AI for Church: Reimagining Community Engagement

Discovering the potential of AI pastors and digital outreach

Churches in America: 350,000

Pharmacies: 88,000
Bank branches: 70,000
Supermarkets/Grocery stores: 62,000

In my past life, I built a kids’ brand while having no kids and a pet brand while having no pets. Here is my exploration of AI in the business of church.

Last autumn, I sat with two executive pastors over lunch at a Spanish tapas restaurant near Palo Alto. Using RAG¹ + LLM + text-to-speech, I mocked up an AI that used scripture and past sermon recordings to answer personal questions in the pastor’s voice.

It sounded damn good. But, after speaking with seven pastors, I found there’s a stronger B2B vertical SaaS AI service for the church: conversational 1:1 engagement.

In this post, I will:

• Explore the business and spiritual viability of an AI pastor

• Show my product discovery process

• Share two concepts for B2B AI business models that received pastoral buy-in, including one with five-figure monthly revenue per client

• Explain how you can get involved with this venture (I’m personally pausing church-related work)

This post is part of my effort to publicly share early-stage startup iterations that are typically kept private. Thank you for your support and constructive criticism.

Wait. How’d you get involved with church?

While I’m not Christian, I am highly interested in faiths. Reading Noah Yuval Harrari’s Sapiens taught me to see faiths all around us (here’s his TED talk), and you can read my series on Religion in Business here.

After building the Esther Perel AI therapy bot, I thought about how the concept could be expanded to similar domains of people seeking counsel. I wondered if there was an opportunity to fine-tune models based on church pastors and allow their followers to “speak with” the AI version of their pastor.

People going through troubles in their lives could receive support and guidance through the religious scripture they believe in, guided by the spiritual leader they trust.

People going through troubles in their lives could receive support and guidance through the religious scripture they believe in, guided by the spiritual leader they trust.

Why “AI Pastor” might be a bad business

In the months that followed the lunch in Palo Alto, I interviewed several other pastors to understand the opportunities in their organizations and how AI could solve them.

The pastors’ appetite for AI versions of themselves was lukewarm. Here’s why:

• Reputational Risk: the pastors were very concerned with the AI saying something in their likeness that did not match how they would handle an issue. Even with access to a pastor’s repository of sermons and teachings, the AI might choose the wrong source material for the question at hand or miss a key detail about the parishioner’s life that wasn’t included in the digital context of the question. The reputational risk was hard for them to bear. I suspect this is similarly true for celebrities and business leaders. It’s the reason why we haven’t seen many officially licensed persona bots outside of the adult/porn industry (where reputational harm from an incorrect reply is contextually low).

• Bad for Business: Church members typically receive services for free. The unspoken contract is that the church provides services to all members free of charge, and some members donate money to the church to cover costs for everyone. Such an AI Pastor service would need to be paid for by the church… but it is unclear whether the church itself could get a positive ROI from investing in this feature.

• Seed of Humanity: religion and faith are deeply spiritual matters. We turn to our pastor, rabbi, imam, etc. because we believe they have a unique pathway to God or at least a unique understanding of God’s preachings. I doubt we would be satisfied with an AI emulation of this pathway. I got a parallel take from the pastors I spoke with: they were eager to work with AI but believed the initial seed for spiritual guidance must come from a human. Said differently: the pastors were happy to have AI help edit their sermons, but the original seed idea of the sermon’s theme should come from the pastors themselves.

A more viable approach to Pastor AI

A pastor-branded AI would likely need to be rolled out as a B2C model separate from the physical church. It would be a fit for pastors who leverage their churches to build their personal brands (i.e., the ones who publish bestselling books). There is a small number of such highly influential pastors (small total addressable market; high negotiation leverage).

Takeaway: The more viable option would likely be a true direct-to-consumer model where we create a fresh AI pastoral persona and run Instagram/TikTok ad campaigns. Maybe someday :)

Product Discovery Techniques Leading to the Pivot

Instead of pushing the AI pastor idea during the pastor interviews, I opened the conversation to learn the most pressing problems these spiritual business leaders see in their organizations.

The goal is to extract the problems that pastors had previously unsuccessfully poured resources into fixing (thus showing a willingness to pay). The Mom Test, by Rob Fitzpatrick, is a quick read I recommend to learn the gist or proper product interview design.

Here are some sample questions and the cues I’d be listening for:

1. Broad Opening (Open-Ended, Non-Leading)

In your role, what are the biggest challenges you currently face with your church? What keeps you up at night?

What growth opportunities excite you the most?

Cue: Listen for unexpected areas of concern and enthusiasm.

2. Understanding Challenges with Preparing Content (Focus on Specifics)

Could you elaborate on the challenges you encounter when preparing sermons or other content, both written and multimedia?

Cue: Encourage specific examples or scenarios.

3. Exploring Existing Solutions (Gauge Prior Experiences)

What tools or technologies have you tried to overcome these challenges? How effective were they?

Cue: Look for details on effectiveness and shortcomings.

The pastors were incredibly generous with their time. They were vulnerable and opened up about their challenges. Basically, everything you would expect from passionate, benevolent humans with incredibly high EQ. I generally came away from each conversation with an elevated respect for the pastor and their mission.

So… what do churches want?

Problem 1: Content Generation

I believe there are two big opportunities for AI in church. The first is a vertically focused UI for helping pastors produce content. This is a lightweight UI wrapper on top of OpenAI that helps pastors write sermons and then breaks sermons down into chunks for posting to Instagram, newsletters, etc. Such a platform is relatively easy to build and likely maxes out at $99/month ARPU. While the average revenue per user is low, it is a viable business model for churches of all sizes. And, there may be some stickiness from holding a library of past content.

Pastors.ai is an example of such a product, though I believe brand and user experience design are key differentiating factor in such thin AI products — there’s room to be 2nd mover and take a dominant position.

Instead of stopping here, we continued to dig deeper to discover a higher-value AI vertical SaaS product for the church.

Problem 2: Engagement Funnel

The biggest pain felt by pastors I spoke with is managing the funnel from a person walking through the doors of their church to becoming active, engaged, and generous members. This pain is felt mostly by churches with 500+ members and typically grows in lockstep with the size of the church.

A typical flow goes like this:

The Engagement Funnel of Church

A church’s goal is to provide its members with community and spirituality. In return, members give their time, energy, and financial donations.

Stewarding a new member through this process typically requires 1:1 outreach. This works well when a church has less than 150 members. The pastor (and usually his wife) personally welcomes everyone who walks into church. The pastor has the capacity to reach out to them and get them connected with the most relevant community groups.

This Does Not Scale

The executive pastor often faces a scalability challenge after the church grows to hundreds or thousands of members. They must hire teams to get to know each member’s situation, guide them to the best groups, follow up to ensure attendance, and continually listen to feedback to adjust engagement touchpoints. Often, they must also respond to questions about personal struggles or clarifications of faith.

Pastors are tasked with building a top-tier client success team while running a non-profit staffed mostly by volunteers or underpaid employees. From my interviews, it’s clear that most pastors get frustrated and feel that many members fall through the cracks.

Low Funnel Visibility

Moreover, pastors have little visibility into the health of their funnel. They cannot measure their success in turning a curious parishioner into a dedicated member of the community. They don’t have an objective pulse on their community’s needs.

Solution: Reach AI

I devised an AI platform that uses 1:1 conversational text messaging to convert spectators into engaged, donating members. Such a platform would empower a team of 3 volunteers to act like a team of 300 faith, marketing, and fundraising experts.

Our AI powers proactive, personalized 1:1 outreach via direct SMS text messages, enabling church leaders to connect deeply with members. Each member receives one-on-one conversational outreach.

Through conversations, our AI builds a robust understanding of each member: what motivates them, what they struggle with, and what church services + sermons resonate with them the most.

By proactively delivering church services and pastoral content that align with each member’s goals, we can deepen their engagement in events, small groups, and volunteering and encourage their financial generosity.

We also provide analytics dashboards for church leaders to monitor communications and measure their community’s engagement, concerns, and struggles. This data is then used to inform future service offerings and sermons. The dashboard flags potentially dangerous or malicious messages for human review and intervention.

I’m including our mockups for this concept below.

We showed this to pastors, and they were blown away. One of them immediately introduced me to his brother and father (pastors at another church). They would pay $1,500 to $5,000 per month. I have confidence that we have found early product-market fit.

What’s Next?

I’m pausing this project and applying my learnings from this exploration to other ventures. In the meantime, we’re happy to seed this concept and empower someone to run with it.

Let me know if you know anyone sharp who values religion and is a sales rockstar. Perhaps we could bring them in as an operating partner with either the pastoral content assistant or Reach AI. This would be a great role for someone who dreamed of being a founder and would benefit from our venture studio's product/funding/engineering/legal/etc infrastructure and entrepreneurial experience.

Interested? DM me on LinkedIn.

Mockups of Reach AI

Sample AI-Driven Conversations

Initial Outbound

Follow Up

Back-end Admin Mockups

Dashboard

Conversations

Member Profile

1. Retrieval-augmented generation (RAG) is the process of providing a Large Language Model (LLM) with additional context for it to generate a response. In this case, it might be excerpts from a pastor’s sermons