has been reduced with AI, the same way it was when WordPress came in, and then website builders came in. We saw a million builders joining our industry. That hadn’t happened before. Maybe before you had a few webmasters here and there, but you didn’t see that influx of new professionals before it became easier to build.

Now that it’s even easier to build, I think there will be at least another million professionals coming in. We call them vibe coders now.

Jesse Friedman: Yeah.

Vito Peleg: And they are us from 15 years ago, just in this new revolution. So I don’t think it’s going to die off. It’s all about the relationship, the trust you build with your client, while aligning your business model and your delivery with the expectations of the market, which is instant. Everything should be instant by default. It can take longer if it’s more difficult, but the default is instant.

Jesse Friedman: No, I think that’s a great point. Also, thinking about hosting companies, you mentioned the 30-day check-in earlier. You’re keeping customers engaged and helping them feel like they’re interacting with something that’s actually helping them achieve their goals.

One of the things we talk about with hosting companies all the time is that you should align yourself with the success of your customer. Not just focused on charging for ancillary things like SSL certificates, but focused on how you make sure they’re growing so you’re growing as well.

It makes me think about what hosting companies should be doing here. Atarim is starting to be bundled with hosting companies, which is great. But I also like this idea of the 30-day check-in because one of the things we see a lot with hosting companies is that DIYers, even though they’re a smaller percentage, still need extra hand-holding. They need that support. They need that level of care.

Maybe hosting companies should be implementing something like this too. An AI check-in, something that can actually listen to what you’re saying. It doesn’t necessarily have to solve all the problems. Maybe it doesn’t have to be just a chat widget where people can tell it’s AI and don’t want an inhuman conversation. Maybe it’s as simple as a 30-day check-in. Tell us how things are going. What are you struggling with? What do you need help with? And be transparent. “I’m an AI. I’m not going to solve your problems right here, right now, but I want to hear what you’re dealing with, how it’s going, and I’m going to bring it to a human who will help you be successful.”

That layer of transparency can go a long way with people who feel vulnerable in those moments. And it creates a real check-in. I think hosting companies need to do a lot more of this. Right now everyone is hyper-focused on how to build a site builder, how to use AI to get people live faster, how to reduce time-to-live, how to reduce churn. But I haven’t seen much focus on how to help them sustain that success. How do we keep them going? How do we make them feel heard? That might be a really great way to implement that.

So tell us about how Atarim is working with hosting companies.

Vito Peleg: Right. Very similar to what you’re describing. If we go back to the beginning of hosting when there were no agencies, just a handful of webmasters, or even before that, the host was doing the professional services. They were offering the website builds, writing the HTML and PHP, all of that stuff in the early days. Until they realized that selling hosting was way easier because it’s software. It’s scalable, it’s repeatable. It has the three pillars of software: unlimited scale, high margins, and low to no touch.

So they pivoted, dropped that model, and allowed web agencies to rise in our ecosystem, because people still needed to be served and not everyone wants to spend 100 hours figuring out how to build their first small business website.

But it’s coming back around. At Atarim, over the past few years, we’ve been powering what are called DIFM operations. Do-it-for-me operations. That’s basically the way to fix the leaky bucket of DIY. Every time someone can’t figure it out on their own, you offer to do it for them for a grand or two grand. And we’re serving some of the largest hosting companies in the world that have been doing this for a while. We came in with the software and cut delivery time in half.

I’ll continue sharing stats since people seem to like it. There was a report last week from CloudLinux and WebPros, a hosting industry market survey. One of the interesting points was that 50% of web hosts are looking into professional services as a way of increasing activation, reducing churn, and dramatically increasing ARPU.

Jesse Friedman: Yeah.

Vito Peleg: ARPU is average revenue per user. The reason for this is that we are in an extremely special time right now. Sequoia was talking about the concept of one-to-six. Every dollar spent on software, if someone is clicking the buttons of that software, is worth $6. That direct correlation of one-to-six from DIY to professional services is true if you think about care plans or a lot of the work we do as professionals. It’s also a nice way of thinking about pricing.

But it also means 6X ARPU on commoditized hosting plans. The $5 or $10 hosting plans. How do you get them to $80? There’s no way apart from providing services.

Now, why is this a special time? It’s because services are starting to act like software. Unlimited scale. We just talked extensively about giving that gift of unlimited scale to our agencies. High margins, because the 6X is no longer directly correlated to payroll. And low to no touch, which is exactly what AI allows us to do.

That’s what brings hosting brands back to looking at professional services and saying, okay, this fits perfectly within our niche. We’re already selling the domain. We’re already selling the infrastructure. We’re already selling the website builder. Why not just give the client what they actually want, which is only two things: more traffic and more sales.

Jesse Friedman: Yeah. No, that’s a great way to sum it up. It’s funny because I think we all come from the WordPress space, and it’s very grounded in open source. This idea that we’re going to give back, write code, and help people get a voice online. We want to empower people. And I think there’s this mentality that there’s something wrong with charging for open source software. That’s the furthest thing from the truth.

But I also think that bleeds into this idea that because it’s open source and you can use it for free, people will just want to use it for free. The reality is that your customer experience improves significantly when you invest in it, and the way you invest in it is by making money. We are a commercial enterprise. We live in an economy that requires us to make some money.

Vito Peleg: It’s a healthy thing.

Jesse Friedman: Right. Unless we’re going to become a post-Star Trek no-money utopia, which maybe AI is leading us toward, who knows. But until then, you need to make money. You need to keep the lights on. You need to reinvest back into R&D and keep things going really well.

The thing I keep coming back to is that all these people who are building software and doing this work in open source tend to forget that the people asking for “do it for me” are focused on getting a gig for their band, producing a new widget, selling cupcakes at their bakery, or trying new hair colors for their salon. That’s where they want to focus their time and passion. Whether it’s five bucks for a service, twenty bucks a month for a tool, or five hundred bucks to have it configured for them, people forget that putting a price tag on something gives the customer an opportunity to say, “I value your product, I just don’t want to set it up,” or, “I value your hosting, I just don’t want to build a website. Give me an opportunity to spend a little money.”

We often think about these customers as if they have no money to spend at all. I think the reason is because of the race to the bottom on hosting pricing. When you charge two bucks a month for hosting, suddenly the website itself costs $1,000, and then they want $50 a month on top of that to manage it.

Vito Peleg: And…

Jesse Friedman: Right, and then things get really complicated. We’re basically saying, “Hey, we’re selling access to the street for pennies,” and then suddenly you have to buy a car to drive on it, and service the car, and everything else. But all they wanted to do was get a website live. Yes, they’re going to choose hosting because that’s usually the first step, or maybe they hire an agency. But when they find out how complicated it is, it’s not that it needs to be so easy they can do it in two minutes. It needs to be easy enough that they can actually learn it and have faith, like we talked about earlier. Or just give them a way to shortcut it, and sometimes the best shortcut is a few bucks. Be proud of that. Don’t hide it. Don’t sweep it under the rug. Don’t be ashamed of asking for money. It’s how you keep the lights on. It’s how you invest back in your product and actually help them succeed faster.

Vito Peleg: Yeah.

Jesse Friedman: So anyway.

Vito Peleg: You know, I look at this from the perspective that I’m not the highest earner in my company and never have been. My team members make more than me. I’ll win in the long run. That’s my perception of it.