PressConf EU 2026: Recap & Highlights

PressConf came to Europe for the first time this year: 27 to 30 October at the H10 Marina in Barcelona, after two editions in Tempe, Arizona.

One stage, 50+ people, 8 sessions over two days, and a Friday of exploring the works of Gaudi.

Nothing was recorded, and nothing was streamed. That was the point of the event. Barcelona surprised us with sunny weather, which was a welcome change for all of us who flew in from colder places.

It was our first time sponsoring PressConf, and the room was the one we had hoped for: agency owners, plugin and product founders, and hosting people, talking about the business side of WordPress.

Four of us, no booth

PressConf has no expo hall at any sponsorship tier, so Ryan Bradbury, Georgi Dimitrov, Sergey Gruntovsky, and Svilena Peneva spent the week the way everyone else did: in the room, over coffee, and at the long lunches.

In Georgi’s words, PressConf was an intimate gathering, which made it a great setting for in-depth conversations and deep networking. Security kept coming up in those discussions, so our Imunify products quickly became the main topic of the day.

Ryan’s most striking takeaway from PressConf was how wholeheartedly the WordPress community is embracing AI rather than resisting it. The mood was practical, not anxious: AI is here, and adapting is not optional. Agencies and product builders alike are already rethinking what they build and how they deliver it.

Less surprising but just as encouraging, security was front and centre in almost every conversation. End customers are pressing agencies to build it into their service and prove its value, and that is turning security from a nice-to-have into a core expectation across the ecosystem.

Hector de Prada Gonzalez, Modular DS CEO, chatting with Ryan and Georgi

Three sessions we kept talking about

Thierry Muller, WP Engine: the web is not ready for agents

Thierry opened with a question that is hard to put down:

How does our work impact the world?

People consume information in three ways:

• Through the open web
• Walled gardens 
• Answer engines (mostly fed by the open web)

In other words, the world still relies on the open web, and companies keep investing in it, but differently than before.

The part that will stay with us is Thierry’s three layers.

• An AI answer first touches a website through training data, pages the model read months ago.
• Then the agent reads the live site through a web search (agents searching the live web with their own index or somebody else’s)
• The third layer is through agentic interaction, where the agent acts on the site itself.

This third layer is the one to prepare for, and we are not ready.

It sits on a stack of young standards (MCP, WebMCP, UCP, ACP, AP2, and Web Bot Auth), and the answer engines want it badly. They want users to stay with them, and that means sites an AI agent can use without a person clicking around.

Scored today, Thierry said, the web would not pass an AI-readiness audit, and a growing set of tools now runs exactly that audit.

AI agents also change the load on a site, because their requests arrive uncached and in bursts. WordPress can expose tools for agents to use, but serving those requests through PHP doesn’t hold up at scale: agents and human visitors compete for the same PHP workers, so a burst of agent traffic can slow the site down for everyone. Agents also need a reliable way to find accurate, site-specific content, and WordPress core has no built-in search or retrieval layer for that. 

His conclusion was that the smaller sites, those beyond the top million, can switch to a different platform if WordPress doesn’t meet their needs, but larger sites can’t and shouldn’t. So the infrastructure gap has to be closed by hosts and edge providers, not only in the software.

Flexibility is the main reason people choose WordPress, and once the infrastructure catches up, WordPress will be the only platform that offers both flexibility and the performance to handle agent traffic. 

He ended on the issue the industry discusses least: what happens to the web if everyone asks an AI answer engine instead of visiting sites. The answer engines depend on content that publishers create, but fewer people click through to the original sites, so publishers are already losing traffic. AI crawlers take far more content than they send visitors back, and no standard for permission or payment is finished.

The standards we choose today will decide whether the web stays sustainable, and they should fit what each site does: an online shop and a paywalled publication need different answers.

Those standards will be tested on real servers first, and keeping those servers fast, secure, and open is our job

Remkus de Vries, Within WordPress: tenacity is not perseverance

For almost two decades, Remkus said, WordPress was forgiving. You could succeed without asking why things worked.

Those days are behind us, and Remkus’s talk was about two ways to keep going when conditions change. Tenacity says, “I will keep doing this because I refuse to stop.” Perseverance says, “The conditions changed, so I will learn, adjust, and find a way forward.” His advice: keep the experience, and drop the assumptions that stopped being true. 

The line that stuck: sticking with WordPress is not a strategy.

Familiar effort feels productive. “Try harder” is an incomplete answer when AI changes how the work gets done, and clients expect more.

He asked the room to look outward first, at what has changed in:

• How clients buy
• Who you compete with
• What your current skills are worth
• How products get discovered
• How the market defines good work

Learn from your customers, and from people outside WordPress, because they see the problem from an angle you don’t. Then look inward: your learning habits, your assumptions about the market, your confidence, your willingness to have a distinct voice.

Reinvention, Remkus said, is rarely proof that the earlier path failed. More often it is how you keep going responsibly, and it starts with the smallest meaningful change rather than a rebrand.

His homework: write down the conditions behind your current success, cross out the ones that are no longer true, then pick one thing to level up and put it in your calendar like client work.

Mark Weisbrod, Greyd: don’t produce for everybody

Mark took the other side of the same problem: what goes wrong in a saturated market when everyone builds more of what they already have and improves it a bit with AI.

He walked through the textbook options for a crowded market (cost leadership, differentiation, niche focus) and landed on an uncomfortable conclusion. To stay in business, the industry has to solve much bigger problems for its customers, with a competitive advantage that produces high lifetime value.

Two lines that stood out. “Either solve the problems of a few rich people or a lot of poor people.” And the one that pairs with Remkus: don’t produce for everybody, because that is not a strategy. 

What we took home

Across the different sessions, the message of the week was consistent. The conditions that made WordPress businesses successful for 20 years are changing, from the agents arriving at the server to the clients who expect more and a market that is full.

The answer in every talk was the same: keep the experience, drop the assumptions, and pick the next change rather than the next rebrand.

Thank you to Raquel Manriquez and the PressConf team for a warm, well-run first European edition, and to everyone who sat down with us during the week.