Senior Field Marketing Manager

Accessibility checks now happen inside your AI tools. Siteimprove's new MCP Server brings real-time accessibility auditing into Claude, Figma, and VS Code - catching problems before they ever go live. The problem with fixing accessibility after the fact. Picture a web team that spends three weeks building a new product page. They launch it, feel great, then run an accessibility audit. The results come back with dozens of issues - missing alt text, poor color contrast, unlabeled form fields. Now someone has to go back and fix everything. This is how most teams have handled accessibility for years. Build first, audit later. It's slow, expensive, and frustrating for everyone involved. Siteimprove is trying to flip that model entirely. Its new MCP Server doesn't wait until after you publish. It checks your work while you're still making it. What is an MCP Server and why does it matter? MCP stands for Model Context Protocol. Think of it as a universal connector that lets AI tools talk to each other and share context in real time. Without something like MCP, your design tool, your code editor, and your AI assistant all live in separate bubbles. They don't know what the other is doing. MCP breaks down those walls. Siteimprove's MCP Server plugs its Accessibility Agent directly into tools like Anthropic's Claude, the Figma design platform, and VS Code. When you're working inside any of these environments, the agent is working alongside you. It's not waiting for you to export a file or run a separate scan. It's watching, checking, and flagging issues as they appear. This kind of real-time, embedded checking is a big shift from how accessibility tools have traditionally worked. Most accessibility platforms operate as standalone dashboards you visit after publishing. The MCP approach turns accessibility into part of the creation process itself. How it works across different tools. The Siteimprove Accessibility Agent behaves a little differently depending on which tool you're using - and that's by design. For designers using Figma. Siteimprove built a dedicated Figma plug-in that surfaces accessibility issues while a designer is still working on a mockup. If a color combination doesn't meet contrast requirements, the designer sees that immediately - not after a developer has already coded it up. The plug-in also runs a color blindness simulation with screenshot capture. A designer can see exactly how their work looks to someone with deuteranopia or protanopia before a single line of code gets written. Audit reports can be saved directly into the Figma canvas, so accessibility findings become part of the design file itself rather than living in a separate document nobody reads. For developers using VS Code. Developers working in VS Code get accessibility feedback woven into their coding environment. If an AI coding assistant generates a component with a missing ARIA label or an inaccessible interactive element, the Accessibility Agent can flag it before the developer even saves the file. This is especially valuable as more developers lean on AI code generation. AI tools can write code fast, but they don't always write accessible code. Having an accessibility layer that reviews AI-generated output in real time closes a gap that's growing larger as AI coding assistants become mainstream. For AI builders using Claude. Claude users - whether they're marketers drafting content, builders creating digital experiences, or developers testing prompts - can now have the Accessibility Agent running as part of their workflow. The agent-to-agent capability is particularly interesting here. Imagine an AI coding agent generating an entire landing page. Before that page ever hits production, the Siteimprove Accessibility Agent can autonomously audit it and apply fixes. No human has to manually review every element. The agents handle it between themselves. Siteimprove's MCP Server connects to over 40 partner integrations, which means this kind of agent-to-agent workflow can extend across a wide range of tools in a typical enterprise tech stack. Two forces pushing this change right now. This isn't just a product launch happening in a vacuum. Two real pressures are making real-time accessibility tools more urgent than ever before. The European Accessibility Act is now in effect. The European Accessibility Act took effect in June 2025. It requires businesses operating in Europe to make their digital products and services accessible. Non-compliance carries real penalties. For Siteimprove's customer base - which includes Global 2000 companies like Barclays, Shell, BlackRock, Harvard, and GSK - this isn't optional. These organizations need to demonstrate ongoing compliance, not just pass a one-time audit. Embedding accessibility checks into the creation workflow makes compliance much easier to maintain at scale. The regulation also created urgency for smaller businesses serving European markets. If you're building digital products for European users, accessibility is now a legal requirement, not a nice-to-have. AI is changing who gets to build. Siteimprove CEO Nayaki Nayyar put it directly: "AI has put the power to build digital experiences in the hands of everyone." That's true - and it creates a new accessibility challenge. When only trained developers built websites, you could enforce accessibility standards through code reviews and established workflows. Now marketers, product managers, and non-technical creators are building digital experiences using AI tools. They may not know WCAG guidelines. They may not know what an ARIA label is. Embedding accessibility checks into the tools these new creators are already using is the only practical way to maintain quality. You can't train everyone on accessibility standards overnight. But you can put smart guardrails inside the tools they're already using. The bigger shift: from audit tool to intelligence platform. Siteimprove's move here is part of a broader repositioning. The company started as a web accessibility and SEO point solution. It's now building toward what it calls an "agentic content intelligence" platform. That platform bundles accessibility, analytics, search optimization, and orchestration agents into one connected system. Recent additions include a Content Writing and Optimization Agent and a PDF Remediation Agent that automatically tags documents to meet WCAG and PDF-UA standards. The MCP Server launch extends this vision into the tools where content actually gets made. Instead of being a platform you check in on, Siteimprove wants to be a layer that runs underneath everything you create. This approach mirrors what's happening across the SaaS industry more broadly. Standalone tools are giving way to embedded intelligence. The most valuable software isn't the app you open separately - it's the capability that shows up inside the tools you're already using. What teams should think about before adopting this. Real-time accessibility checking inside creation tools sounds great. But there are practical questions worth thinking through before rolling this out across a team. First, consider workflow disruption. Accessibility flags that appear constantly while someone is mid-design can feel like interruptions. The best implementations of this kind of real-time feedback let users set when and how they see alerts. Teams should check whether the tool's notification behavior is configurable. Second, automated checks don't catch everything. Tools like Siteimprove's agent are excellent at catching technical issues - color contrast ratios, missing labels, structural problems. But some accessibility issues require human judgment. Does the image description actually convey meaning? Is the reading order logical for a screen reader user? Automated agents are a powerful first line of defense, not a complete solution. Third, think about who owns the fixes. When an agent flags an issue during design, does the designer fix it? Does it get logged for a developer? Does the agent fix it automatically? Teams need clear ownership rules, or flagged issues will fall through the cracks even with great tooling in place. Why catching issues early changes everything. There's a well-established principle in software development: the earlier you catch a bug, the cheaper it is to fix. The same logic applies to accessibility. An accessibility issue caught during design takes minutes to fix. The same issue caught after development might take hours. Caught after launch, it could require coordination across multiple teams, a content freeze, and a re-deployment cycle. The cost multiplies at every stage. Shifting accessibility checks upstream - into the moment of creation - isn't just about compliance. It's about making the entire content production process more efficient. Teams spend less time in rework cycles. Developers don't have to undo design decisions. Marketers don't have to delay campaigns while accessibility issues get resolved. The Siteimprove MCP Server is a practical attempt to make that upstream shift real inside the tools teams are already using. Whether it becomes a standard part of how digital teams work will depend on how well it fits into existing workflows without adding friction. But the direction it's pointing - accessibility as a built-in layer rather than a bolt-on audit - is clearly where the industry is heading. #SaaS & Tech Trends #GZOO #BusinessAutomation

Siteimprove has launched its Model Context Protocol server, integrating its accessibility agent into Anthropic Claude, Lovable, VS Code and Figma. The system automatically scans, detects and remediates accessibility issues as digital content is created, rather than flagging problems after publication. The Copenhagen-based company, founded in 2003, aims to address the surge in inaccessible digital content produced by AI-powered creation tools. The MCP server extends connectivity across Siteimprove's existing 40-plus partner integrations, enabling agent-to-agent workflows where AI-generated experiences are automatically audited before reaching production. A new Figma plug-in surfaces accessibility issues during the design phase, whilst allowing designers to save audit reports and explore colour blindness scenarios directly within the canvas. The platform also includes search and content agents for AI discoverability optimisation.

Siteimprove has launched two AI agents to enhance its content intelligence platform: the Content Writing & Optimization Agent and the PDF Remediation Agent. The announcement was made at Gartner Marketing Symposium in Denver. The Content Writing & Optimization Agent helps organisations create content at scale whilst maintaining brand standards and incorporating Answer Engine Optimisation insights. It supports AI-assisted outlining, drafting and optimisation workflows. The PDF Remediation Agent automatically fixes accessibility issues by auto-tagging PDFs, adding document structure and correcting heading hierarchies to meet WCAG and PDF-UA standards. Siteimprove also offers expert-led and hybrid remediation options for complex needs. Founded in 2003, Siteimprove provides a unified platform combining accessibility, analytics, SEO/AEO and content strategy for enterprise clients across manufacturing, government, healthcare and education sectors.