Speakers

Talk Description

Content Management Systems are long-term architectural decisions, while AI evolves at high speed. This mismatch creates a real challenge for modern content platforms: how to integrate AI without locking the system into today’s assumptions. In this talk, Adrian Stelmach from explitia.com explores how AI should be positioned within CMS ecosystems as a replaceable architectural layer, not a built-in feature. The focus is on system design, clear boundaries between CMS, integrations, and AI services, and architectural decisions that allow content systems to evolve over time. The presentation is based on real-world CMS projects and discusses practical patterns and anti-patterns for AI-driven enrichment, search, and personalization—without vendor hype.

Talk Description

A lot of great Tech Leaders are great with architecture, but struggle with "human problems". Especially on the edge of tech and business (i.e., when implementing CMS systems for non-technical clients). This keynote reveals why this is happening and provides a single, battle-tested framework to diagnose and prevent team miscommunication. Instead of generic advice on empathy (sic!), I'll give you the framework based on Process Communication Model® (PCM®), and 10+ years of my work with tech leaders & their teams. You will discover how to diagnose the real root causes of the people-related issues, and how to build a bridge between analytical devs and action-oriented managers and clients.

Workshop Description

CMS backends bring together navigation, forms, dialogs, notifications, and editorial controls in workflows that people use every day. A problem in one interaction can make the entire task confusing or impossible, while the source of that problem may be difficult to recognize from the visual interface alone. In this hands-on workshop, developers and UI designers will test realistic editorial tasks with a keyboard and will additionally get a practical introduction to testing with voice control and a screen reader. We will follow complete workflows, from navigating the backend and editing content to handling errors and publishing a page. What happens during those tasks will help us trace accessibility issues back to the design, HTML, ARIA, and interactive behavior. Developers will learn what to look for when they investigate what failed, recognize the likely cause in the code, apply an appropriate fix, and find reliable implementation guidance when the answer is less obvious. For UI designers, the workshop adds an accessibility perspective to decisions they already make about interaction states, focus behavior, contrast, navigation order, instructions, and feedback. This shared understanding improves the design and developer handoff. Expected behavior can be defined earlier, developers spend less time filling in the gaps, and potential accessibility issues are caught before they begin delaying development or testing. Attendees will leave with a practical checklist they can use during design reviews, development, testing, CMS configuration, and product selection, along with reliable sources for investigating and fixing the issues they find. The workshop is relevant to developers and UI designers working with any CMS. No previous accessibility specialization is required. Participants should bring a laptop. Bringing a test instance of a CMS they work with will make the exercises more relevant, but it is not required.

Talk Description

Most accessibility conversations stop at the frontend. The backend, the interface where content actually gets made, is almost never part of that conversation. That is a problem, and it is becoming a legal and financial one. In this session, Anne-Mieke Bovelett makes the case that an inaccessible CMS backend is not just bad UX. It is a barrier to employment, it causes high support costs, it's a compliance risk under disability employment law, …and it is a gap in the market that developers with the right skills can fill. The session uses a demo to show what happens when a content editor tries to do their job using only a keyboard, or with voice control. Not because those users are rare, but because they are more common than most developers assume. These are the same people who end up calling support when the interface breaks on them, and those support hours are not free. A CMS backend built with accessibility as a baseline means fewer mistakes during content entry, fewer support tickets, and less time spent explaining where things are or why something is not working. The return on that investment is measurable, and it lands directly on the client's bottom line. The session also covers disability employment law. Many countries legally require companies above a certain size to employ a set percentage of people with disabilities, or pay a financial levy instead. If a company is required by law to employ someone with a disability, and their internal tools are inaccessible, they cannot fulfil that obligation. The CMS backend is part of that picture. And then there is AI. A lot of people in the industry are quietly assuming that AI will eventually take care of accessibility, or replace the human doing content entry altogether. This session explains exactly why neither of those things is true, and why the demand for developers who understand accessible interface design is only going to grow.

Talk Description

Everyone is adding AI to their content stack. Few teams are getting results. The difference isn't the AI tool, it's what's underneath it. This talk explores why content structure is the hidden dependency in AI-powered workflows, and what marketing and content teams actually need to do before AI can deliver on its promise. Drawing on real work building AI-ready content infrastructure, you'll walk away with a practical framework for auditing your content model, understanding where headless and composable CMS architecture fits, and knowing what "AI-ready content" actually means in practice. No hype, no demos. Just the lessons learned from teams trying to make AI work with the content they already have.

Talk Description

In a CMS landscape often cluttered with superficial AI "add-ons," the real challenge for architects is building an intelligent orchestration layer that remains independent of the delivery platform. The goal is not just to generate text, but to eliminate the friction between human experience and digital publication. This session presents a blueprint for AI-assisted storytelling: a framework-agnostic architecture that treats the CMS as a modular destination rather than a constraint. Using the "Nomad Pipeline" as a real-world case study—developed to automate high-frequency publishing during a motorcycle journey across borders—we will explore how to decouple content intelligence from the final stack. What we will cover: The Intelligence Layer: Orchestrating GitHub Actions, OpenAI’s Whisper, and Google Gemini to transform raw, multimodal inputs (voice and imagery) into structured, context-aware data. Architecture vs. Plugins: Why moving the AI logic outside the CMS improves scalability, data governance, and system evolution. Agnostic Distribution: A practical demonstration of pushing the same AI-refined content to diverse environments: the dynamic ecosystem of WordPress (via REST API) and the high-performance static architecture of Astro.js. Human-Centric Automation: How to use AI to amplify the author's voice and authenticity while the system manages the technical "heavy lifting."

Workshop Description

Obsidian trzyma wiedzę w zwykłych plikach Markdown na dysku, a Claude Code potrafi te pliki czytać, porządkować i pisać od nowa. Razem działają jak CMS, tylko bez panelu i bazy danych: notatki, research, oferty i treści na stronę leżą w jednym miejscu, a agent AI pracuje na nich bezpośrednio. Pokażę na żywym przykładzie, jak taki system jest zbudowany i co robi za mnie każdego dnia i jak uzupełnia moją stronę www

Talk Description

For two decades, the industry chanted one mantra: write great content, and top rankings will follow. That mantra is now wrong. Users have moved to ChatGPT, Perplexity, Claude, and Google's AI Overviews. What those systems recommend isn't determined by content quality alone, but by what your CMS outputs and how. We're at the beginning of a transition from human-readable content to machine-interpretable content. Most CMSes are quietly working against the teams using them: no or weak schema, fragmenting content into WYSIWYG soup, lagging on Core Web Vitals, and shipping no native path for AI integration. Great work, invisible to the systems that now drive discovery. This session gives CMS teams the exact technical checklist to fix it: the schema types that matter, how to model content as atomic blocks AI can chunk, the speed thresholds that decide whether crawlers stay, and the MCP and API patterns that keep your content live inside AI. You'll leave with a crash audit you can run on your own CMS and a clear answer on whether it's helping you or holding you back.

Talk Description

When organisations plan a content migration, one of the first challenges is understanding how their content should be structured. Ideally, that thinking happens before a platform is selected. In practice, the tools we use often pull us towards either abstract diagrams or the specifics of a particular CMS. Both are valid starting points, but getting from an initial idea to a detailed, usable content model can mean revisiting decisions and repeating work. Does getting specific really have to mean getting platform-specific? Drawing on real-world project experience, this talk explores a missing link in content modelling: a way to think through content architecture in detail while keeping implementation choices open. It looks at why recording intent matters—not just what fields exist, but how content fits together, what must be present, and what should be allowed—and how that understanding can support better-informed platform decisions. The session introduces CMMF (Content Model Master Format), a portable format that supports a platform-neutral way of thinking about content structure and intent in detail. It explores how capturing that intent independently of a CMS can make modelling work more useful during platform evaluation, through implementation, and as systems and requirements change.

Talk Description

Most CMS architecture in 2026 stacks the same blocks: a JS framework, a runtime, a bundler, a couple hundred npm packages, a container, a database service. Publr asks the opposite question — how far can you push each layer of the stack before the whole thing breaks? — and ships the answer as a single ZIG binary with embedded SQLite, no npm, no frameworks, and no external Zig packages. The whole CMS builds with one command and deploys as one file.

Talk Description

Feel like AI is about to replace everything you do? You’re not the only one. But most WordPress freelancers are still waiting and watching instead of acting. This talk is about one specific problem: how to scale yourself as a one-person business or small team without hiring, without burning out, and without letting quality slip. The question isn’t whether to use AI. It’s how to use it sensibly when your business grows, clients keep coming, and you can’t afford to bring someone on. I’ll show how I changed the way I work as a WordPress freelancer. No theory, just real examples from my daily work. You’ll see how I automate repetitive tasks and manage project context, and how that lets me look after clients better and deliver higher quality without working through the night. I’ll also show how I make sure my business doesn’t end up depending on AI. This isn’t a talk about magic AI agents that will change your life. It’s also not about finding clients or generating LinkedIn content. We’ll look at one concrete problem, concrete tools, and real examples. Data security and client confidentiality get full attention throughout, because in our industry they’re not optional.

Talk Description

CMS landscapes are diverse: headless, hybrid, visual editors, embedded content tooling, and bespoke APIs all have different expectations from the frontend. In this talk I’ll show how Astro’s framework-agnostic rendering and island architecture help unify these patterns, what architectural tradeoffs matter in real projects, and what we learned integrating a visual CMS into an Astro-based workflow — focusing on practical decisions and not on product marketing.

Talk Description

There’s a 21-year-old house in the open source neighbourhood. Millions of people live in it. Almost nobody has opened every door. Some rooms are hidden so well that even people who’ve lived there for two decades have never found them. Behind others are things the rest of the web is only just discovering. Today, the doors open. 20 minutes, a few “wait, WHAT?” moments, and a house you’ll never look at the same way again. Joomla users and curious neighbours welcome.

Talk Description

What if you could spin up a full CMS --- with a real server-side language, a database, and the entire plugin ecosystem --- right in the browser, with zero setup? No Docker, no hosting, no installs. Just click a link and you're in. WordPress already does it. WordPress Playground uses WebAssembly-compiled PHP paired with a SQLite backend to runs the entire CMS inside a browser tab or a CLI. It powers real-world workflows. Live plugin demos, PR previews, security research, translation tools, one-click Blueprint-based environments, and AI-assisted site building. But this isn't a WordPress-only story. Any CMS can become a client-side app thanks to WebAssembly. In the PHP world, the hard part — the WASM runtime — is already done. Typo3 and Drupal have started exploring this path. Laravel and Symfony components already work inside the Playground runtime. Composer runs in the browser. On the JavaScript side, there is prior art on running Node.js in the browser that, with a collective open-source effort, could unlock the same possibilities for JS-based CMSes. In this talk I'll show what WordPress Playground makes possible today, explain the architecture that makes it work, and make the case for why every CMS community should be investing in browser-native runtimes. Let's grow this ecosystem together.

Talk Description

I am a security researcher focused on finding vulnerabilities in Wordpress plugins. In just 3 months I have found 50 vulnerabilities, 11 cve’s and 39 awaiting diaclosure. I have acheved 2nd place in Patchstack March Leaderboard. I would like to deliver a talk mostly about broken authentication vulnerabilities in plugins, showing examples, real world impacts and remediations as well as use of AI in hunting and fixing vulnerabilities.

Workshop Description

MDCMS is an open-source, Markdown-first headless CMS where AI agents are first-class users rather than a bolted-on feature. In this hands-on workshop we'll integrate it directly into a real Next.js codebase: define schemas in TypeScript, sync via CLI, open a visual Studio for editors, and connect agents through MCP to bulk-edit content, update config from prompts, and extend the platform via its type-safe module system.

Talk Description

Teams should not have to choose between local Markdown/MDX editing and enterprise-level CMS collaboration. Local files are fast, transparent, easy to diff, and natural for developers and AI tools. But real content operations also need permissions, environments, version history, review flows, and safe publishing. In this talk, I’ll share how we designed MDCMS from the ground up to be AI-native across the whole system, while combining the best of both worlds: a local file-like editing experience and the controls expected from an enterprise-ready CMS.

Talk Description

Managing SEO and content across multiple products sounds manageable, until scale turns it into an operational challenge. In this session, I’ll share how we transformed our SEO and content workflows across a 4-product ecosystem (three WordPress plugins and one SaaS platform), managing more than 1,500 content pieces across multiple domains. As our content library grew, manual workflows became increasingly difficult to sustain: keyword research slowed down, content production took time and lacked consistency, audits became time-consuming, and identifying content gaps required significant manual effort. We needed a smarter, more scalable way to manage content operations without sacrificing quality or strategic direction. This talk explores how we introduced AI into our daily SEO and content workflows, not to replace people, but to remove bottlenecks and improve consistency across the ecosystem. I’ll walk through real examples of how AI supported us along the way and where human expertise remained essential.

Talk Description

Many SEO problems do not start with content, links, or individual page optimization. They start much earlier - in the way the CMS is designed. URL generation, redirects, canonical tags, indexation controls, XML sitemaps, pagination, multilingual setup, structured data, media handling, and content templates can either support organic visibility or limit it from day one. In this talk, I will show which CMS-level decisions have the biggest impact on SEO, what issues are commonly built into content management systems, and why some of them are difficult to fix after launch. The goal is to show how SEO can be designed into the system itself, instead of being treated as a patch after the website is already live.

Workshop Description

MDCMS is an open-source, Markdown-first headless CMS where AI agents are first-class users rather than a bolted-on feature. In this hands-on workshop we'll integrate it directly into a real Next.js codebase: define schemas in TypeScript, sync via CLI, open a visual Studio for editors, and connect agents through MCP to bulk-edit content, update config from prompts, and extend the platform via its type-safe module system.

Talk Description

Jak WooCommerce radzi sobie w erze "vibe codingu" i narzędzi AI? Dlaczego coraz mniej firm decyduje się na oparcie biznesu o sprawdzoną przez wiele funkcjonujących firm platformę? Jak to się stało, że przez lata budowana solidna podpora setek tysięcy sklepów na całym świecie traci pozycję "pierwszego wyboru" i staje się nieatrakcyjną propozycją? A może jednak WooCommerce ma jeszcze coś do zaoferowania?

Talk Description

Klient nie kupuje od Ciebie systemu ani linijek kodu. Kupuje poczucie bezpieczeństwa – zrozumienie, przewidywalność i współpracę z kimś, kto potrafi przeprowadzić go przez technologiczny chaos. W tej prelekcji pokażę, co naprawdę stoi za pytaniami, obawami i frustracjami klientów oraz dlaczego problemy w projektach IT rzadko wynikają z technologii, a znacznie częściej z braku zrozumienia. Opowiem, jak tłumaczyć decyzje techniczne w ludzki sposób i budować zaufanie w relacji z klientem. I dlaczego empatia to nie „bycie miłym” ani czytanie w myślach – a przekonanie, że „emocje zostaw za drzwiami”, to jeden z największych mitów biznesu.

Talk Description

Wnioski z godzin audytów stron klientów, tygodni wysłuchiwania historii o „poprzednich wykonawcach” i miesięcy wprowadzania drobnych poprawek, które potrafią zrobić ogromną różnicę. Porozmawiamy o tym, co najczęściej psuje strony internetowe i co możemy zrobić, żeby: •⁠ ⁠poprawić wyniki biznesowe klientów, •⁠ ⁠uratować dobre imię deweloperów, •⁠ ⁠i przywrócić trochę zdrowego rozsądku po obu stronach.

Talk Description

PDFs are where great marketing goes to die — and yet most of us keep making them. We decided to stop. Instead, we used AI to transform our static materials into dynamic, interactive tools that turn "just trust us" into hard data people can actually explore. We built the whole thing ourselves, with prompts, no developers, in a fraction of the time. And then we had to maintain it. This talk is about both halves. The exciting part — using AI to build marketing that actually works — and the part nobody puts in the conference abstract: what happens after you ship. The security scan with 8 warnings. The codebase that became a black box. The very confident AI that left the door unlocked. I'll show you exactly how we did it — from auditing the assets that were silently failing, to using AI to convert qualitative content into quantitative proof points, to the unglamorous work of keeping it secure once it's live. No fluff. No theory. Just what worked, what broke, and what we'd do differently. Come for the AI tips. Stay for the PDF funeral.

Talk Description

The fundamentals of a CMS were built around three assumptions: humans author content, the system stores and publishes pages, and frontends are built once and rendered many times. By 2030, on a ride that has already started, all three will have inverted, and the architectural decisions we make now will determine whether our platforms travel with the shift or get left behind. Production inverts. It's already happening: bots draft and humans orchestrate. Quite very often, the author's job is no longer to "create" content, but to describe intent, voice, and acceptance criteria — then direct the agents who draft, fact-check, localize, and adapt. WYSIWYG gives way to WYDIWYG: (What You Describe Is What You Get...most of the time, at least). So, new CMS primitives are emerging: intent objects, brand contracts, agent identities, evals, and cost budgets per piece. Substrate inverts. What we used to call CMS becomes invisible. It becomes a semantic layer of authored and derived content (embeddings, summaries, translations, agent-ready chunks, voice variants), addressable by agents over MCP, where "publish" is continuous and the URL is one address among many. Most content in 2030 won't be authored; it will be derived. CMSs will become a materialization engine rather than a repository. Surface inverts. Frontends are synthesized on demand. Design tokens and component logic become runtime contracts that an agent assembles into an ephemeral UI per request — voice, AR, web fragment, embedded inside someone else's agent. The frontend stops being something you build and becomes something that's requested. This talk walks all three inversions as one connected architectural shift, with concrete patterns for each, what survives (the content model, governance, the design system as contract), what dies (the page, the template, the publish button, the URL as primary address), and what every CMS team should be doing now to prepare.

Talk Description

An exploration of the evolution of open source CMS platforms - from the early days of the web to today’s dominant solutions. This talk looks at once-popular systems that have faded away, the reasons behind their decline, and the key factors that allowed WordPress to rise and become the leading CMS. Along the way, we'll examine how technology choices, community, and ecosystem shaped the winners and losers of the CMS landscape.

Talk Description

Jakiś czas temu skończyłem duży projekt – napisałem skomplikowany tool do migracji danych w WordPressie. I wtedy usłyszałem: „Przechodzisz do supportu." Moja pierwsza myśl? Ja, senior developer, będę teraz aktualizował wtyczki? Przyjąłem to jednak z pokorą jako wyzwanie. I dobrze, bo rzeczywistość szybko zweryfikowała moje wyobrażenia o tym, czym support w ogóle jest. Dziś jestem Head of WLC Growth & Care – działu, który w White Label Coders odpowiada za rozwój i utrzymanie projektów WordPress. I mogę powiedzieć jedno: support to jedno z najtrudniejszych środowisk, w jakich pracowałem jako developer. W tym wystąpieniu opowiem o tym, czego nie widać z zewnątrz: dlaczego dobry support wymaga szerszej wiedzy niż niejeden projekt greenfield, jak debugowanie cudzego kodu w środowisku produkcyjnym – pod presją czasu i oczekiwań klienta – uczy precyzji i pokory jednocześnie, i dlaczego wyzwania komunikacyjne w tej pracy są równie złożone jak techniczne. Na konkretnych przypadkach z WLC – bez lukrowania.

Talk Description

Open source is democratic infrastructure: built in public, improved by many, and trusted because it is transparent. But infrastructure does not maintain itself. In this talk, Marcel Bootsman shares a practical playbook for how companies and individuals can support open source without taking control, reduce burnout risk, and turn good intentions into sustainable contribution.

Talk Description

Modern frontend architectures promise the best of both worlds: static performance and dynamic data. But in practice, many teams still rely on time-based revalidation or request-triggered regeneration, leading to stale content and unpredictable cache behavior. This talk introduces a publish-driven approach to static frontends, where data changes—not time—become the primary trigger for cache invalidation and regeneration. Using patterns built on top of Next.js (ISR, API routes, and on-demand revalidation), we’ll explore how to turn your frontend into an event-driven system. By leveraging webhooks from your CMS or backend, you can precisely control when content updates propagate, while still benefiting from CDN-level performance and cache-first delivery. We’ll walk through the architectural shift from time-based to event-based rendering, discuss trade-offs, and demonstrate how this approach generalizes beyond Next.js to any framework that supports static generation and server-side hooks.

Talk Description

In the era of AI and self-proclaimed marketing gurus, many creators are questioning the point of creating and selling online courses. The truth is far more complex though - there's still demand for online courses, but the market has changed. Learners’ expectations are much higher, both in terms of course quality and the way it is sold and delivered. This brings us to the choice of CMS. Many solutions are available on the market, and we're spoilt for choice. In my presentation, I will highlight the features and criteria to look at when choosing the best CMS to sell and deliver your online course. There is no “one size fits all”, and cost optimisation is just one of the points I will make.

Workshop Description

Everyone's convinced a few good prompts can build a real platform. In this workshop, we test that assumption by looking at the pieces we maybe shouldn't vibe-code. TYPO3 CMS already gives developers a rich set of APIs and features, shaped by decades of real-world use and security review. We'll develop a website with custom functionality together and see, example by example, just how much you can do — securely and maintainably — when you're building on an enterprise platform that's already battle-tested. Bring some PHP knowledge and a laptop with a DDEV installed.

Talk Description

Digital sovereignty is the biggest topic right now, but true digital sovereignty cannot happen without the freedoms of open source software. Let's look at what digital sovereignty really is and how you and your agency can become providers of true "digital sovereignty by design." We'll look at how digital sovereignty affects different CMS users depending on size and financial strength, and how choosing an open source CMS can be one of the keys to greater independence. How do you talk about digital sovereignty? Some voices make it seem like controlling your data is everything. We often hear this from proprietary vendors, whose only solution is to restrict the conversation, because digital sovereignty challenges their business model. Delivering digital sovereignty requires understanding how to cope with your own digital sovereignty challenges. In the future, you’ll be advising your clients and choose the right kind of pragmatism.

Workshop Description

In this two-hour workshop, you'll go from zero to a fully deployed, agent-powered CMS site, and learn how to extend it with your own plugins. We'll start by deploying an EmDash blog to production, covering configuration, content modeling, and the deployment pipeline. From there, we'll explore EmDash's agentic capabilities, using AI agents to configure and customize the site, and to generate, enrich, and manage content without manual busywork. Finally, we'll roll up our sleeves and build a custom plugin, covering the plugin lifecycle, APIs, and how to package and ship your extension.

Talk Description

WordPress powers 43% of the web but inherits a security model from 2003. When you install a WordPress plugin, you grant it unrestricted access to your database, filesystem, and user data. The result: 96% of WordPress security vulnerabilities originate in plugins. At Cloudflare, we asked: what would WordPress look like if built today? Our answer is EmDash — a full-stack TypeScript CMS powered by Astro and Cloudflare Workers. EmDash keeps what made WordPress successful (extensibility, plugin ecosystem, familiar admin UX) but rebuilds it on serverless, type-safe foundations. In this talk, I will explain how EmDash uses sandboxed Worker isolates to solve the fundamental security problem of plugin architectures, why Astro's content-driven approach makes it ideal for modern CMS frontends, and how we've designed EmDash to be AI-native from day one — complete with built-in MCP servers, agent skills, and x402 payments for the agentic web.

Talk Description

For years, a CMS solved a real problem: how to build fast, without writing everything from scratch. Mature ecosystems, proven solutions, predictable delivery costs. It made sense. Today, that equation is starting to look different. AI Assisted Engineering is changing the accessibility and cost of software development in ways that were simply unrealistic just a few years ago. Solutions built for a specific problem, without the overhead of functionality nobody needs. Sites that are faster, leaner, and easier to maintain – because they contain only what they were built to do. In this talk, I'll share what this looks like in everyday practice. Not as an argument against CMSs, but as context for a more deliberate decision: when an off-the-shelf solution still has the edge – and when it becomes a debt you'll be paying for years. This session is for tech leads and software house owners thinking about whether – and how – to evolve their teams and processes.

Talk Description

AI is already changing how we build for CMS platforms: generating code compatible with any CMS, drafting API integrations, and prototyping entire features faster than ever. But while AI can speed up early delivery, shipping that work to real clients is a different story. We’ll share real examples of AI-assisted projects: what worked, what broke, and what still needed a human. We’ll cover the second wave after the first draft (debug, refine, test) and the gap between AI promises and client expectations around time, cost, and “doing it cheaper”.

Workshop Description

A workshop where every attendee leaves with a working visual regression tool running locally on their own laptop. Add your client URLs, capture baselines, simulate updates, and watch the tool flag broken layouts in side-by-side comparisons. The full source code is yours to keep, extend, and ship into your agency workflow.

Talk Description

In recent releases, WordPress has started introducing small portions of real gems that will have an immense impact on the way we develop websites and think about essential aspects of them that go beyond the code, such as performance, user experience, developer experience, and maybe the most important of all, accessibility. Did you know that WordPress' HTML API powers Interactivity API? Did you know it is delightful and fun to develop with? And did you know that it can be essential for improving accessibility in every step of your development workflow? Milana can show you how.

Talk Description

LLMs are the most confident unreliable narrators we've ever put into production. They'll invent court cases in beautiful prose, cite papers that don't exist, and tell you the capital of Australia is Sydney with the same tone they use for things they actually know. The scary part isn't that they're wrong — it's that they're wrong in a voice we've spent a hundred thousand years learning to trust. We will explore bad sides of LLMs — hallucination, bias, silent failure, and the confidence trap — and what critical thinking actually looks like when your main source of information has started freestyling in complete sentences. Equal parts for the engineers wiring these things into products and the PMs writing the specs, with jokes at the model's expense and none at yours.

Talk Description

Ta prezentacja opowiada historię migracji edytora Zendesk Knowledge z wersji legacy do nowego edytora. To, co miało być stosunkowo prostym przejściem z programu Early Access do publicznego udostępnienia, przerodziło się w niemal dwuletni projekt realizowany przez zespoły inżynieryjne, produktowe i operacyjne. Prezentacja skupi się na najważniejszych decyzjach, napotkanych wyzwaniach, tym, co przeoczyliśmy, oraz wnioskach z wdrażania dużej migracji edytora w istniejącym produkcie.

Talk Description

Sprzedaż biletów jest trudniejsza, niż się wydaje. Gdy 200 osób klika "kup" w tej samej milisekundzie, każda warstwa stosu - baza danych, kanał real-time, zakładka w przeglądarce - musi podejmować decyzje, których normalnie nie musi. Zbudowaliśmy własny CMS w Laravelu pod Safe Tickets właśnie po to, żeby sobie z tym radzić. Nie dlatego, że gotowe platformy nie potrafią - ale dlatego, że poradzenie sobie z tym dobrze wymagało decyzji architektonicznych, które chcieliśmy podejmować sami. W tej prelekcji pokazuję trzy z nich - i nie udaję, że któraś jest idealna. Po pierwsze: jak serializujemy współbieżne zakupy na poziomie bazy, używając czterowarstwowego pattern'u idempotentności łączącego Cache::lock z UNIQUE constraint w DB jako finalnym fallbackiem, który przeżywa zarówno wyczyszczenie cache, jak i przerwanie locka. Po drugie: jak utrzymujemy spójny stan miejsc we wszystkich otwartych zakładkach przy użyciu Laravel Reverb - pięć typów eventów, kanały publiczne i prywatne, oraz świadome trade-offy przy broadcasting wewnątrz transakcji DB. Po trzecie: co się dzieje, gdy użytkownik zamknie zakładkę z zablokowanymi miejscami. Nie karzemy go - dajemy mu czas. sendBeacon, dwuminutowy grace period i state restore, który przeżywa pełne odświeżenie strony. To nie jest prelekcja "zbudowaliśmy idealny system". To prelekcja "oto trade-offy, które zrobiliśmy, oto co zmienilibyśmy, oto co zrobilibyśmy znowu".

Talk Description

What does it look like when a CMS is best placed to be extended by AI and used to build agents and rich integrations. MCP API and embeddings allow you to interface with your data in some new and creative ways without removing your guardrails or reinventing the wheel.

Talk Description

For fifteen years, the CMS industry has built tools for one job: publishing content to a website. Meanwhile, every enterprise has quietly accumulated a parallel content universe - webinar recordings, internal knowledge-sharing sessions, sales calls, support tickets, training materials, customer Q&As. This knowledge sits in silos. It rarely makes it to the website. It almost never trains a chatbot, feeds a sales enablement system, or becomes a learning module - even when it should. AI just changed the economics of turning that knowledge into anything an enterprise needs: an article, a chatbot dataset, a knowledge base, a sales deck, a training course. The bottleneck is no longer creation. It's orchestration - knowing what content exists, where it sits, what context it carries, and how to route it to the right audience in the right format. This is not what today's CMS does. This talk argues that "content management" is about to be redefined: from a publishing tool into the operational backbone of how enterprises capture, contextualize, and deploy their own knowledge.

Talk Description

A CMS that naturally supports accessibility is the difference between a sustainable architecture and a continuous battle against your own tech stack. With the European Accessibility Act (EAA) elevating digital inclusion, web accessibility has evolved from a simple frontend developer's checklist into a core architectural strategy. We can craft WCAG 2.2-compliant templates, but true success happens when the CMS empowers content editors to maintain that standard effortlessly. In this session, we’ll embark on a journey to discover the true accessibility potential of today's popular CMS platforms. We will explore the unique strengths, built-in guardrails, and integration paths of traditional monoliths (like WordPress and Drupal), agile SaaS builders, and modern Headless architectures. Together, we will build The CMS A11y Matrix. Join me to uncover how choosing the right tech stack can streamline your compliance, boost your team's efficiency, and help you deliver a truly inclusive, high-performance web experience.

Talk Description

WordPress can do quite a lot. But how are other plugins, applications, or AI agents supposed to know what it can do - and how to make use of those abilities? That's where the WordPress Abilities API comes riding to the rescue - a standardized way to define, discover, and expose what WordPress can do. In this talk, we'll explore what Abilities are, why they were introduced, and how to use them in practice. We'll build our own Ability, define its input and output, set permissions, and implement the execution logic. Expect code and practical examples! We'll also dive into REST APIs, automation, and AI to see why there's so much buzz around Abilities right now.

Talk Description

After onboarding over 6 million users at WPBakery and teaching hundreds of students, one insight stands out: most CMS-based companies focus on onboarding users into their products, not the CMS itself. This session explores why educating users about the broader ecosystem is critical for growth and how shifting onboarding beyond your product can expand your market and create more successful customers.

Talk Description

They didn't guess the password. They didn't break 2FA, bypass the firewall, or trigger a single alert. The activity log shows no failed attempts, no unusual login times, no new accounts created. And yet, the attacker is inside your WordPress site, accessing your WordPress dashboard. Session hijacking leaves no login trace, because the attacker never logged in. They took an authentication cookie and reused an existing session. To WordPress, every request looks legitimate; the real user's credentials, their 2FA, their login restrictions, none of it applies. The session was already authenticated. Most WordPress security advice is built around hardening and protecting, however, that leaves a critical blind spot: what happens after authentication succeeds? In this talk, we'll walk through how WordPress creates and validates authentication cookies and session tokens, explore the real-world scenarios where sessions get stolen, and map out what an attacker can do with a hijacked session before the damage becomes visible. Then we'll get practical. We will look into how session visibility, timeout policies, concurrent login controls, device recognition, and activity logging that captures post-login behaviour are the controls that catch what firewalls and 2FA miss. Firewalls, 2FA, and passkeys are essential. This talk covers what comes after them.

Talk Description

We often evaluate CMS platforms by features, integrations, and publishing workflows. But the strongest CMS ecosystems are built on something less visible: community. In this talk, I’ll share why conferences, meetups, educational content, and user communities are often the real drivers of adoption, retention, and growth. Drawing from years inside the WooCommerce and WordPress ecosystem—and organizing industry events myself—I’ll show why software alone is never enough, and what content professionals, agencies, and CMS teams can do to build stronger ecosystems around their platforms.

Talk Description

Vibe coding has found its way into CMS workflows. Extensions and Designs are prompted into existence, or assembled from loose suggestions and thoughts. Intentions turn into somewhat shippable products without a full read-through. That's not necessarily a problem. But it does shift something in how responsibility is distributed, and that shift hasn't been talked about much yet. This talk comes from looking at that question from three distinct angles. As a product person, a developer, and a security company. From watching how AI-assisted development interacts with CMS security, and where the gaps tend to appear. Let's take a look at how shit can hit the fan and more importantly how to avoid it.

Talk Description

The old King lies on his deathbed. And before his last breath, the game for the throne has already begun. A centuries-old heritage luxury brand hangs in the balance. The web of intrigue spreads fast, and three factions have emerged — each with a claim, each with a price. The Heir: rightful, proven, cast from the same mold as the dying king — but is continuity enough? The Brother: revolutionary, restless, offering a clean break from the old ways — but at what cost? The Pretender from the North: an outsider, here to shatter the old order entirely and rewrite the rules of the realm. Strengths and weaknesses. Alliances and betrayals. And only one throne. Who will inherit the empire?

Talk Description

Most people who use a CMS do not want to use a CMS. They want the page right, the post out, whatever they came to do, done. They want content that converts, that earns its place, that turns into value. The system we built to help them get there is fixed: one shape, drawn once and shipped to everyone, and it is starting to show its age. It is not just content anymore. It has to function across the work the business actually does, including commerce, customer data, and the systems around them. The CMS the human actually needs is alive. It is shaped to the person, the action, and the place the work is happening. It travels with them, into chat, into email, into the tool they were already in. It brings only what the moment requires. What we need today is not screens. It is the right experience in the right format, and the right format might be voice. Agentic architecture is what makes that possible, and it asks something different of the system underneath: a content model an agent can reason about, and capabilities that know the difference between what is safe to do automatically and what needs a human to approve it. The human stays in the loop on the things that matter: brand voice, factual claims, and the decision to publish.

Talk Description

Sometimes we think that modern technology will keep us safe online. But sometimes the human factor can "break down" even the best security solutions. I'll show what to look for to prevent big money from being stolen, and at the same time, how to give your hosting server administrators peace of mind, so they don't need to explain the theft methods to the police again. And a strong password is only the beginning...

Talk Description

UX content (microcopy) – field labels, validation messages, empty states – may look simple. It doesn't. When you try to manage it through a classic CMS, something breaks. Not because the tools are bad, but because they were built on a different model of what content is. Classic CMS architecture treats content as an autonomous object with metadata. Design systems treat content as a context-dependent element: its meaning, constraints, and validity are defined by its position in a component hierarchy. These are two different ontologies that no amount of configuration bridges that gap cleanly. This talk uses UX content as a stress test for CMS architecture. We'll examine: * where the model breaks, * why translation management systems have accidentally become the closest thing to a UX content layer, * what it costs teams to work around a structural mismatch – including encoding design hierarchy into key names just to know what a string is for. The takeaway is a model for the synchronization layer that doesn't yet exist as a product, but whose requirements we can now name precisely.