Last year I tried spearfishing for the first time.
The instructor took us out into the middle of the ocean and taught us the basics: how to breathe efficiently, slow our pulse before diving, and manage the pressure changes as we descended. Before we could do anything interesting, we spent time practicing a simple maneuver. We’d dive straight down to the ocean floor about 30 feet below, pause briefly, then come straight back up.
We did it over and over.
Straight down. Small pause. Straight back up.
Only after we had that basic skill down did we move on to the more interesting stuff: lying motionless beside a rock shelf waiting for fish to swim by, exploring underwater grottos, and learning techniques that actually make spearfishing exciting.
The conversational chatbot is the basic dive
Building AI features into your software is a lot like this.
Today, the conversational chatbot has become the default starting point. Users increasingly expect to be able to ask questions in plain English instead of hunting through menus, settings, reports, and dashboards to find what they need.
Most companies end up building a chatbot first. That’s fine. It’s a useful capability. But it’s also only one AI pattern.
AI can do far more than answer questions in a chat window. In many cases, the highest-value AI experiences don’t involve a chatbot at all.
Examples in the wild
The fastest way to see what’s possible is to look at what’s already working in real products. So in that spirit, here are nine user-facing AI features already proving their value in the wild.
1 Next Best Action
A recommendation that looks at the user’s situation and suggests the single most important thing to do next.
Example: Salesforce’s Einstein Next Best Action, which recommends actions to service agents based on customer context.
Why: Users often know they need to do something, but they don’t know where to focus. This removes the frustration of sorting through a pile of information just to figure out what matters most.
2 Help Inside the Work Rather Than Beside It
AI assistance that appears exactly where the user is already working instead of requiring a separate conversation.
Example: Grammarly, which suggests clearer wording while you’re writing.
Why: Context switching is mentally expensive. Instead of pulling users out of their flow to ask for help, the help comes to them.
3 Instant Summaries
One click turns a long thread, meeting, report, or document into the handful of points that matter most.
Example: Slack AI summarizing long channels and discussions.
Why: Catching up takes time, and users constantly worry they’ve missed something important. Summaries reduce both the effort and the anxiety.
4 Start From a Draft Instead of a Blank Page
AI generates starting points that users can react to, edit, and improve.
Example: Canva’s Magic Design, which turns a short prompt into design options.
Why: A blank page creates friction. Many projects stall before they start because users don’t know where to begin. Starting from a draft makes momentum easier.
5 Documentation That Writes Itself
The user’s actions automatically become documentation.
Example: Scribe, which records clicks and generates step-by-step instructions.
Why: Documentation usually doesn’t get written because it feels like a separate job. This eliminates the burden by making documentation a side effect of doing the work.
6 One-Click Cleanup
AI handles repetitive cleanup work while leaving the important decisions to the user.
Example: Descript removing filler words like “um” and “uh” from audio and video.
Why: Users rarely enjoy repetitive editing work. This removes a tedious task without taking away creative control.
7 Anomaly Detection
AI identifies unusual activity and brings attention to it automatically.
Example: Stripe Radar flagging suspicious transactions or monitoring tools highlighting abnormal system behavior.
Why: Users can’t investigate problems they don’t know exist. Instead of constantly watching for issues, they can trust the software to alert them when something needs attention.
8 Automatic Organization
AI sorts, categorizes, tags, and organizes information without requiring manual effort.
Example: Gmail categorizing email, photo libraries grouping similar images, or note-taking tools organizing content automatically.
Why: Information becomes overwhelming surprisingly quickly. Automatic organization removes the ongoing maintenance work required to keep things usable.
9 Predictive Insight
AI analyzes historical patterns and helps users see what’s likely to happen next.
Example: HubSpot’s predictive lead scoring, which ranks leads by likelihood to close.
Why: Users are often drowning in data but starving for direction. Predictive insight reduces the uncertainty of deciding where to focus their time and attention.
So what about your users?
Are your users:
- Hunting through data trying to figure out what to do next?
- Getting pulled out of their flow whenever they need guidance?
- Spending too much time catching up on information?
- Staring at a blank page trying to get started?
- Avoiding documentation because it feels like extra work?
- Getting stuck with tedious cleanup tasks?
- Missing problems until those problems become urgent?
- Drowning in information that isn’t organized?
- Struggling to decide where to focus because the future is unclear?
If any of those frustrations sound familiar, don’t immediately ask:
“How can we add a chatbot?”
Instead ask:
“Where are our users getting stuck?”
Clarity on friction leads to clarity on solution
Once you identify the friction, the right AI capability often becomes obvious. Sometimes the answer will be a chatbot. But just as often, the better solution is a recommendation, a summary, a draft, an alert, an organization system, or a prediction delivered exactly when the user needs it.
The companies getting the most value from AI aren’t necessarily the ones building the most sophisticated models. They’re the ones removing the most friction.
The chatbot may be where the dive begins. But it’s far from everything that’s waiting beneath the surface.