Key Points
- Practical AI leads. Workflow automation (24%), natural language search (20%) and localization (17%) are buyers’ top AI priorities.
- Hyper-personalization ranked last. It tied with content auto-tagging at the bottom of the list, at just 6% each.
- Ease of use still wins. It’s the #1 driver of platform satisfaction (57%), so AI that adds complexity works against what users value most.
- Personalization still matters, but it depends on the foundations. It needs connected customer data, and integration remains a persistent challenge for many teams.
- There’s a simple test for any AI feature: what will it save my team this week?
Walk any tech expo floor or scroll through vendor websites and you'll hear the same promise: AI will transform your digital experience. Hyper-personalized content for every visitor. Pages that write and tag themselves. Experiences that adapt in real time.
The people who actually buy and run website platforms are asking for something more modest, and more useful.
In a survey of 100 US IT and marketing leaders, conducted by Liferay with research firm INFUSE, we asked which AI features matter most when adding AI to a website management platform. The answers were practical. Buyers want AI that removes friction from the work their teams already do. The more ambitious capabilities that dominate vendor marketing landed at the bottom of the list.
The findings: automation and search first, personalization last
Workflow automation topped the list, chosen by 24% of respondents. Natural language search came second at 20%, and multi-language support for localization and translation third at 17%. At the other end, hyper-personalization and content auto-tagging were each chosen by just 6%, the lowest of any option. The pattern is consistent: the higher a feature ranked, the more directly it saves time on tasks teams already do every day.

Why buyers are pragmatic about AI
The AI results fit a pattern that runs through the whole survey. Website platform buyers are risk-aware and efficiency-driven, and they judge new capabilities by whether they make daily work easier.
Ease of use matters most. When asked what drives user satisfaction with their platform, 57% named ease of use and learning curve, the top answer. Performance and stability followed at 52%. AI that makes the platform harder to use, or adds another tool to learn, works against what users value most.
Fundamentals come first. When choosing a platform, the top considerations were security and vulnerability management (40%), licensing costs (39%), and scalability and performance (32%). AI features have to earn their place alongside those priorities, not replace them.
Budgets demand clear returns. Only about a third of respondents rated their website platform as a critical or high budget priority (32%), while 39% called it moderate. When budgets are measured, AI that saves hours on routine work is easier to justify than AI that promises a transformation years down the road.
What the top three look like in practice
Workflow automation
Most website teams lose time between steps, not within them. A page waits for review, an approval sits in someone's inbox, or a content update needs a developer ticket. AI-assisted workflow automation targets those gaps: routing content to the right reviewer, flagging what's missing before approval, and scheduling or publishing routine updates without manual hand-offs. Across hundreds of pages and campaigns, those small savings add up.
Natural language search
Keyword search fails when people don't know the exact term. Natural language, or semantic, search understands what someone means rather than just what they typed. For site visitors, that means finding the right answer on the first try. For internal teams, it means locating existing content and assets instead of recreating them.
Localization and translation
For organizations operating across regions, every piece of content can multiply into several versions. AI-assisted translation speeds up first drafts so teams can spend their time reviewing and adapting content for each market, rather than starting from a blank page.
How Liferay approaches practical AI
Liferay builds AI into the everyday content workflow rather than bolting it on as a separate tool.
For workflow automation, built-in or custom publishing workflows automatically route content to the right reviewers, such as legal or design, before it goes live. Teams can set publication, review and expiration dates to manage the content lifecycle, and assets are tagged and categorized automatically on upload. Integrated AI helps generate text, images and meta descriptions, and marketers manage SEO tasks like sitemaps and redirects without filing developer tickets.
For natural language search, Liferay's semantic search uses large language models via Hugging Face, OpenAI and txtai to match the meaning behind a query, not just its keywords. Synonyms, spellcheck and federated search across multiple sources round it out. On the public side, Liferay structures content with metadata that AI answer engines such as ChatGPT, Perplexity and Google's AI Overviews can parse and cite.
For localization, Liferay treats translation as part of the content lifecycle rather than a separate project, and supports three routes so each market can pick what fits the content. Teams can translate in the platform, with source and target side by side and untranslated fields clearly marked. They can automate translation in bulk with AI, using services such as Amazon Translate, Microsoft Translator, DeepL or Google Cloud Translation, and add AI Hub for translations that follow brand tone and guardrails. Or they can hand content off to an agency: export it as XLIFF, the translation industry's standard format, and import the completed translations back in, without the usual spreadsheet-and-email shuffle. Liferay supports more than 50 languages out of the box, including regional differences in date formats and spelling, and generates hreflang tags automatically for multilingual SEO.
For tailored AI, pre-built features cover common needs, but no platform can anticipate every organization's own processes. Liferay AI Hub lets teams tailor AI to their content operations, using a visual, no-code engine to define custom agents and automations, and choosing the best large language model for each task through a standards-based layer built on the Model Context Protocol (MCP). For example, an agent can analyze a new draft and suggest optimized titles, meta descriptions and alt text for human approval, or scan submitted content for prohibited claims and flag specific sentences for the compliance team.
Does this mean personalization doesn't matter?
No. Personalization itself is still a priority. 35% of respondents already use their website platform for content personalization, making it the second most common use after content management. And 25% named it as an area they want to improve.
What ranked last was AI-driven hyper-personalization as a top AI priority. The survey doesn't ask why, but other findings offer a likely explanation. Effective personalization depends on connected customer data, and integration is a persistent challenge. Among organizations that have migrated platforms or plan to, integration with existing systems ranked as the second biggest challenge, behind only content and data migration.
In other words, many teams may need to get the foundations right, such as connected data, efficient workflows and findable content, before advanced AI personalization can deliver on its promise. That's consistent with the order buyers chose.
Questions to ask any vendor about AI
The simplest test is this: what will this AI feature save my team this week? If the answer is vague, or depends on a long implementation first, treat it with caution. A few more questions help separate practical AI from demo-day features:
- Is it built into the platform, or does it require third-party tools? Every add-on adds cost, integration work and another system to maintain.
- Which of our current workflows does it improve? Look for answers tied to tasks your team does today, not future scenarios.
- How does it handle our data? With 62% of respondents naming data protection and privacy compliance as an essential security capability, they know where data goes and how it's governed before turning any AI feature on.
- Can non-technical users use it? If only developers can configure it, it won't relieve the bottlenecks that matter most.
- What does it need from our data and integrations to work? Advanced features often depend on connected systems. Know the prerequisites before you commit.
Is your platform creating friction or removing it?
The buyers in our survey aren't asking AI to reinvent their websites. They're asking it to remove the friction that slows their teams down. The first step is knowing where that friction is.
Get your website management score in 2 minutes. Six questions cover content velocity, integrations, localization, governance, AI and automation readiness, and platform flexibility, with an instant score and a breakdown of where to focus first.
About the research
Liferay, in partnership with INFUSE, surveyed 100 IT and marketing leaders in the United States between January and March 2026. Respondents represent organizations across life sciences, technology, retail and consumer goods, public sector, nonprofit, financial and professional services, and manufacturing, with annual revenue ranging from under $100 million to more than $10 billion. AI feature percentages are based on all 100 respondents; 26 were not asked that question.