AI Trends Redefining Employee Experience in 2024 

Each year I conduct research into the employee experience software market and when I was writing the 2023 report, Artificial Intelligence was barely mentioned. Also, where there were AI features, they were rarely labelled as such. Throughout 2023, software vendors developed and quickly introduced AI features, so it’s now unusual for a platform not to include something-AI.  

Today, AI offers an opportunity to support organisations and improve various aspects of their operations. The features I’ve seen so far fall roughly into four categories, described below, and AI overall is one of the biggest overall trends in the industry at present.  

Generative AI for content creation 

Over half of the products we’ve reviewed now have generative AI features for written content, although only a handful included generative image creation. The way they approach these features does vary, however, offering some or all of the following: 

• A natural language prompt for publishers to instruct AI to create a page of text 
• Settings for tone, length, formality – applied as part of generation and / or to guide the AI to edit text* 
• Generated summaries of written text*. 

* Either that the AI has created, or a publisher has written. 

A few of the features I’ve seen particularly stand out as being innovative (in an already innovative space!). The first example is from Oak Engage, which has an inbuilt policy template that is applied when a publisher wishes to create a new policy page. The AI then completes the template, generating a draft policy for review and editing. This is a helpful approach, although the template might feel too restrictive (and overly comprehensive) for some. 

 Microsoft Word introduced an ‘inclusivity’ language checker during 2022 and I’m surprised that only Interact have replicated this in their page builder. The sentiment of the page’s tone is also checked in the same way, allowing publishers to consider how their piece might be perceived before it’s published. This is a small yet supportive feature that I think will help publishers finesse their pages. 

Where publishers want to use AI to generate the majority of their content, there are some tools like Unily that not only ring-fence organisations’ data (a more secure approach than a simple integration with ChatGPT for example) but also apply an ‘internal audience’ filter to ensure generated content is more audience-appropriate first time. Generally, I’ve seen security-conscious features, but feel that greater ring-fencing needs to be introduced to make content more closely tailored to organisational needs. 

Powell and Sociabble are the first two products I’ve seen to introduce configuration options for admins, which provides context for the AI to refer to. Powell applies intranet location context on top of any generated content, while Sociabble has sophisticated settings for organisations and individuals to write context statements that any generated content is passed through.  

Overall, generative AI features in these tools are now fairly common and I expect remaining vendors to add generative AI features in the coming months. I would advise that organisations should decide whether they want to use generative AI features and how, before exploring what’s available. 

Search and information finding 

Very few products have introduced AI into the search experience, which is a surprise given how often ‘search’ appears high on the list of improvements employees wish to see. Granted, issues with search aren’t typically associated with the technology itself, but there is a lot of potential for AI to make real improvements here.  

A handful of products have introduced ‘likely answers’ into search results that are generated by AI, including ahead, Atlas, Haiilo, MangoApps, Microsoft (via Copilot), Engage (from Sorce) and Workgrid. This replicates a Google-like experience, where the search presents the answer to a question rather than merely lists of results of pages where the individual would have to continue their hunt for information.  

The different products vary in the way they provide ‘likely answers’ though, typically only offering some of the following (I think it would be beneficial for all these features to be present): 

• ‘Teaching’ the AI when it has provided an inaccurate answer 
• Integrations with third-party tools, such as Salesforce or Workday 
• Presenting additional resources, such as an associated file, within the answer or providing a link to it 
• Analytics to help admins see what is or isn’t working within search 
• A chatbot-style interface and / or making this part of the standard search experience 
• Contact information for subject experts should someone want more information. 

Using AI to generate content tags is another good example of aiding information finding and search, as this can be something publishers forget to do themselves. Akumina includes a tool to assist with this and I think this is really helpful. Likewise, effective titles can be tricky for some people to create – products like Staffbase have introduced AI tools to assist with this part of the process too. 

In terms of information finding, some products have (e.g. Interact) AI-generated answers. These take the content of a page and summarise the information in a question-and-answer format that publishers can then edit. This approach provides a bite-sized way for people to consume the information and allows for a pleasant browsing experience to find what they’re looking for. There’s a risk that every page could become an FAQ though, so good governance and guidance is needed. 

As a slight variation of this, Copilot in Viva Engage will present ‘readers may ask’ prompts to help publishers cover all relevant information in their initial post. This helps with comprehensiveness but may stifle questions, making the intranet feel less conversational and open. 

There is a lot of potential to dramatically improve employee experiences of products if AI is effectively applied to search. Debate has raged for years as to whether intranets and associated tools can be ‘front doors’ into organisations’ digital workplaces. If AI-driven search is developed, then these products are in a very good position to offer that way in – potentially more so than other digital workplace tools. 

Workgrid is a pop-up assistant that lets people ask natural language questions, then will present ‘likely answers’ to them.

AI support for Admins

This aspect is getting less attention, but for intranet managers there is great potential for automating some routine tasks.

For example, Copilot in SharePoint will soon be able to generate a SharePoint site from a prompt such as “Create an employee onboarding site featuring our company logo and using the Welcome to Contoso.pptx file.” Just like fine-tuning text, additional prompts can be used to adjust the layout, branding and so on. Think of it as being a little like the ‘Designer’ feature in PowerPoint – but how many times has the Designer suggested something inappropriate? The technology will certainly help here, but it will only be as good as the starting content.

I like the concept of AI-driven analysis to give greater depth to analytics, so that true actionable insights are provided to admins and content owners. Sadly, I’ve not seen any vendors doing this just yet. Analytics are typically the weakest part across these platforms, so there is huge potential to develop strong AI-driven approaches.

Finally, there is scope for AI to assist with content governance and site life cycles. Detecting duplicate sites, searching for outdated content, and checking for copyright or brand infringement would benefit greatly from AI attention. For example, LumApps’ forthcoming ‘governance engine’ automatically identifies stale content and suggests areas where there may be content gaps.

Overall, there are few AI approaches for admins on the market just yet; likely because these are less glamorous than eye-catching generative AI features. However, this is where real time saving can be made and employee experiences improved, so I hope to see more features introduced over the next year or so.

General UX enhancements 

The final area is a catch-all for other AI features, which are smaller tools that still contribute towards improving the overall user experience. Auto translation is the most common of these features, yet is still not common, which I think is a shame. However, Unily offers speech to text to automatically generate captions in video and these can also be auto-translated – so there are some vendors taking steps in the right directions here. 

Similarly, MangoApps has an exciting AI audio generation tool that renders an audio version of a page without the need for publishers to record anything. This is a great accessibility feature and is a pseudo podcast feature, that will allow people to keep up to date with news without having to read something. I hope that the next step is audio translations. 

A handful of vendors are using AI to assist with personalisation and audience targeting too, which means employees are more likely to see news that’s relevant to them. Firstup’s Orchestration Engine is a good example, which has been around for a little while and surfaces content to people at a time that’s right for them and appropriate for the message. 

For more information on the AI features I’ve mentioned, please download our latest Intranet and Employee Experience Platforms Report. If you’d like help selecting the right platform for your organisation, then please get in touch. 

If you’re a software vendor and wondering why your AI features haven’t been mentioned, please visit our information pages and get in touch if you’d like us to review your product.

This article was originally published on reworked.co in May 2024.