In my previous three articles, I covered the “AI Checkbox Fallacy,” the “Seduction of the Sparkle,” and the importance of “Building for Tenants, Not Tourists.” I’ve been pretty critical of how companies are implementing AI, but criticism is easy. What’s harder is actually putting principles into practice.
That’s why I wanted to share how I’m trying to walk the talk with my Bible study app, Harvous. My approach boils down to one simple rule: App first. AI second.
How We’re Doing It at Harvous
At Harvous AI stays in the background. It never takes center stage. This drives everything else we do and really captures what I mean by “App first. AI second.”
Here’s what this looks like in real life:
- Contextual Scripture Support: When someone in your group shares a thought, AI quietly analyzes for relevant cross-references and historical context. You can see these through a simple “More Context” button if you want. The AI doesn’t butt in or take over—it just makes extra info available when you ask for it.
- Discussion Help: The AI can suggest questions to deepen the conversation, but only when you specifically ask. These aren’t generic AI questions, but thoughtful prompts designed to help your group explore the text further.
- Personal Learning: The system looks at your notes to create custom review exercises based on your own insights. It’s not teaching you new stuff; it’s helping you remember and connect your own observations.
Notice what I deliberately left out: There’s no “AI Bible Teacher” feature. No “Ask the AI Pastor” button. No automated theological correction system. Sure, these would be flashy and marketable, but they’d completely undermine what Harvous is actually about.
Our Actual Implementation Choices
Here’s how this translates to specific choices we’ve made:
- Everything is opt-in: All AI features are opt-in, never automatic. You have to actively choose to use AI assistance.
- Suggestions, not directions: AI suggestions are framed as considerations, not authoritative instructions. “Have you thought about…” rather than “This means…”
- Enhancing what’s already there: We integrate AI capabilities into existing user flows rather than creating separate “AI experiences.”
- You’re in control: Both groups and individuals can adjust how much AI help they want, putting you in control.
- Finding connections, not creating content: The AI surfaces relevant scripture and connections rather than generating new theological content.
None of this is particularly flashy. None would make for a great investor demo. None would probably generate press coverage. But all of these choices serve the people who actually use Harvous for Bible study—rather than random “AI tourists” who might be briefly impressed by something more attention-grabbing.
This Isn’t Just About Bible Apps
What makes our approach different isn’t specific to Bible apps—it’s a principle that could apply to any product: App first. AI second. The core purpose and what users actually need should drive everything, with AI serving that purpose rather than becoming its own thing.
- For a Bible study app, that means AI should enhance study and discussion, not replace it.
- For a photo editing app, that means AI should make editing more powerful, not take away creative control.
- For a writing tool, that means AI should support your writing process, not take it over.
This seems obvious, but it’s surprisingly rare in practice. Most companies aren’t asking “How can AI make our core product better?” It seems they’re asking “What AI features can we add?” It’s a subtle but critical difference.
Practicing What I Preach
It’s easy to criticize how companies are implementing AI. It’s way harder to actually build something better. With Harvous, I’m really trying to put these principles into practice—building for the people who actually use it, using AI as a subtle enhancement rather than a flashy feature. In every decision, I remind myself: App first. AI second.
I won’t get it perfectly right. There will definitely be missteps and adjustments. But by starting with clear principles and focusing relentlessly on what Harvous is really for, I hope to avoid the traps I’ve been writing about.
The future of AI isn’t about products that show off their AI capabilities. It’s about products that use AI so thoughtfully and subtly that you barely notice it—you just notice that the product works better for you.
Because while tourists might be impressed by the AI label, tenants just want a better place to live. And that comes from putting your app first, and AI second.