Nathan Wrigley hosts Miriam Schwab on the Jukebox Podcast from WP Tavern to discuss her long WordPress history, Elementor’s growth, and how AI is shaping the platform’s future.
Miriam Schwab’s WordPress story spans roughly two decades. She began offering WordPress as a service, ran a custom WordPress agency for tech companies and nonprofits, then founded Strattic, a startup delivering static WordPress architecture with retained dynamic capabilities. Strattic solved speed, security, and scalability issues, raised venture capital, and was acquired by Elementor in June 2022. After leading Strattic’s team, Miriam took on a role now called Head of WordPress at Elementor, acting as the liaison between Elementor and the wider WordPress community across strategy, product, community, and partnerships.
Elementor is approaching its 10th anniversary in June and has seen rapid growth from day one. The free version played a large role in adoption by offering a compelling product at the right time. According to W3Techs figures cited by Miriam, Elementor’s share of the entire web grew from about 11.7% to 13.1% over 2025 — meaning it now powers over 13% of websites. In 2025 W3Techs named Elementor the top “CMS” by absolute site growth, ahead of platforms that previously dominated those lists.
With that scale comes responsibility: backwards compatibility and careful QA are critical. WordPress sites are highly diverse in themes, plugins, server setups, and PHP versions, so every release must be tested widely. Elementor v4 represents a significant overhaul of a decade-old codebase. Migration between versions will need thoughtful approaches; existing sites may continue running older versions while new content uses v4, so the two approaches will coexist for a period.
Elementor has grown to roughly 350 staff, enabling larger engineering efforts, QA, and product development to support big updates like v4.
AI has been a major focus in 2025 and into 2026. Miriam says Elementor pursues multiple AI approaches:
– In-editor AI: Generating images and copy directly inside the Elementor editor for typical content tasks (titles, body text, images).
– Angie: A standalone, free WordPress plugin applying agentic AI capabilities across WordPress, not limited to Elementor. Angie exposes tools to manage sites — create posts, manage users, change categories, handle WooCommerce and products, and more. The team built about 200 tools and also added support for WooCommerce and ACF to give Angie broad reach. Angie includes guardrails and safety checks and is available for users to install and test.
– Site Planner: A conversational AI that guides users from idea to a wireframe website. Rather than a single prompt, it asks targeted questions, creates a brief, produces a sitemap with content hints, and generates wireframes (about 80–90% complete). Users can tweak structure, drag-and-drop pages, change palettes, upload logos, and export to hosting — Elementor’s hosting, a connected account, or as a ZIP. Site Planner is aimed at beginners and professionals alike, accelerating research and client sign-off by moving past the blank-canvas stage.
Integration with WordPress’s new Abilities API is a key area. Angie was developed before the Abilities API was finalized, so the team built its own system to expose WordPress capabilities. That work is substantial and, in many cases, covers more than the Abilities API initially offered. The plan is to sync with and leverage the Abilities API as it matures. Miriam praises the Abilities API as a smart standardization that makes agentic AI easier to interface with a 20+ year legacy platform like WordPress, preventing fragmented plugin-specific approaches and helping ensure long-term compatibility.
Miriam emphasizes that agentic AI for WordPress is now realistic and valuable. When AI can act on site-level functions — create, publish, manage users, adjust WooCommerce products — it expands beyond in-editor content tools into full site management and workflows. That shift opens many possibilities but also complexity in debugging and support when multiple AI-driven plugins interact.
Site Planner’s conversational flow: you can begin with a short prompt, but the Planner will ask clarifying and suggestive questions, producing a “brief” and scoring its strength. It then creates a site map indicating page hierarchy and content chunks. The wireframe stage assembles page structures (testimonials, galleries, events, forms) tied to the content generated from the conversation. The output is editable and can be published to hosting or exported. Currently Site Planner projects are private until exported or published; the team plans deeper integration with Elementor, enabling easier duplication and regeneration within the environment.
Design granularity and future AI capabilities: today’s tools can assemble structures and content; Elementor v4 and future integrations aim to let AI exercise much more granular creative control — custom code snippets, widgets, and fine-grained styling decisions. Miriam envisions AI acting like a design collaborator where a user can request small iterative changes (swap an image, tweak padding, add box shadow) conversationally, and the system implements them. That trajectory suggests creators will shift from manual tinkering to directing and producing outcomes via AI.
Miriam reflects on professional roles changing but remaining crucial. AI accelerates work and elevates quality, but humans must still guide direction. The “human-in-the-loop” will define what excellent websites are — strategy, UX judgment, and domain expertise remain valuable. AI makes many tasks faster and higher quality, but skill will pivot from execution details to how well someone directs AI and defines outcomes.
Support and safety: AI tools require extensive guardrails. Miriam recounts that early Angie use could remove a user; the team added protections to prevent destructive actions without clear user confirmation. Support workflows will change: AI can generate logs and help analyze issues across knowledge bases, but debugging interactions among multiple AI agents and plugins will be a new challenge. Determining whether a problem was caused by Elementor, Angie, WordPress core, or a third-party plugin will be a more complex variant of plugin conflict resolution.
Accessibility: AI is not a magic bullet. Miriam notes that telling AI “make my site accessible” isn’t sufficient; AI needs proper tools, direction, and measurable standards. Elementor built an accessibility plugin, Ally, which scans sites, reports issues, suggests fixes, and can use AI-assisted implementation. Ally aims to improve accessibility continuously and can help site owners address issues iteratively, though total compliance guarantees remain complex.
Looking ahead, Miriam is excited about 2026: AI has reignited innovation in WordPress, making new capabilities more accessible. She expects significant progress and many creative releases throughout the year from Elementor and the wider ecosystem. While the landscape will change rapidly and present challenges for builders, support teams, and professionals, she believes AI will provide powerful tools that, when guided well by humans, elevate the quality and reach of WordPress sites.
Nathan and Miriam close acknowledging the uncertain, fast-moving nature of AI and its potential to reshape WordPress development, design, and management while underscoring the continuing importance of human judgment and strategy.