Key takeaways
- AI is changing how companies handle support, and for many users, human contact is disappearing from the experience entirely
- Automation has real benefits, but it works best when it complements human judgment, not replaces it
- At Shift, every support interaction involves a real person who actually knows the product
- Real support builds trust, surfaces patterns, and helps shape what gets built next
- The Shift Community forum is an extension of that philosophy, a space where users help each other and influence the product
Something has been shifting recently in how companies handle customer support. The live chat button that once connected you to a person now routes to an AI automated robot. The help center that used to feel human is now a decision tree with a friendly avatar. Even calling up your local mechanic might land you on an AI answering and booking system. This change has become widespread, and has occurred mostly without announcement.
The numbers back it up. In 2020, only 5% of customer service teams were using AI-powered automation. By 2025, that number exceeded 80%, one of the fastest adoption curves in enterprise technology (ChatMaxima, 2026). That's a near-complete transformation of what a customer service conversation looks like for most people in only a handful of years.
For those in support, I get it. AI customer service tools have real advantages. They scale without adding headcount, they respond instantly at 3 a.m., and they handle repetitive questions efficiently. If your product has thousands of users asking the same five questions, automation makes sense. The business case is clear.
But something gets lost in that trade-off, and we're starting to feel it.
The gap automation leaves
When support is fully automated, you start losing something valuable: the texture and essence of what users are actually experiencing. Automation can close tickets but it can't tell you that three different users this week described the same frustration using completely different words. They can't pick up on the thing someone mentioned in passing that turns out to be a real product insight. They can’t assess a user’s unique account needs and use judgement and empathy to determine next steps. They can't read between the lines.
There's also the element of trust to consider. 79% of Americans still prefer interacting with humans over AI for customer service (Lorikeet, 2026). That preference doesn’t exist in a void. When a user is stuck, frustrated, or confused, they want to know someone is paying attention. That the company behind the product they may love values them as a user and is tuned in to their experience. That’s been the unspoken contract between businesses and consumers for ages.
A resolution matters, of course, but so does the experience of getting there and if that experience feels cold or impersonal, it affects how users feel about the product overall, regardless of whether the issue was solved.
That's not necessarily a knock on the tech. It’s just the reality of a legacy of human service and support that isn’t going to change overnight.
What do people actually want from AI? The 2026 AI Insights Report is here
AI is everywhere, but trust is fragile. Our 2026 AI insights report breaks down how people are using AI, what concerns them most, and why transparency and human centered AI design will define the next wave of technology trends in 2026.
Read MoreWhy we keep humans in every conversation
At Shift, there's a real person on the other end of every support interaction. That's a deliberate choice, and one we've maintained even as the team has grown. AI for customer care has its place in our workflow, but it's never the thing standing between a user and a human who can actually help.
Here's what that means in practice:
- Faster context. Our support reps know the product deeply. They can quickly understand what a user is trying to do, or test for a workaround, not just match keywords to an FAQ.
- Better signal. When our team has real customer conversations every day, patterns surface. We hear what's confusing, what's working, and what's missing. That information then flows directly to our product team and informs product decisions. We’re always working to improve and speed up the process of getting our customer’s voices into the ears of our product team and decision makers.
- Actual accountability. When something goes wrong (we’re human, so things will) users deserve an authentic response from someone who can own it. That's hard to replicate with a scripted flow.
- Relationship. This sounds soft, but it's real. Users who feel valued and heard stay. They also tell you things they wouldn't bother typing into a no response bug form.
I've been doing this long enough to know that support conversations are one of the most honest data sources a company has because they preserve the subjectivity, the sentiment, the urgency. You can't automate your way out of needing that.
Support as a feedback loop, not just a function
It might be a cliche, but the best products I've seen aren't just built for users. They're built with them, and that requires ongoing contact with the people actually using your product.
That's part of why we built the Shift Community forum. It's a space where users can share workflows, surface ideas, flag bugs, and support each other. It extends what our team does every day in direct support conversations into a broader, more collaborative space where the community itself becomes part of the feedback loop.
Inside the Shift Community, you can:
- Share your setups and workflows with other users
- Offer peer support and learn from how others use Shift
- Surface feature ideas directly to the team
- Report bugs and follow known issues
- Stay up to date on releases and product changes
The goal isn't to replace direct support, but add to it. It was time for users who want to go deeper, connect with others, and contribute to what Shift becomes next, to have somewhere to do that.
What "human support" actually means
I want to be clear that prioritizing human support isn't a statement against technology. We still use tools that help our team work better. Shift AI, which is built directly into the browser, is a good example of how we think about this: intelligence that assists, not replaces. The same logic applies to support.
The question isn't total automation vs. humans only. It's about identifying where human judgment and connection adds irreplaceable value, and being honest about that. Consumer sentiment isn't necessarily anti-AI, it's anti-bad-AI, where reducing costs beats out customer experience. When an optional AI resolves something quickly and well, people are fine with it. Some users are always going to default to self-serve when available, which is why options for support matter. But, when automation becomes a wall between a frustrated user and someone who can actually help, it erodes trust and sours the experience.
In support, the value of ‘human’ shows up in empathy, in context, in fast testing to find a workaround, in a personalized message that acknowledges a user’s hardship, in knowing when a question is actually a sign of a bigger problem, and, above all, in leaving someone knowing that their experience with the product actually matters to the people who built it.
If you're a Shift user and want to connect with others, share your setup, or just see how people are working, the Shift Community is worth checking out. And if you ever need support, there's a real person waiting.
