Osman Koç is the CEO of UserGuiding, a platform that helps teams improve user onboarding and build in-app experiences that drive adoption.
We’ve seen many product teams pioneer AI adoption in their organizations, especially with agentic AI tools and coding copilots. AI has enhanced and sped up feature development. However, all these new features and products your team releases one after another require some follow-up.
You need to create onboarding that introduces them to your users, explains why you created them, how they’re useful, how to get started and how to achieve the best outcomes.
Let’s talk about where and when you can utilize AI tools while designing your onboarding flows and communication.
AI Can Handle The Mechanics
The most obvious and quickest way to delegate tasks to AI tools from your onboarding workflow is in creating modals and flows. Just as you use coding agents to develop features and tools, you can use them to develop onboarding flows like guides, checklists, tooltips and banners. One tricky aspect of this approach is that copilots aren’t great at the front end, so UI design and UX still require a lot of polishing and planning from your developers.
Third-party specialized AI tools, like AI-powered onboarding tools, often offer better-trained agents, along with no-code customization and design elements. These tools take the task out of your developers’ hands and put it in your product team’s, who may not be coders or have access to the codebase. Many AI-powered or agentic AI onboarding tools offer onboarding flow optimization, copy/messaging creation, material and flow creation with prompts, and engagement analytics along with suggestions to improve certain aspects of your onboarding.
Getting back to agentic AI tools and copilots, you can also set up workflows that go over the engagement of your announcement modals, banners, guides, checklists or whatever your developers created for onboarding. Copilots are also great at running automated workflows and checkups and reporting their findings via dashboards. Just as third-party specialized AI tools offer AI engagement analytics, you can set up a similar dashboard and reporting workflow yourself. This will let you see what parts of your onboarding work and what parts don’t click with your customers.
You can take this one step further: Instead of just bringing in raw engagement data, you can use conversational chatbots to interpret it and even “talk” to it.
Nevertheless, you need to remember that some of these workflows, like setting up analytics dashboards or custom alerts for certain onboarding engagement trends, require development effort, so much so that at some point, managing all this setup can make sense only if you have a dedicated AI adoption team or AI engineers who solely work on these kinds of tasks.
AI Can’t Decide What ‘Value’ Means For Your Product Or Users
Now, as someone who’s been in the SaaS industry for several years, I can comfortably say that successful, long-term adoption of a technology or tool requires knowing not just where it can be used but also where it can’t (or, at least, shouldn’t) be used. If you’re trying to increase AI adoption across your organization, delegating every task to AI tools isn’t the way to go. This won’t help your team, nor will it make your users happy or customers loyal.
“Good” onboarding heavily depends on context:
• Who are you onboarding? Are they technical or nontechnical users?
• What features/tools has this specific user been using within your app?
• Why are they using your product in the first place?
• Why did you develop this feature/tool, and why are you promoting it to this specific user?
Your answers to these questions will shape where you announce your new feature and how much hand-holding you offer during onboarding. It also affects how you introduce the new feature. Will you highlight your feature’s AI-poweredness? Will you lead with “advanced” capabilities and cool tricks? Or will you position the new feature as a “slight” and “not-so-dramatic” jump from an existing feature because your users consist of non-tech-savvy people who can easily be scared away by new tools?
What I’m trying to say is that you know your users. You know their use cases, their roles and their personas. Your sales team speaks to them; your success and support teams listen to their needs, expectations and problems; and your product team develops new features or updates based on that feedback and insight.
So, yes, AI can easily create an onboarding guide or an announcement pop-up for you, then analyze its engagement and give you stats. But it won’t be able to tweak your messaging or your value proposition for your different user segments and personas—at least, not before you curate it specifically for them first. It can then, of course, grammar-check or give you slightly different options to run A/B tests.
The same goes for onboarding tutorials and guides. You need to decide which parts of a new feature’s workflow, or your product’s UI, to show during a (new) user’s first few interactions, depending on who they are. Your copilot won’t be able to prioritize your features/capabilities/workflows, at least not as well as your product managers.
Key Takeaway
If there’s one thing that stays with you after this article, let it be this: First, you need to decide what’s worth showing and to whom. Only then can AI help speed things up and whip up an onboarding flow. AI tools are good for execution, but it’s you (and your product team) who need to define the onboarding strategy and translate it into value propositions for different user segments.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?