Every update, improvement, and new feature — in order.
Two timely posts plus an accuracy update. (1) 'Want the Presumption of Conformity? You Have Until July 22 to Sign the Code of Practice' — to make the published list of initial Code of Practice signatories and gain a presumption of Article 50 conformity, providers and deployers must submit by July 22, 2026, 18:00 CEST. Covers what the presumption means, who can sign, and how to decide. (2) 'The EU Is Softening the AI Literacy Rule. Here's Why You Shouldn't Stop Training Yet' — the Digital Omnibus rewrites Article 4 from 'ensure' to 'take measures to support' AI literacy, but it's a political agreement, not yet law, so the current obligation still binds. Breaks down the two opposite misreads. (3) Accuracy update: the existing Article 4 AI literacy post now includes a section on the pending Omnibus softening with correct 'agreed but not adopted' framing, and the June 16 Code of Practice post now flags the July 22 signing deadline. Blog count now at 53.
Search Console showed several blog posts ranking on page 2 with strong impressions but few clicks — a position problem, not a copy problem. Two of the highest-impression posts (the open-source AI exemptions guide and the EU database registration guide) had zero internal links pointing to them, leaving them orphaned. Added contextual in-body links and related-article links from the most topically relevant posts: the GPAI and classification guides now link to open-source exemptions, and the conformity assessment and compliance checklist guides now link to database registration. More internal link equity signals importance to Google and helps these near-winners climb from page 2 to page 1.
Two new guides. (1) 'EU AI Act for Insurance: Why Life and Health Pricing AI Is High-Risk by Name' — the companion post for the /for/insurance page. Covers why Annex III point 5(c) names life and health insurance directly, which systems are in scope versus out (P&C, motor telematics, internal actuarial models), the separate creditworthiness trigger for premium financing, and how Solvency II model governance overlaps with Articles 9 and 10. (2) 'There's a Free EU AI Act Regulatory Sandbox Coming, and Most Startups Don't Know It Exists' — Article 57 requires every member state to run an AI regulatory sandbox by August 2, 2026, with free priority access for SMEs and startups. Covers what a sandbox is, who qualifies, the written guidance and exit report you get, and when it's worth applying. Cross-linked from the startups post and the insurance industry page. Blog count now at 51.
Two timely posts covering the EU AI Act transparency rules. (1) 'The Final Code of Practice on AI Content Marking Is Here — What's Actually In It' — the Commission published the final Code of Practice on June 10, 2026. Covers the two sections (marking/detection for providers, deepfake and public-interest-text labelling for deployers), the layered metadata-plus-watermarking approach, why the voluntary Code still matters for demonstrating Article 50 compliance, and what to do before August 2. (2) 'The Article 50 Watermarking Grace Period Everyone's Misreading' — the Digital Omnibus deferred only Article 50(2) machine-readable marking to December 2, 2026, and only for systems already on the market. The disclosure obligations (chatbot notices, deepfake labelling) still apply August 2. Breaks down the two opposite ways people misread the grace period. Both posts cross-linked from the existing Article 50 cluster. Blog count now at 49.
Three conversion improvements. (1) Newsletter subscribers now get a welcome email immediately: the two deadlines that matter, three starter guides, and a link to the free classifier. Previously, subscribing produced silence. (2) New public page at /sample-report — real ActReady output for a fictional CV-screening startup: the high-risk classification result with Annex III reasoning, the 10-obligation tracker, and an excerpt from a generated Article 14 Human Oversight Plan. We don't have customer logos yet, so we show the actual product instead. Linked from the homepage and footer. (3) Pricing page now explains why ActReady costs a fraction of alternatives: most EU AI Act tools are enterprise GRC platforms at €130K+/year or €299+/month behind a sales call. ActReady is self-serve software for SMBs — and we say plainly when you'd still want specialist counsel instead.
Three improvements. (1) Funnel event tracking added — classifier starts and completions (with risk tier), newsletter signups, email popup conversions, and checkout starts are now measured via Vercel Analytics. Events respect cookie consent: nothing is tracked unless the visitor accepted. (2) llms.txt published at /llms.txt — a machine-readable index of all 47 guides, tools, and industry pages, generated automatically from the blog data so it never goes stale. Helps AI assistants like ChatGPT, Claude, and Perplexity find and cite our content when people ask EU AI Act questions. (3) Fixed 5 stale tests that still referenced the comparison table removed in v2.56, and replaced a banned phrase in the blog newsletter signup heading.
Two new vertical landing pages. (1) /for/insurance — life and health insurance risk assessment and pricing AI is named directly in Annex III point 5(c), making insurance one of the few industries that's high-risk by name rather than interpretation. Covers what's in scope, what's not (P&C pricing, actuarial modeling), the 10 mandatory obligations, and how Solvency II model governance overlaps. (2) /for/customer-service — support chatbots, AI agents, AI-drafted replies, and voice bots all trigger Article 50 transparency obligations on August 2, 2026, making customer service teams among the first in any company to face the EU AI Act. Covers disclosure requirements, what limited-risk does NOT require, and deployer responsibilities when using platforms like Intercom or Zendesk. Both pages added to the sitemap and footer.
Four improvements. (1) www.getactready.com now 301 redirects to getactready.com — eliminates duplicate content for SEO. (2) Internal cross-linking expanded — 7 existing blog posts now link to the 3 new posts via related articles, helping Google discover new content from already-indexed pages. (3) New blog post: 'EU AI Act Instructions for Use: What High-Risk AI Documentation Actually Needs to Say' — covers all 12 mandatory elements from Article 13(3), who the audience is, common mistakes, and a practical 10-section template structure. Blog count now at 47. (4) Dynamic Open Graph images — every blog post now generates a unique OG image with the post title, category badge, and ActReady branding. LinkedIn, Twitter, Reddit, and Slack previews now show distinct images per post instead of the generic logo.
Two new blog posts. (1) 'EU AI Act Article 4: AI Literacy Training Is Already Mandatory' — covers the obligation that's been in force since February 2, 2025 and that most companies are ignoring: structured AI literacy training for all staff involved in operating AI systems. Practical approach, what to cover by role, and how to document compliance. (2) 'EU AI Act and AI Agents: When Autonomous Systems Trigger Compliance Obligations' — where AI agents fit under the EU AI Act's definitions, why autonomy makes compliance harder (not easier), when agents become high-risk under Annex III, the human oversight challenge for autonomous systems, and multi-agent system liability questions. Blog count now at 46 published guides.
Batch of under-the-hood improvements. (1) Auth protection expanded — all authenticated API routes now require Clerk auth at the proxy level, not just dashboard pages. (2) Blog index converted to a server component for better SEO — search and filters still work client-side. Blog post data extracted to a shared module so the sitemap and blog index stay in sync automatically. (3) Sitemap now uses real last-modified dates per page instead of today's date on every crawl. (4) Content Security Policy header added — whitelists Clerk, Stripe, Supabase, and Vercel Analytics; blocks everything else. (5) Stripe webhook now properly downgrades users to free when a subscription goes past due or is cancelled. (6) Dashboard error boundary added — a single bad API response no longer crashes the entire dashboard. (7) Email signup popup now shows an error state instead of falsely claiming success when the API call fails.
Replaced the deadline-only countdown bar with trust signals that are actually verifiable: link to the official EUR-Lex regulation text, the transparency deadline date with days remaining, and the live count of free compliance guides published on the blog.
Two new blog posts shipped. (1) 'The EU AI Act Transparency Deadline Is 63 Days Away and Nobody's Ready' — covers what actually happens on August 2, who's affected, why most companies haven't started, and a practical checklist. (2) 'Your Chatbot Needs an AI Disclosure. Here's Exactly What to Write.' — specific copy-paste disclosure examples for chatbots, support tools, content generators, and internal AI tools, plus common mistakes to avoid.
Completed the i18n wiring that was pending from v2.53. The translated strings for Regulatory Alerts (16 strings), AI Training (14 strings), and Vendor Risk (6 strings) are now actually used in the page components via useI18n() and t() calls. Page chrome — titles, subtitles, progress labels, button text, empty states, footers — now renders in all 7 supported languages when the user changes their language setting.
Major landing page rework to remove patterns common to AI-generated SaaS sites. Removed: 'Three steps to compliance' numbered icon section, 'Featured on' logo bar (Product Hunt, PeerPush, LinkedIn), comparison table with checkmarks (ActReady vs Consultant), three-icon trust signal row, 'Everything you need' headings, 'Whether you're...' copy. Replaced with: deadline status bar showing real countdown data, prose-style 'Here's the thing about the EU AI Act' explainer section, simple two-card cost comparison, and direct 'What ActReady actually does' heading. Also cleaned banned phrases across features page, FAQ, changelog, legal-tech industry page, and the /eu-ai-act-compliance-tool landing page. Writing style guide saved to project memory for future sessions.
Multiple SEO improvements. (1) New landing page at /eu-ai-act-compliance-tool targeting 'eu ai act compliance tool' keyword — includes feature list, enforcement timeline, how-it-works section, pricing CTA, and related blog links. (2) SiteNavigationSchema added to homepage structured data with 6 key pages described for Google sitelink eligibility. WebSiteSchema upgraded with SearchAction and alternate names. (3) Google Search Console audit revealed only 2 of 62 pages were indexed — 5 pages 'crawled but not indexed', 2 redirect errors. User manually submitted 20 priority pages for indexing across two days. (4) Sitemap updated with new page URLs and missing blog slugs.
Major SEO upgrade to the free risk classifier page (/classify). (1) Rich metadata — title, description, and keywords optimized for search queries like 'EU AI Act risk classifier', 'AI risk classification tool', and 'EU AI Act compliance tool'. Open Graph and Twitter card tags updated. (2) JSON-LD structured data — WebApplication schema for Google rich results plus FAQPage schema with 5 common questions for featured snippet eligibility. (3) SEO hero section — trust badges ('Free tool', 'Updated for 2026 deadlines'), expanded description with keyword-rich copy, and three feature highlights (4 risk tiers, 60-second completion, instant obligations). (4) Educational content below the classifier — 'How the EU AI Act Risk Classification Works' section explaining all four risk tiers with deadlines, enforcement timeline with visual status indicators, and an interactive FAQ accordion covering classification, deadlines, extraterritorial reach, pricing, and fines. (5) Bottom CTA section driving visitors to sign up for the full compliance dashboard. (6) Sitemap updated with the latest blog post slug.
Three features shipped: (1) New blog post — 'EU AI Act Article 50: The Multi-Layered Marking Strategy You Need Before August 2026' covering the Commission's draft guidelines on metadata marking, watermarking, and content fingerprinting with a practical action plan. (2) Command Palette (Cmd+K / Ctrl+K) — instant search and navigation across all dashboard pages. Type to filter, arrow keys to navigate, Enter to open. Supports fuzzy matching on page names, descriptions, and keywords. Fully i18n-aware. (3) i18n expansion — added 60+ translated strings for Regulatory Alerts, AI Training, and Vendor Risk pages across all 7 languages (EN, DE, FR, ES, IT, NL, PT). Page chrome (titles, buttons, labels, empty/error states) is now translatable on these pages.
Major update to keep ActReady current with the latest EU AI Act developments. (1) Regulatory Alerts — 4 new entries: two new Article 5 prohibited practices (non-consensual intimate AI imagery and AI-generated CSAM, effective December 2, 2026), the Annex I product-regulated AI deadline extension to August 2, 2028, compressed legacy system grace period (now 3 months, legacy systems must comply by December 2, 2026), and the Commission's draft Article 50 transparency guidelines detailing the multi-layered marking strategy (metadata + watermarking + fingerprinting). Updated the August 2026 enforcement alert to correctly reflect that only transparency and GPAI obligations apply on that date (Annex III high-risk was extended). (2) Risk Classifier — added the two new prohibited practices to the Article 5 step in the guided wizard, so systems generating non-consensual intimate content or CSAM are immediately flagged as prohibited. (3) Deadlines — added the Annex I August 2, 2028 deadline constant to the deadline system and countdown hook.
Major internationalization push: (1) Portuguese (Portugues) added as the 7th supported language — full translation dictionary covering all 190+ UI strings, flag picker updated in Settings. (2) Dashboard home page fully internationalized — all stat card labels, onboarding steps, deadline banners, compliance progress section, quick actions menu, action items, and recent activity headers now use translated strings across all 7 languages. (3) Compliance Tracker page fully translated — section headers, progress summary, obligation card controls (mark as, notes, assignments, status buttons), empty/error states, deadline text, and guide labels all wired to i18n. (4) AI Systems page fully translated — page title, add system form (all labels, hints, placeholders), empty/error states, delete confirmation modal, and system card labels. (5) Documents page fully translated — title, generate button, search placeholder, filter labels, empty/error states. (6) Weekly digest cron endpoint hardened — CRON_SECRET environment variable set in production, auth check now rejects requests when the secret is missing.
The entire dashboard interface now supports 6 languages: English, German (Deutsch), French (Français), Spanish (Español), Italian (Italiano), and Dutch (Nederlands). (1) Language selector in Settings — flag-based grid picker, instantly switches the UI, auto-detects browser language on first visit, persists choice to localStorage. (2) Full i18n system — 120+ translated strings covering sidebar navigation, resource links, deadline countdown, page titles, team management, notifications, settings labels, tracker statuses, risk levels, and common actions. (3) Lightweight architecture — React Context + dictionary approach, no heavy i18n libraries, type-safe translation keys with TypeScript, lazy-loaded per locale. (4) Sidebar fully translated — all 17 nav items, 4 resource links, deadline text, and theme label switch language in real time.
Full Slack webhook integration: (1) Settings UI — paste your Slack incoming webhook URL, test the connection with a single click, enable/disable with a toggle. Guide link to Slack webhook setup. (2) Rich Block Kit messages for 5 event types: obligation assignments (shows assigner, assignee, article, system, with View Tracker button), status changes (done/in-progress with emoji indicators), new AI system registrations (risk level colour dots, domain info), team member additions (role, email), and weekly compliance digest (grade, score, progress bar, deadline countdowns, Open Dashboard button). (3) Wired into all existing APIs — systems, team, obligations, and weekly cron. All Slack calls are fire-and-forget so they never slow down the UI. (4) Test webhook endpoint (/api/slack/test) validates the URL format and sends a confirmation message to the channel.
Full notification system: (1) Notification Center (/dashboard/notifications) — in-app feed of all compliance events including obligation assignments, status changes, deadline warnings, and team updates. Mark individual or all notifications as read, delete notifications, relative timestamps, and type-coded icons. Live unread count badge on sidebar. (2) Assignment email alerts — when you assign an obligation to a team member, they get an email with the obligation details, article reference, system name, and a direct link to the compliance tracker. (3) Weekly compliance digest emails — every Monday at 9 AM UTC, org owners receive a summary email showing compliance score, grade, obligation progress bar, deadline countdowns with urgency colours, and recent activity. Powered by Vercel Cron. (4) Notification preferences in Settings — three toggles (assignment alerts, deadline reminders, weekly digest) that control which emails are sent. All enabled by default, instantly togglable. (5) API infrastructure — /api/notifications (GET with unread filter, PATCH to mark all read), /api/notifications/[id] (PATCH read, DELETE), /api/cron/weekly-digest with CRON_SECRET auth.
Major new feature: team collaboration. (1) Team management page (/dashboard/team) — add team members by name and email, assign roles (Admin, Member, Viewer), change roles on the fly, remove members with automatic obligation unassignment. Initials-based avatars, role permission legend, empty state with onboarding prompt. (2) Obligation assignments — every obligation in the Compliance Tracker can now be assigned to a team member via dropdown. Assigned member name shows on collapsed cards and full details in the expanded view. Unassigning is also supported. (3) API infrastructure — new /api/team endpoints (GET, POST), /api/team/[id] (PATCH, DELETE), updated obligations API to support assigned_to field with team_members join. Activity log tracks team member additions and removals. (4) Database migration for team_members table with org-scoped unique email constraint and assigned_to foreign key on obligations.
New dashboard page: a horizontal Gantt-style timeline showing every compliance milestone as a bar spanning from start date to deadline. Tasks are grouped by phase (Foundation, Transparency, High-Risk) with color-coded bars by status (green = done, amber = in progress, grey = upcoming). Features include: scrollable 12-month viewport with pan controls, vertical 'today' marker, transparency and high-risk deadline markers, month grid columns with current month highlighting, hover tooltips with date ranges, inline labels on wide bars, and a bottom legend. High-risk phase milestones only appear if you have high-risk systems. Added Timeline link to sidebar navigation.
Four improvements: (1) New purple shield favicon — the browser tab now shows the ActReady brand icon (purple rounded square with white shield and checkmark) instead of the default Next.js icon. Also added matching Apple touch icon. (2) Dedicated Features page (/features) — comprehensive listing of all 18 platform features with plan badges, descriptions, and feature points. Replaces the anchor link to the landing page section. (3) Dedicated FAQ page (/faq) — 24 questions organised into 4 sections (About the EU AI Act, About ActReady, Compliance Process, Deadlines & Enforcement) with jump links and dynamic deadline countdowns. (4) Cleaned up leftover Next.js boilerplate files from public/ folder (next.svg, vercel.svg, globe.svg, file.svg, window.svg).
Three additions: (1) Compliance Roadmap — a new dashboard page showing your step-by-step path to EU AI Act compliance, with milestones grouped by phase (Do Now, Before Transparency Deadline, Before High-Risk Deadline). Milestones dynamically reflect your actual progress — systems registered, documents generated, obligations completed. High-risk milestones only appear if you have high-risk systems. Includes a progress summary bar and phase-colored timeline. (2) Three new blog posts: 'EU AI Act for Startups' — when obligations kick in, SME provisions, pre-launch checklist, and what investors are asking; 'Article 13 Transparency and Logging' — output interpretability, automatic logging requirements, instructions for use, and a 6-week implementation plan; 'Incident Reporting (Article 62)' — what counts as a serious incident, reporting timelines, provider vs deployer responsibilities, and how to build an incident response process. (3) Added Roadmap link to dashboard sidebar navigation. Total blog posts: 40.
Four changes driven by user feedback: (1) 'Am I affected?' quick-check on the landing page — 3 yes/no questions that honestly tell visitors when they DON'T need to worry, building trust. (2) Sharpened hero copy — changed from broad urgency to specific value ('Know exactly where you stand'), acknowledges not every system needs full compliance. (3) Provider vs deployer role card in classifier results — clearly explains the difference and links to the guide, addressing the most common point of confusion. (4) New blog post: 'What Most Companies DON'T Need to Worry About' — honest breakdown of when obligations are light, when they're heavy, and who actually needs to act. Total blog posts: 37.
Published three new blog posts: (1) Article 10 Data Governance — training data quality criteria, bias examination requirements, data lineage tracking, and documentation obligations for high-risk AI datasets. (2) Article 17 Quality Management Systems — the 12 required QMS components, overlap with ISO 9001 and ISO 42001, documentation structure, and a 12-week implementation timeline. (3) Education AI — which school and university AI is high-risk (admissions, grading, proctoring), what's outright prohibited (emotion recognition), and provider vs deployer obligations for edtech companies and institutions. Total blog posts: 36.
Three improvements: (1) Compliance Score Card — a prominent letter-grade score (A–F) on the dashboard showing weighted compliance across three dimensions (inventory, documentation, obligations), with mini progress bars and intelligent next-step recommendations. (2) Compliance Report PDF — one-click download from the dashboard generates a comprehensive PDF summarising your organisation's compliance status: grade, dimension breakdown, deadlines, system inventory table, obligations progress, documentation status, and prioritised next steps. (3) Added 16 new tests for the compliance score calculation (93 total).
Published three new blog posts: (1) Open-Source AI Exemptions — which open-source activities are exempt, the lighter GPAI regime for open-source models, and why integrating open-source into high-risk systems still requires full compliance. (2) Biometric AI — the complete breakdown of prohibited biometric practices, high-risk biometric systems requiring notified body assessment, and emotion recognition rules. (3) AI Governance Framework — how to set up internal governance with roles, processes, and a 4-week implementation plan aligned to Article 17 quality management requirements. Also enhanced the system detail page with a visual compliance ring — SVG donut chart showing compliance percentage with animated progress, color-coded by completion status, plus risk level icons, approved document counts, and deadline info cards. Total blog posts: 33.
Two new product features: (1) Compliance trend chart on the dashboard — SVG area sparkline showing compliance percentage over time, derived from obligation completion dates. Renders below the progress bar with grid lines and an animated current-state dot. (2) Obligation notes & evidence — each obligation in the tracker now has a notes textarea for recording compliance evidence, decisions, and links. Auto-saves after 1.5s of inactivity with save/auto-saved indicator. Shows 'has notes' badge on collapsed cards. Also includes the performance audit (theme FOUC fix, Hero server component, lazy email popup, DNS prefetch).
Published three new in-depth blog posts: (1) EU AI Act for Healthcare — which medical and clinical AI systems are high-risk, MDR overlap, and clinical decision support classification. (2) Article 14 Human Oversight — the three oversight levels (HITL, HOTL, HIC), specific product features to build (confidence scores, override mechanisms, review queues, emergency stop), and a 12-week implementation roadmap. (3) EU Database Registration — who must register under Article 49, what Annex VIII information is required, step-by-step process, and ongoing obligations. Total blog posts: 30.
Completed a full performance audit of the landing page. (1) Fixed theme flash-of-unstyled-content (FOUC) — added inline script in <head> to set the light class before React hydrates, eliminating the dark→light flash for light-mode users. (2) Converted Hero section from client component to server component by extracting the countdown badge into a tiny client island — reduces client JS bundle on the landing page. (3) Lazy-loaded the email popup via next/dynamic with ssr:false through a client wrapper — popup JS only downloads when the component mounts, not on initial page load. (4) Added DNS prefetch hints for Clerk CDN domains. (5) Made font display:swap explicit on both Geist fonts. Also verified: heavy libraries (docx, jszip) are server-only, jsPDF is already dynamically imported, 8/12 landing components are already server components, no raster images to optimize.
Seven improvements from competitive research: (1) AI System Register — dense table view at /dashboard/inventory with sortable columns, filters, and CSV export. (2) FRIA template already available in document generation. (3) Evidence/proof layer — Audit Trail at /dashboard/audit-trail with timestamped compliance events, date grouping, type badges, search, and CSV export. (4) ISO 42001 Mapping — 12 control-to-article mappings with gap analysis, cross-linked with GDPR/ISO 27001 page. (5) Enhanced regulatory alerts with 3 new updates and per-system deadline readiness monitoring cards. (6) Trust Center prominence — new landing page section with live mock preview of the trust center. (7) Footer and sitemap updates. 21 new tests (77 total).
New System Register page (/dashboard/inventory) provides a dense, table-based view of all AI systems. Columns: name, risk level, domain, compliance %, obligations progress, document count, and applicable deadline. Features: column sorting, search by name/domain/description, risk and domain filters, CSV export for audit packs. Summary stats bar shows totals, high-risk count, average compliance, and obligation progress. Added to sidebar navigation and dashboard quick actions. 11 new tests (67 total).
Added 4 new industry pages (Healthcare, Education, SaaS, Recruitment) for a total of 7 verticals. Enhanced 404 page with navigation cards. Security: HSTS and Permissions-Policy headers. Mobile: responsive grids across dashboard, trust center, training, comparison. Accessibility: skip-to-content, WCAG AA contrast, aria-labels on all icon buttons, aria-expanded on FAQ/nav, search labels. Error monitoring: global boundaries, client listener, /api/report-error. Tests: Vitest + RTL with 56 tests across 8 files (utils, deadlines, error reporter, Button, Card, Badge, FAQ, Comparison).
Published four new blog posts: 79-day week-by-week transparency compliance plan, EU AI Act vs GDPR comparison, compliance cost breakdown, and AI content labeling guide. Fixed double .png.png OG image extension on blog pages. Made hardcoded '83 days' references dynamic so they auto-update on every deploy. Added all new slugs to sitemap.
Verified classify and pricing pages have proper SEO metadata via layout files. Audited all page-level metadata across the app — all key pages have custom titles and descriptions.
Replaced oversized centered warning icons in all error states (systems, documents, tracker, risk map, vendors, global error boundary) with compact inline icons next to the error heading. Same treatment applied to the email signup success state and trust center empty state. All populated dashboard views verified clean with real data.
Cleaned up the risk classifier: removed centered icon boxes from sign-in and upgrade gate cards, tightened gate card copy, replaced the oversized centered result card icon with a compact left-aligned icon+badge header, removed verbose 'Save to Dashboard' description. Applied same result card fix to the shared classification result page. Contact page cards now use inline icon+heading instead of centered icon boxes. Sidebar deadline clock no longer pulses infinitely.
Removed generic subtitles from every dashboard page (they just restated the page title). Removed decorative header icons from pages where they added no value. Replaced all oversized centered-icon empty states with left-aligned, functional copy. Document generation wizard now uses a visual stepper with circles, connecting lines, and checkmarks instead of plain text. Removed redundant Back buttons from detail pages that already have breadcrumbs. Shortened all status guidance banners, editing hints, and AI disclaimers — the product speaks for itself. Obligation tracker cards now show step count and template availability when collapsed. Dashboard onboarding step descriptions are now concise and specific instead of generic marketing copy. Quick actions panel no longer repeats descriptions that match their labels.
Redesigned the landing page to break the repeating 'heading + 3 cards' template pattern. Social proof section now shows product stats alongside logos. 'Who it's for' uses a full-width featured card plus a 2-column grid instead of 3 identical cards. Section headings vary between left-aligned and centered. Tightened section spacing throughout for a more cohesive flow. Hero subhead now separates the CTA line as a bold callout. Feature cards have hover micro-interactions (icon scale, accent border glow). CTA section has enhanced gradient accents. Dashboard onboarding now tracks step completion with progress bar, checkmarks, 'Up next' highlighting, and strikethrough for completed steps. All empty states now include effort hints ('Just 3 fields — takes under a minute'). Sidebar deadline clock subtly pulses to reinforce urgency.
Fully centralized risk level, document status, obligation status, and vendor status configurations — removed duplicate definitions from 10+ files. Clerk auth components (sign in, sign up, user menu) now use CSS variables and adapt automatically to dark and light mode. Vendor checklist and training progress are now stored per-organization in localStorage, preventing data from leaking across accounts. The trust center page derives its risk config from the shared constants.
Aligned Pro and Enterprise plan features between the pricing page and upgrade page so customers see identical information everywhere. Fixed incorrect 'Growth' plan reference on the training page. Removed non-functional sidebar search bar (Ctrl+K hint that did nothing). Added vendor delete confirmation modal. Edit modes now warn before navigating away with unsaved changes. Long system names truncate with hover-to-reveal. Blog listing page now has proper SEO metadata. Pricing billing toggle has correct ARIA switch role. Native select dropdowns now match the theme in both dark and light mode. Replaced inline status label ternaries with centralized constants. Added ARIA labels to sidebar toggle buttons.
Fixed broken OG image for social previews, dead /signin links, and incorrect pricing metadata. Every API mutation (add system, delete vendor, save settings, Stripe checkout, billing portal, trust toggle) now catches errors and shows error toasts instead of silently failing. All data-loading pages (systems, vendors, risk map) now show error states with retry buttons instead of infinite spinners. The useApi hook no longer flashes a full-page spinner when refetching after mutations. Card components are now keyboard-accessible (Enter/Space, focus ring) when interactive. Progress bars across the app have ARIA progressbar roles. Vendor summary cards are now responsive (2-col on mobile). Document detail action buttons wrap on small screens. Removed double padding from vendors, training, and alerts pages. Centralized the timeAgo utility and removed 3 duplicate implementations. Removed duplicate DOMAIN_OPTIONS. Dashboard pages removed from sitemap (conflicted with robots.txt). useCountdown now fires hourly instead of every 60 seconds.
Added breadcrumb navigation to document and system detail pages for clearer wayfinding. Forms now show inline validation errors with red-highlighted fields — system add requires name, domain, and description; vendor add validates email format and URL format. Vendor status changes now trigger toast notifications. Documents list page filters are now synced to the URL, so search queries and status filters persist when navigating back. Created a centralized constants file for risk levels, statuses, and deadlines used across the app.
Added a toast notification system — all save, delete, and status change actions now show instant visual feedback. Every page now handles API errors gracefully with retry buttons instead of infinite spinners. Dashboard shows loading skeletons instead of a plain spinner while data loads. Document viewer adds word count, reading time, and the ability to move documents backwards through the review workflow (back to draft, reopen for review). Textarea is now disabled during save to prevent edits while persisting. Added aria-labels to interactive icon buttons. Fixed og-image metadata double file extension.
AI Systems page now shows a compliance progress bar on every system card and a deadline countdown banner showing days to both the transparency and high-risk deadlines. Compliance Tracker adds deadline context to the progress summary. Documents list page gets section completion progress bars and improved card layout with color-coded icons for approved documents. Generate Document page fixes a bug where doc counts showed total across all systems instead of per-system, and template 'Generated' badges now correctly reflect per-system status. Document cards replace the Export button with a cleaner chevron navigation pattern.
Dashboard now shows both deadlines (transparency + high-risk) when you have high-risk systems, instead of a single generic target date. Documents page gets a search bar and status filter pills (Draft/Review/Approved). Compliance Tracker adds a system filter dropdown to view obligations for a specific AI system. System detail pages now have inline obligation status selectors — change status without navigating to the tracker. Vendor compliance checklists now persist across page refreshes. Risk Map completely overhauled: added a compliance timeline showing both EU AI Act deadlines, clickable system cards with compliance percentages in each risk tier, and an obligation heatmap showing coverage by article with progress bars.
Training page now shows an Article 4 AI literacy legal context banner, estimated time remaining, a 'Continue' button that jumps to your next incomplete lesson, and clickable module progress indicators. Vendor Risk page gets per-vendor compliance checklists (6-item EU AI Act response tracker), a 'Send via Email' button that opens a pre-filled questionnaire with your company name, auto-filled questionnaire templates, an at-risk vendor warning banner, and a 4-column summary row with pending count.
You can now regenerate individual sections of a generated document without regenerating the whole thing. Hover over any top-level section in the Document Outline and click the sparkle icon to improve just that section with AI — content streams in live and the document updates automatically. Also added inline markdown editing (click Edit to modify raw content directly), a placeholder counter that shows how many [INSERT] fields still need your input, and a status guidance banner for draft/review/approved workflows.
Complete upgrade to document generation. Documents now render with proper formatting — headings, bold text, bullet lists, numbered lists, and highlighted [INSERT] placeholders — instead of raw markdown. Added a collapsible document outline for navigating long documents. Section progress tracking shows which section is being generated in real time. Rewrote all 6 document prompts to reference specific EU AI Act articles with paragraph numbers, use the company name throughout, include regulatory context (Digital Omnibus dates), and produce regulation-ready content. Added 2 new document templates: Fundamental Rights Impact Assessment (Article 27) for deployers, and Instructions for Use (Article 13) for providers. Generation now passes company name and country from account settings.
Major visual upgrade to the blog and overlap mapping pages. Blog listing now features category tags with color-coded badges, a featured card for the latest article, category filter pills, and 'Updated' badges for recently revised posts. Blog posts now show key takeaways boxes, collapsible table of contents, callout components (info, warning, tip, important), and related articles at the bottom. Overlap mapping page gets a new coverage summary section with progress bars showing GDPR (~25%), ISO 27001 (~20%), combined (~40%), and net-new AI Act work (~60%) at a glance.
Published two new guides responding to the Digital Omnibus deal. 'Two Deadlines, Two Compliance Tracks — What to Do First' breaks down the priority matrix: what to do immediately, before August 2026 transparency, and before December 2027 high-risk. 'Article 50 Transparency Obligations: Your August 2026 Compliance Checklist' focuses on the deadline that didn't move — covering chatbot disclosure, AI content labeling, deepfakes, and emotion recognition with a concrete 4-week implementation plan. Both added to sitemap.
The EU Digital Omnibus provisional agreement (May 7, 2026) extended Annex III high-risk AI obligations from August 2, 2026 to December 2, 2027. Updated every deadline reference across the entire site to reflect the new dual-deadline reality: transparency obligations still hit August 2, 2026, high-risk obligations now apply December 2, 2027. Updated the countdown hook to support both deadlines, rewrote the Digital Omnibus blog post from speculative to confirmed, updated all three industry landing pages (HR tech, fintech, legal tech), hero messaging, CTA section, pricing trust signal, classifier results, FAQ, structured data schemas, login/signup pages, email templates, dashboard onboarding, alerts page (marked Omnibus as confirmed), training module, vendor questionnaire, and all blog post content referencing the old August 2026 high-risk deadline.
Replaced the text-only landing page with live CSS-rendered product mockups. The hero now shows a classifier result in a browser frame. The features section was completely redesigned from six identical icon cards to three large alternating showcases: the classifier wizard with sector selection, the document generator showing Annex IV sections being built, and the compliance dashboard with obligation tracking. Each showcase has text on one side and an interactive-looking product preview on the other. Remaining features (Regulatory Alerts, Training, Audit Pack) moved to a compact row below.
Three new targeted landing pages at /for/hr-tech, /for/fintech, and /for/legal-tech. Each covers which AI tools in that sector are high-risk under Annex III, the specific obligations that apply, provider vs deployer responsibilities, and links to relevant blog posts. Added all three to the footer and sitemap for SEO. Also enhanced structured data across the site: added BreadcrumbList schema to all blog posts, added LinkedIn/ProductHunt/PeerPush to Organization sameAs, fixed inaccurate free plan descriptions in SoftwareApplication and FAQ schemas, corrected Starter plan feature count, and added 14 missing blog slugs to the sitemap.
Fixed four inaccuracies caught in a full codebase audit: FAQ answer incorrectly stated the free plan included system tracking and document generation — corrected to accurately describe what's free. AI Classifier upgrade prompt said 'Available on Growth and above' — plan is called Starter. How It Works step said '5 plain-English questions' — the classifier has 6 steps. LinkedIn in the 'As featured on' section was a non-clickable div while ProductHunt and PeerPush were proper links — fixed with the correct URL.
Published four new guides: what HR software buyers must do under the EU AI Act (Workday, HireVue, and other off-the-shelf tools), EU AI Act obligations for fintech covering credit scoring and fraud detection, a full breakdown of deployer obligations for companies using third-party AI, and a practical guide to building an Article 9 risk management system. Also fixed the comparison table in light mode — consultant column checks were nearly invisible.
New signups now receive a welcome email with their first four steps — classify, add systems, generate documents, complete training. Paid subscribers get a separate upgrade confirmation email listing everything unlocked and the trial end date. Both emails are plain, on-brand, and include the August 2 deadline. Powered by Resend; gracefully skipped if the API key isn't set.
The dashboard upgrade page now has a working monthly/annual billing toggle. Annual pricing (Starter €23/mo, Pro €63/mo, Enterprise €159/mo — ~20% off) was already wired into the checkout API but had no UI to reach it. Toggle is centred above the plan cards and prices update live. Strikethrough of the monthly price makes the saving immediately obvious.
Added two-tier rate limiting to the AI Classifier and Document Generation APIs — 5 requests per minute (burst protection) and 50 requests per 24 hours per organisation. Returns a clear error message with retry timing if the limit is hit. Prevents runaway Anthropic API costs from accidental loops or misuse.
Login page now matches the signup redesign — two-column layout with a deadline countdown and welcome back messaging. Blog listing now has a live search bar that filters by title and excerpt. Blog posts now include Open Graph and Twitter card images so shares on LinkedIn and Twitter show a proper preview. Added Digital Omnibus amendment to Regulatory Alerts with action guidance explaining why you shouldn't wait for the proposed extension.
Signup page now has a two-column layout with value props, a deadline countdown, and free trial messaging — no more bare Clerk form. Added an email capture popup on the landing page (appears after 40s or 60% scroll, respects dismissal preference). Notifications in Settings now correctly marked as 'Coming soon' instead of showing fake interactive toggles. Risk map shows a proper empty state with classifier and add system CTAs. Training page badge fixed from 'Growth+' to 'Starter+'. AI classifier prompt significantly upgraded with more precise Annex III point references, clearer GPAI obligations, and stricter JSON output instructions.
Published a breakdown of the EU AI Act Digital Omnibus proposal and why waiting for the December 2027 extension is still the wrong move. Also published a detailed guide for legal tech companies covering which AI tools are high-risk, the contract review / e-discovery gray areas, and the specific obligations that apply.
Risk Classifier is now a first-class nav item in the dashboard sidebar — no more hunting for it. The Add System form now shows a 'Classify first' prompt with a direct link instead of opening a new tab and breaking the flow. Pricing page now correctly labels the free plan as 'guided risk classifications' and lists the AI Classifier as a paid feature on Starter and above. Blog posts with deadline day counts now update dynamically on each deploy instead of going stale.
The Manage Billing button in Settings was silently doing nothing. Fixed by building the /api/stripe/portal endpoint and wiring it to Stripe's billing portal session API. Paid subscribers can now cancel, change plan, or update payment details directly from Settings.
Classifier results now generate a shareable link — copy it and send to your legal team, investors, or co-founders without them needing to re-run the classifier. Also added a dedicated Upgrade page inside the dashboard showing all plans, feature comparison, and one-click Stripe checkout.
Published a deadline urgency guide explaining how long compliance actually takes and why procrastinating is dangerous, plus a practical breakdown of what enforcement actually looks like and what happens if you miss the August 2 deadline.
Paid users can now describe their AI system in plain English and get an instant, detailed EU AI Act classification powered by Claude — citing specific articles, obligations, and next steps. The free guided wizard has also been fully rebuilt: EU scope check moved to step 1, new concrete examples under every option, new 'Who does it affect?' question, dedicated GPAI/foundation model detection step, and expanded prohibited practices check including emotion detection in workplaces and schools.
Paid users with Trust Center enabled can now embed a live EU AI Act compliance badge on their website, pricing page, or security docs. The badge updates automatically as you complete obligations and links to your public trust page — giving enterprise customers instant proof of compliance.
Non-signed-in users who complete the free classifier now see a personalised email capture based on their risk level — deadline reminders and updates tailored to whether they're high-risk, limited, or minimal.
Added a side-by-side comparison of ActReady vs hiring a compliance consultant — cost, time, and features. Added a 'Who ActReady is for' section on the homepage covering founders, product teams, and compliance leads. Added trust signal stats to the pricing page.
Added an email newsletter to every blog post. Subscribe to get EU AI Act updates, enforcement news, and compliance guides delivered to your inbox — no spam, unsubscribe any time.
Added an 'As featured on' section to the homepage highlighting ProductHunt, PeerPush, Reddit, and LinkedIn appearances. Updated the CTA section deadline copy to accurately reflect the sub-100 day countdown.
Published the complete list of EU AI Act Article 5 prohibited practices — every banned AI use case explained — and a full guide to conformity assessments: who needs one, whether third-party assessment is required, and what the process involves.
Published a practical guide answering whether the EU AI Act applies to US-based companies (it does), and the clearest breakdown available of the provider vs deployer distinction — including how obligations shift between roles.
Updated meta descriptions across the site to be more specific and keyword-rich. Fixed metadataBase URL to correctly reference getactready.com, improving how the site appears in Google search results and when shared on social media.
The classifier now detects when your product is built on or integrates a general-purpose AI model (Claude, GPT-4, Gemini, etc.) and surfaces the specific obligations that apply to GPAI integrators under Chapter V of the EU AI Act — including transparency requirements, usage policy obligations, and copyright compliance. Document generator updated with a GPAI Integration Notice template.
Classification results now generate a unique shareable link. Send your risk tier result directly to your legal team, investors, or co-founders without requiring them to re-run the classifier. Links are read-only and include your full classification rationale.
Published a practical breakdown of Article 50 transparency requirements for chatbots and conversational AI, and a developer-focused guide to GPAI model obligations for teams building on Claude, GPT-4, and Gemini.
New dashboard page tracking every EU AI Act milestone — enforcement dates, official guidance updates, GPAI Code of Practice drafts, and national authority designations. Filterable by type (deadline, enforcement, guidance, amendment) with action-required guidance for each alert.
Five interactive training modules covering EU AI Act fundamentals, risk classification, high-risk obligations, transparency & disclosure, and GDPR/ISO 27001 overlap. Module 1 is free for all users. Growth plan unlocks all 5 modules and a downloadable Certificate of Completion — shareable with clients, investors, and regulators.
Added a Resources section to the dashboard sidebar with quick links to Blog, Changelog, and Support — accessible without leaving the app. Fixed step number placement on the landing page How It Works section.
The free classifier now generates ready-to-paste disclosure text for limited and minimal risk systems. Copy the exact wording you need for chatbot banners, AI content labels, and terms of service clauses — no legal drafting required.
Analytics now only loads after explicit cookie consent. All AI-generated documents include a disclosure notice per Article 50. ActReady now meets the same transparency standards it helps customers achieve.
Added two new guides — a detailed breakdown of EU AI Act fine tiers with real numbers for SMBs, and a full compliance checklist covering every obligation for high-risk AI providers and deployers.
Generated compliance documents can now be exported as .docx (Word) or PDF directly from the dashboard. Word files use proper heading structure and formatting. PDF opens a print-ready view.
The obligation dropdown in the compliance tracker now correctly maps to all 11 high-risk AI obligations. Previously some items weren't expanding — this is now resolved.
Published five in-depth guides covering the EU AI Act for SMBs, risk classification, Annex IV technical documentation, HR tool compliance, and SaaS company obligations.
Initial launch with free AI risk classifier, AI-powered document generator, compliance tracker, and dashboard. Covers all 11 high-risk AI obligations under the EU AI Act.