Claudes Live Artifacts show what a platform looks like

Anthropic shipped Live Artifacts in Cowork on 20 April 2026. Dashboards and trackers built by Claude now persist across sessions. They refresh with current data every time you open them. This is not a minor feature update. It is the moment Claude's scattered capabilities clicked into a single working surface.

What changed

Until now, artifacts in Claude were static. You asked a question. Claude built something. You closed the window. It was gone. Next session, you started from scratch.

Live Artifacts change that in three ways.

First, they connect to your apps. Gmail, Outlook, Slack, Stripe, Notion, HubSpot; whatever connectors you have wired up. When you reopen an artifact, it pulls fresh data and rebuilds.

Second, they persist. Every artifact saves to a dedicated Live Artifacts tab with full version history. Come back tomorrow or next month, from any session, and pick up where you left off.

Third, they are available on all paid Claude plans. Pro, Max, Team, Enterprise. No gating behind the top tier.

The practical difference is significant. A weekly dashboard that pulls your calendar, email, and task data can now exist as a single artifact you open on Monday morning. It rebuilds itself. You do not re-prompt.

The platform play

Rewind 16 months. In early 2025, Claude was a model and a chat window. A strong model, arguably the strongest for business use. But still a chat window.

Look at what Anthropic has built since then.

Claude Code arrived for developers. Terminal-based, agentic, capable of writing and shipping code across entire repositories. It gave Anthropic a foothold with technical teams.

Cowork launched in January 2026 for knowledge workers. Not developers. Office professionals, operations managers, project leads. People who work in documents, emails, and spreadsheets. Cowork gave Claude the ability to take on multi-step tasks, access local files, and work without constant supervision.

Then came the layers. Over 50 connectors bridging Claude to tools like Google Drive, Slack, Jira, Salesforce, and Microsoft 365. Scheduled tasks that run automatically. Plugins that bundle skills and connectors into role-specific packages. Skills that turn Claude from a writing assistant into something that produces deliverables.

Live Artifacts is where all of that converges.

Connectors give Claude access to data. Schedules let Claude run without you. Artifacts give Claude a place to put its work. Live Artifacts make that work persistent and self-refreshing. Each feature on its own is useful. Together, they form a platform.

This fits a pattern we keep seeing across the industry. There are two groups in this race. Model-first companies like Anthropic and OpenAI started with strong AI but no platform. Platform-first companies like Microsoft and Google had millions of business users but weaker models. Both groups are now converging. Anthropic is building the platform around the model. Microsoft is upgrading the model inside the platform. The gap between them is closing fast.

What this means for UK SMEs

The pace of these announcements is disorienting. A new feature every week. A new integration every fortnight. It can feel like you need to understand all of it before you start.

You do not.

At gecco, we have watched this play out across dozens of SME engagements. The businesses that get the most from AI are not the ones who adopt every feature on launch day. They are the ones who started with their people.

We used to build weekly briefings for clients using ChatGPT Projects and Deep Research. It worked, but it required manual prompting every Monday. Live Artifacts makes that same output persistent and automatic. The tool caught up with the workflow.

That is the pattern. The technology keeps improving. Your job is to make sure your team is ready to use it when it arrives.

Here is where to start.

Get a secure AI workspace and licences for everyone. This should be happening now. Every month without a managed platform is another month of shadow AI, where your staff use free tools with no governance, no data controls, and no visibility.

Find your team's friction points. Where do people waste time on repetitive work? Which reports get rebuilt from scratch every week? Those are your first candidates for AI tools.

Train your team properly. Onboarding, practical sessions, department-specific training, and AI Champion programmes give people the confidence to experiment. Without training, licences sit unused.

Build custom tools that fit your workflows. Schedules, skills, projects, Cowork folders. These are not one-size-fits-all. The value comes from tailoring them to how your team already works.

Start small and measure. AI adoption is not one massive project. It is dozens of small ones. A dashboard here. An automated report there. A scheduling assistant for one department. Each one delivers a measurable result. Each one builds confidence for the next.

Whether you are for or against AI

This is the part that matters most.

The technology conversation is loud. New features, new models, new benchmarks every week. It is easy to feel overwhelmed. It is tempting to wait until things settle down.

They will not settle down. This is the pace now.

But here is the thing. Whether you are enthusiastic about AI or sceptical of it does not change the trajectory. AI is here. It is not going away.

What matters is what you do with it.

Will you use it to save time on the work that does not matter, so you can focus on the work that does? Will you give your team the tools and the training to work with AI confidently and safely?

If not, this moment will pass. And you may find yourself in a role that is being reshaped by AI, with no say in how.

Remember: 80% of AI adoption is the people. The technology is 20%. Live Artifacts is a good example. The feature is impressive. But without trained, engaged staff who know what to build with it, it is another tab nobody opens.

Next step

gecco works with UK SMEs to make AI adoption practical, structured, and people-first. From onboarding and training through to custom Assistants, Automations, and Agents, we help your team get value from AI without the guesswork.

If you want a clear read on where your organisation stands, take the AI readiness assessment. It takes a few minutes and you will receive a tailored report with honest, practical recommendations.