Umbraco 17 pairs .NET 10 Long-Term Support with genuinely built-in AI and Generative Engine Optimisation (GEO) readiness. Real enterprises are already running on it - Carlsberg Group, Thames Water, Baker Tilly and the Council of the European Union among them - which says more about its production-readiness than any feature list. Below, we break down what it means for three roles inside a typical enterprise: CTOs weighing infrastructure risk, Heads of Content managing editorial governance, and Digital Marketing Managers chasing visibility in an AI-mediated search landscape. 

Two audiences, one platform decision 

Technical teams want predictable maintenance and airtight security. Content and marketing teams want speed - and visibility not just on Google, but inside ChatGPT, Claude and Perplexity, where more of their audience's research now happens before a click ever occurs. Those two demands used to pull platform decisions in opposite directions. Umbraco 17, released as a Long-Term Support version on .NET 10, is one of the first CMS releases where they don't have to. 

For the CTO: stability first, AI second 

As Umbraco development partners, we spend most of our upgrade technical conversations on risk, not features. Three changes in Umbraco 17 move the needle on that front. 

• .NET 10 and three years of LTS: guaranteed security patches and platform stability through to Q4 2028. Because Umbraco has aligned its own release cadence to .NET's, 17 and .NET 10 are dual LTS releases - a genuinely rare alignment that removes a whole category of upgrade-timing risk. 
• Backoffice load balancing: previously, only the public-facing site could run across multiple servers - the editorial backoffice was stuck on one machine, which became a real bottleneck for large content teams. Umbraco 17 removes that single point of failure entirely. 
• Data governance: The third piece, and it's the one which clients ask about first. Umbraco.AI connects directly to your own enterprise tenant - Azure OpenAI, Anthropic, or AWS Bedrock - so proprietary content never trains a public model. 

It's also worth knowing that Umbraco now ships a Developer MCP and a library of backoffice-specific Agent Skills, so tools like Claude Code or Cursor understand Umbraco's extension points natively - a smaller thing than load balancing, but one that genuinely speeds up custom development. 

Worth flagging: load balancing only pays for itself once you're running a large, high-concurrency editorial team. If you're a five-person content team, it's not the reason to upgrade - the LTS runway and governance model are. 

For the Head of Content: governance without the friction 

A CMS should make editorial teams faster without loosening the reins on brand and accuracy. Umbraco 17 leans into what the industry broadly calls a "centaur" model of human-AI collaboration -a metaphor borrowed from freestyle chess, where the best results come from humans and machines working together rather than either working alone. In practice, that means every AI suggestion in the backoffice still needs a human to approve it before anything publishes. 

• Prompt Management (the feature is officially named this, not "Property Prompts"): editors can trigger predefined prompts directly against specific document types and fields - generating a meta description, translating a summary, or tightening alt text in one click, scoped to exactly where you allow it to run. 
• Umbraco Copilot: a chat-style assistant inside the backoffice that understands what page or field you're working on, can restructure content or clean it up on request, and - as of the mid-2026 release - can save and publish on your behalf, but only with a manual approval step before anything goes live. 

What's newer, and worth knowing about, is the guardrail layer sitting underneath both of these: prompts and AI output are now inspected before or after generation to keep content on-brand and sensitive data out of the wrong place, and a testing framework lets you grade AI output against your own standards before it ever reaches an editor's screen. That's the difference between "AI-assisted" as a marketing line and AI-assisted as something a compliance team will sign off on. 

This is a core part of our own Artificial Intelligence services work - helping teams set the guardrails before the tools go live, not after something goes wrong. And both items are free as part of Umbraco 17, however there are token costs depending on what AI model provider you go with – we can help you calculate the rough costs. 

For the Digital Marketing Manager: winning in the era of GEO and AEO 

Search behaviour is shifting from blue links to AI-synthesised answers, and that shift shows up in the traffic data we're reviewing with clients most weeks now - non-branded queries increasingly resolve inside a chat interface before a visitor ever lands on a website. Building websites and digital products today means designing for AI agents as a second audience, not an afterthought. 

• AI-readable content architecture: structured Block Grid layouts, paired with the free, open-source AI Visibility for Umbraco package, generate clean llms.txt files and structured endpoints so crawlers from ChatGPT, Claude, Perplexity, Gemini and Bing Copilot can actually parse, cite and attribute your content correctly - and the package shows editors who's crawling the site, which is more visibility than most teams have today. 
• Add-on semantic vector search: Umbraco.AI.Search - which is a free add-on from Umbraco - moves past keyword matching entirely. Content is chunked and stored as vector embeddings, so a visitor searching "flexible working options" surfaces the right policy document even if it never uses those exact words. 

One caveat: none of this replaces the groundwork. Vector search only surfaces good content - it doesn't fix thin or outdated content. We typically run a content audit before switching this on, which is where most of the actual visibility gains come from. 

This all reinforces the case for integrated content visibility and analytics, so you can see where your brand is being cited - not just where it's ranking. 

Aligning your platform strategy 

Umbraco 18 has since arrived as a shorter-support release, but Umbraco HQ has been clear that 17 remains the recommended landing point for enterprises - especially anyone migrating off Umbraco 13, where the LTS-to-LTS upgrade path is direct. If you're weighing platform options against the cost of enterprise incumbents, Umbraco 17 is a legitimate answer to both the engineering and the marketing brief at once, which is not something we could confidently say about most CMS releases a year ago. 

Platform choice, AI governance and upgrade sequencing are all pieces of a broader digital transformation strategy - and they're easier to get right when they're planned together rather than bolted on in sequence. 

How we work: with you, not for you 

An upgrade like this is easy to treat as a technical checklist - patch, migrate, ship. We'd argue that's the wrong frame. Every one of the changes above is also a decision point: how much editorial autonomy to hand to AI, which content is worth making machine-readable first, whether load balancing is solving a real bottleneck or a hypothetical one. Those aren't questions a vendor can answer by reading your ticket; they need someone in the room who understands both your platform and your business. 

That's the basis of how we work as Australia's leading independent customer-led technology business - we partner with clients across the full lifecycle, from strategy through build, launch and optimisation, rather than handing off a scoped deliverable and moving on. In practice, that means an Umbraco 17 upgrade with us isn't just a migration: it's a chance to look at what the upgrade unlocks for you specifically - where AI governance should sit in your content workflow, which pages are worth prioritising for GEO visibility, whether your current information architecture will actually benefit from semantic search. We've taken that same approach with organisations like the NSW Department of Finance, Services and Innovation, Queensland Public Trustee, and the Royal Flying Doctor Service - treating platform change as an opportunity to find, not just a project to complete. 

If you'd like a second opinion on your own upgrade path - one partner, one process, no handovers - that's the conversation we're set up to have.

Frequently Asked Questions 

1. What's the difference between Umbraco 17 and Umbraco 18? Umbraco 17 is a Long-Term Support (LTS) release built on .NET 10, with guaranteed security updates through Q4 2028. Umbraco 18 has since shipped as a shorter Standard-Term Support release. For most enterprises — especially those migrating off Umbraco 13 or earlier — 17 remains the recommended landing point, since the LTS-to-LTS upgrade path is direct. 
2. Does Umbraco's AI use our content to train public models? No. Umbraco.AI connects to your own enterprise AI tenant — Azure OpenAI, Anthropic, or AWS Bedrock — so your content is never used to train a public model. Every AI suggestion in the backoffice also requires human approval before it publishes, keeping editorial control with your team rather than the AI. 
3. What is GEO, and why does it matter for our website? GEO (Generative Engine Optimisation) is the practice of structuring content so AI tools like ChatGPT, Claude, and Perplexity can find, parse, and cite it correctly — not just so it ranks in Google. As more research happens inside AI chat interfaces before a visitor ever clicks through to a website, GEO is becoming as important as traditional SEO for brand visibility. 
4. Do we need to redo our content before switching on AI-driven search? Often, yes. Native semantic vector search surfaces whatever content already exists — it doesn't fix content that's thin, outdated, or poorly structured. A content audit before enabling this feature is usually where most of the real visibility gains come from, so it's worth doing first rather than after. 
5. How long does an Umbraco 17 upgrade typically take? 
   It depends on your current version and how much AI/GEO capability you want to enable alongside the migration itself. The best starting point is an Umbraco 17 & AI Readiness Audit, which assesses your upgrade path, AI governance requirements, and content/schema readiness before committing to a timeline.