If you’re in touch with the Digital Employee Experience (DEX) community you probably already know about the latest offering from Nexthink— a generative AI tool called Nexthink Assist.
This is a laudable innovation, and it’s sparked some very interesting conversations around the future of artificial intelligence (AI) in the DEX space.
Generative AI is without a doubt a powerful tool, and the possibilities are as endless as they are exciting.
But I want to cut through the hype.
Today I’ll be sharing what excites me most about the intersection of generative AI and DEX. I’ll also be sharing what concerns me the most.
Let’s take a realistic, pragmatic look at what generative AI is capable of, and what we can expect next from innovators in the DEX space.
What Can AI Do for DEX Professionals? Let’s Ask ChatGPT
Who better to ask about the capabilities of an AI than the most popular AI in the world right now?
I asked ChatGPT-4 a few questions about AI and DEX. Here are its responses, and my follow-up thoughts:
How could generative AI be a useful tool for DEX analysts, whose job is to improve DEX within an organization?
ChatGPT says: Generative AI can be a powerful tool in the hands of DEX analysts in multiple ways:
Data Analysis
Generative AI can assist in analyzing vast amounts of data from different digital touchpoints within the organization. By identifying patterns and providing insights, AI can help DEX analysts understand where improvements are needed most.
This is certainly one of the most exciting capabilities, and it’s the angle Nexthink has taken to promote Nexthink Assist.
Natural language processing allows AI to understand simple, human requests like “Get me a list of all devices with cylanceUI.exe that have experienced device crashes in the last 3 hours.”
Nexthink used to run on a query language called NXQL, and a couple of years back they transitioned to the simpler, easier-to-use successor called NQL. If that was a step in the right direction, then this is an Olympic leap.
Data democratization— the process of making data accessible to non-technical, everyday business users— is a critical element in driving the adoption of DEX practices.
If you want to make people care about the data, make it easy for them to care about the data.
Predictive Modeling
ChatGPT says: Generative AI models can predict future trends based on current and historical data. This can help DEX analysts understand how changes they make could impact the digital experience in the future, and identify potential issues before they become problems.
That sounds great, but I don’t expect it to be useful.
Let’s say your AI assistant identifies that every time an update to your VPN client is rolled out, there’s a spike in connectivity issues. So, since you’ve just rolled out a new update, you can expect issues and address them proactively, right?
The problem is that we, as IT professionals, intuitively know this. More often than not the challenge for DEX professionals isn’t identifying these kinds of issues— it’s getting anybody to care enough to do something about it.
Maybe predictive modeling capabilities will help DEX analysts provide proof where intuition isn’t enough.
But I don’t think we’re going to see anything revolutionary like predicting the future.
Personalization
ChatGPT says: Generative AI can be used to create personalized experiences for employees. This includes tailoring digital interfaces to individual preferences, suggesting relevant resources, and delivering personalized learning and development opportunities.
I love this.
Imagine if every IT professional could have their own AI-generated dashboard of metrics they care about at that point in time.
The moment a major incident occurs, you could ask your AI assistant to do something like “create a dashboard of all the metrics we measure related to connectivity in the New York office” and at the snap of your fingers you’d have a personalized, relevant dashboard from which to track the issue.
Automation
ChatGPT says: Generative AI can be used to automate repetitive tasks, thus improving the digital work environment. This automation can make work more efficient and free up employees to focus on more meaningful, complex tasks.
Automation is always a slam dunk. A study by PWC claims that “labor productivity improvements” will be the first way in which AI tools drive financial value.
But even with AI technology, it will take a little bit of work to automate even simple processes. It’s not clear to me what AI tools have to offer on top of existing automation platforms.
Generating Reports
ChatGPT says: AI can automatically generate reports highlighting the key metrics, trends, and insights regarding the digital employee experience. This can save DEX analysts time and allow them to focus on interpreting the data and planning improvements.
I like this use case, but I have reservations.
Anyone who’s used a large language model (LLM) like ChatGPT will know that sometimes it goes completely off the rails.
As your LLM grows and matures, it becomes more reliable. But that takes a lot of time, and I haven’t seen any AI tools sophisticated enough to not require proof-reading and fact-checking at every turn.
In short, you could use AI tools to generate reports, but you’d have to check every detail of every report. It might save you some time, but it certainly isn’t doing the job for you.
UX/UI Design
ChatGPT says: Generative AI can provide suggestions for improving the user interface and user experience based on data gathered from employee interactions with digital tools. This can help DEX analysts build more user-friendly digital environments.
Replace the words “generative AI” with “DEX analysts” and this sentence still makes sense.
Nothing ground-breaking here. But it’s a win if it makes the job easier.
Testing and Feedback
ChatGPT says: Generative AI can be used to simulate different scenarios and collect feedback on potential improvements. This can help DEX analysts make data-informed decisions about which changes to implement.
What I have in mind when I hear this, is that AI could simulate user behavior to identify potential use-patterns and issues before even pilot-users get their hands on new tools.
That sounds very interesting, but I’m skeptical about whether current AI tools could replicate users.
Users are quite… special…
I don’t think this would replace testing with pilot users, I think it just adds an extra step to the testing process.
Chatbots and Virtual Assistants
ChatGPT says: DEX analysts can use AI to create chatbots and virtual assistants, providing employees with immediate, 24/7 support for any issues they encounter.
Running users through an AI chatbot before they reach the helpdesk is a tricky business.
On the one hand, it’s super useful and works brilliantly with most “how-do-I” questions. It could also help with logging out-of-hours non-priority tickets.
On the other hand, for users who have a genuine issue that’s impacting their ability to work and requires helpdesk support… Well, you’re adding another layer they have to get through before their issue gets resolved.
Making your chatbot AI-powered doesn’t address that issue.
Mental Health Support
ChatGPT says: Generative AI can be trained to provide mental health support to employees, recommending breaks or mindfulness activities when signs of stress or burnout are detected. This can improve the overall digital experience by ensuring employees’ well-being is taken into consideration.
Yikes.
No.
DEX can absolutely help you support the mental well-being of your employees. But AI technology is nowhere near ready to be given responsibility for the well-being of human beings.
Need I remind anyone that a few months ago an AI chatbot drove a man to suicide?
Sure, go ahead and use AI to detect signs of stress or burnout in your employees— but any action you take or advice you give should come from a human. AI isn’t ready.
That said, AI chatbots do make pretty good coaches. They’re good sounding boards for work-related concerns and can give you generic coaching advice that isn’t useless. It’s a slippery slope, though.
Garbage In, Garbage Out?
When it comes to generative AI, just as with humans, diet is important.
And in my opinion, most organizations have a poor data diet.
Before an AI can help you improve your digital employee experience, you must be collecting the right data.
Not just errors, crashes, usage, and other such hard metrics. You need to be measuring qualitative experience through XLAs.
And if you want to train an AI model on your company’s data, you need to have a solid backlog of experience data to feed it.
So while the possibilities of AI in the DEX space are exciting, I think we would all do well to take a step back and make sure we’ve got the basics down.
Generative AI can absolutely take your DEX efforts to new heights, but without a solid foundation to launch from, you won’t get very far.
