Making AI Feel Like a Human Partner: 5 Behavioral Principles
…with inspiring examples from Lovable, Claude, Riverside, ImagineArt, and more
Hello everyone 👋 I’m Kate Syuma, and welcome to Growthmates.news — the newsletter where we explore growth stories to inspire your professional and personal growth. Join the community of 6,500+ Product, Design, and Growth people from companies like Amplitude, Intercom, Miro, Atlassian, Grammarly, Framer, and more.
AI is spreading fast. But it still feels awkward. Robotic. Duct-taped onto existing products.
We’ve all felt it — that moment when a smart new AI feature pops up in your workflow… and instead of feeling helpful, it interrupts your rhythm. Or worse, confuses you.
At first, AI felt like magic. Now we’re at a different point in the curve — one where users are starting to ask new questions.
Do I trust this?
Is this actually helping me?
Do I want to use it again tomorrow?
Before we dive in… This post is brought to you by Amplitude — the digital analytics platform that helps teams turn insights into action. Their new AI Agents work 24/7 to spot drop-offs, suggest optimizations, run experiments, and improve outcomes — all while keeping you in control 🤓
Want to collaborate on future content or share your product story? Check it out 👉
We used to celebrate AI just for existing. Now, it’s not just about whether you have it.
It’s about how it makes your users feel, and if they would like to come back and truly adopt it.
In the last few months, I’ve been looking around in search for examples where companies did an extra mile — not just integrated AI, but really thought deeper about user experience behind that will naturally foster activation and adoption.
Let’s see what I found 👇
1. Investment Loops 🔁
📌 Let users own the creation. AI is just a partner.
A lot of AI tools fall into the trap of trying to impress users with polished, one-click results. But magic wears off fast when the user isn’t part of the process.
That’s where the idea of investment comes in.
It’s a concept from the Hook Model by
(more in this article) — the idea that when users put effort into a product, they’re more likely to return. Not just because of the output, but because they helped shape it.
The best AI experiences don’t just serve answers. They invite users to co-create — to make decisions, add input, shape direction.
It’s not about making AI do everything for the user. It’s about giving the user more agency, with AI as the assistant.
✅ Best Practice → Riverside “Co-creator“ mode.
Riverside is constantly evolving with AI. Instead of just auto-generating a trailer, it lets you guide the outcome. You pick the vibe. The focus. You say: “This is what I want to create.” That’s what builds emotional buy-in. The user isn’t just handed content — they see themselves in it.
That feeling? You value the outcome, as you helped to co-create this.
2. Human Presence 🫣
📌 Make the first experience personal and memorable.
First impressions matter. Especially with AI.
In traditional products, a clunky onboarding might lead to a 40–60% drop-off. With AI, it’s worse — robotic, awkward first interactions can drive away the majority of users for good. That’s why bringing human warmth into the first moments is critical. Not fake-human.
✅ Best Practice → Calude Onboarding.
Take Claude. Right from the start, it says: “Hi, I’m Claude.” Then it asks what you’re into — invites you to choose a few things. Not in a cold, dropdown menu kind of way. It speaks like a human to a human.
That tone builds trust. And it doesn’t stop after onboarding. Hover over Claude’s avatar any time, and you’ll get dynamic, personable intros: “I’m here to help.” “Need anything?” — like a friendly assistant, not a mysterious black box.
✅ Best Practice → Pine x TheyDo onboarding with AI avatars.
Now let’s talk about scaling that human feeling — especially in B2B or enterprise products, where onboarding is often complex.
One team used a tool called Pine to layer a real human avatar into the onboarding flow — in this case, the CEO of TheyDo (more in this article). The face you see isn’t a stock image. It’s a real person, trained on real context, welcoming you in.
The results? Activation shot up. Because it felt like someone cared.
And the beauty of it — this can scale. With AI-powered avatars, you can create variations, go beyond the CEO, personalize the flow for different users.
Because at the end of the day, people still want to feel seen. Even by software.
3. Social Proof & Familiarity 🙌
📌 Human brains look for social context.
No one wants to be the first to try something unproven, especially in AI.
We seek signals that others like us are already using a tool. That people we trust have tested it. That it fits the identity we aspire to.
This is why social proof is so powerful. It reduces doubt. It fast-tracks decision-making. And in B2B especially — where all sorts of reputations are on the line — it creates a sense of safety.
✅ Best Practice → Show real use, from real humans
Lovable pulls in community builds as user-generated content (UGC). Figma and Notion highlight how teams have used their products for many years. When I worked on Miroverse, I learned in practice how UGC plays a role in building trust and unlocks more relevant use cases for a new audience, who still needs extra help to build a habit. And it works — because people don’t just want to read about features. They want to see how someone like them is already succeeding with it.
Even better if those stories reflect a user’s goals or identity, your user will think: “Oh, if that startup, or that PM, or that design team uses it — then I probably should too.”
✅ Best Practice → Speak like a human, not a spec sheet.
Familiarity also lives in the interface.
Notion AI calls itself “your writing partner.” It speaks in everyday words: “Ask,” “Think,” “Find,” not “query the LLM.” It feels like a collaborator, not a command line.
ChatGPT adds personality, remembers past chats, and mirrors tone — all of which reduce mental load. You don’t have to decode robotic phrasing. It just talks like a person.
And sometimes, a little face helps. Notion uses the same friendly face across its product and website — a small signal that says, “You’re in good company.”
Don’t go too far. If people aren’t sure whether they’re talking to a human or a machine, you’ve crossed a line. Familiar ≠ fake. Aim for trust in the first place.
4. Labour illusion ✍️
📌 We value things more when we see the work behind them — even if it’s automated.
That’s the labour illusion at play: revealing just enough of the process makes outcomes feel more thoughtful, more trustworthy. Invisible magic? Impressive, maybe. But not convincing.
✅ Best Practice → Amplidue AI Agents.
Amplitude gets this right. Their new AI Agents are designed to feel like team members, not tools. They are always working in the background — quietly monitoring user journeys, surfacing friction points, suggesting experiments. Even when you’re not actively prompting them, they’re thinking.
And it doesn’t stop there. When it’s time to take action — say, ship a survey or tweak a conversion flow — the Agent loops you in. You get notified. You approve it. You see what it’s about to do, and why.
The message is clear: this isn’t some black box making decisions. It’s a partner doing real work — and showing its thinking.
When AI reveals its effort, users trust the output more. Not because the system is smarter. But because it feels more human.
✅ Best Practice → Lovable “spinning up preview”.
Lovable leans into labour illusion by showing a “spinning up preview” while generating complex outcomes. This visible pause signals that the system is working, thinking, not just instantly responding.
It makes the product feel more capable and thoughtful, helping users trust the result. After all, when AI does in seconds what would take a human hours, showing some effort helps bridge that emotional gap. I’m curious to see how it will evolve further — if the system builds the whole app in 1 second, will the human brain trust it? I don’t think so.
At the same time, this moment is used to surface other potential features to explore. Just like Slack did in its early days, Lovable turns wait time into a moment of discovery. It’s not only about making the output feel valuable — it’s also about creating a more engaging, guided experience.
5. Progressive Disclosure of Intelligence 🧠
📌 We don’t trust things that feel too smart or unpredictable. We trust what earns its intelligence over time.
The best AI experiences don’t flaunt intelligence too early — they ease into it. Amplitude’s AI Agents don’t try to do everything at once. They start with suggestions and only take action when you’re ready.
Over time, you can unlock more capabilities, like:
Spotting friction points and drop-offs in user journeys;
Suggesting real-time optimizations;
Designing and running experiments;
Acting across the platform (within your chosen guardrails);
Learning from past interactions to improve outcomes.
You stay in control. The Agent simply grows with you — building trust by showing what it can do, not just telling you.
That kind of gradual reveal makes AI feel less like a black box, and more like a partner.
✅ Best Practice → ImagineArt progressive disclosure of results.
Another good example of this is ImagineArt — similar to Midjourney. When you ask it to generate an image, it doesn’t just output a final result. Instead, it progressively shows you multiple options, along with evolving prompts. You can influence the direction step by step, seeing how your choices shape the output.
This kind of progressive disclosure not only builds trust, but also reinforces your role in the process. The AI gets better because of your input — not in spite of it.
How to make AI more human? 🫶
Making AI more human isn't about simulating consciousness or fooling users into thinking they're talking to a person. It's about designing with care, clarity, and collaboration.
Across the five principles, a clear pattern emerges:
→ Let people invest in the process — not just consume the output (Investment Loops);
→ Start with warmth, not wizardry (Human Presence);
→ Show that others have walked this path before (Social Proof & Familiarity);
→ Don’t hide the effort — reveal just enough to build trust (Labour Illusion);
→ And earn your intelligence over time, not all at once (Progressive Disclosure).
Each of these helps transform AI from a cold utility into a meaningful partner. One that feels transparent, empowering, and grounded in human context.
Because the future of AI isn't just smarter. It's more relatable, more collaborative — and ultimately, more human.
Bonus: Resources on building AI products 🎁
How to Make AI Feel More Human with Behavioral Science: webinar recording with
and ;How an AI-Powered “User Onboarding Agent” More Than Doubled our Activation Rate (Pyne ✕ TheyDo guest post);
How to Foster AI Adoption: From Isolated Feature to a Core Experience by
;Why AI startups are blowing past revenue milestones that old-school SaaS could only fantasize about by
;3 Behavioural Methods for better Habit-forming experience by
;A new framework for AI agent pricing by
;Everything you need to know about AI (for PMs and builders) by
.
This is all for today, dear readers 💜 If you found this helpful, share it with a teammate, a founder, or a curious builder. My mission is to help you build and grow meaningful products that delight your users. Here are some ways I can be useful:
User-centric Growth Course (June cohort sold out 🏁)
You can bring your growth challenge and build a growth roadmap from “Quick Wins” to “Big Bets” by combining behavioural science and proven growth strategies. Join the August cohort with 100$ off →
Newsletter collaboration 📩
Let’s tell the world about your product and co-create meaningful content for a vibrant product & design community. Book a slot →
Onboarding & Activation Design Sprint 🏃♀️
I will work closely with your team for 2 weeks to help you fix your Onboarding and improve the first user experience that can drive x2 uplift on your Activation. Reserve your sprint →
Coaching and Mentorship 🤝
DM me on LinkedIn if you’d like to chat about your career challenges, motivations, and transitions you’re exploring.
If you find this newsletter valuable, share it with a friend, and consider subscribing if you haven’t already done so. There are some discounts available, and maybe you can expense this :)
Let’s connect with me on LinkedIn and learn more about my work on Growthmates.club.
With best regards,
Kate Syuma