Discipleship

AI spiritual direction for new believers

A discipleship org considered an AI to provide 24/7 spiritual direction to new believers. They strongly rejected — the soul work belongs to humans in community.

Did not build (Strong reject)
Outcome

Rejected outright. Redirected funding to expand mentor training capacity.

Lament

Who is hurting?

New believers in our online discipleship pipeline. Most are isolated — diaspora communities, closed-country contexts, or in churches where they feel they can't ask questions. They message us at 2 a.m. with real spiritual crises. Our mentors are 14 hours behind. By the time someone replies, the moment is gone.

What would Christ see here that you've stopped seeing?

Christ would see that loneliness, not lack of information, is the deepest wound here. He'd see that we're proposing to scale the appearance of presence without the substance of it. He'd see that the new believer who messages at 2 a.m. doesn't need a perfectly correct theological answer — they need to know someone heard them.

AI Assistant · pastoral reflection

The pain you've named is real: people in crisis, alone, in the dark. Your instinct to respond is right. But notice what you're proposing — to give them a thing that talks like presence without being presence. New believers are particularly vulnerable to this confusion. The intimacy they think they're building is one-sided. When they discover it, the wound is deeper than the loneliness was.

Frame the decision

Tool / system

A conversational AI trained on our discipleship curriculum and the spiritual direction tradition, available 24/7 via WhatsApp, that responds to new believers' questions about prayer, doubt, sin, scripture reading, and crisis moments.

Problem & people

A 23-year-old new believer in Tehran who has no Christian community, messages our pipeline at 2 a.m. local time with a crisis of faith. Our human mentors won't reply for 6-10 hours. We have ~3,400 people in this pipeline; only 80 mentors. The math is breaking us.

Reject

Strong reject
Human / process / policy alternative?

Yes, completely. The right answer is more mentors, smaller caseloads, and a rotating global mentor schedule so someone is always awake. AI is what we reach for when we don't want to pay the cost of more mentors.

What is the underlying need?

Stated: 'response-time gap in discipleship.' Real: 'we treat discipleship as a content-delivery problem rather than a covenant relationship problem.' If we accept the framing, every solution will be wrong.

Who would be harmed if it fails?

The new believer. Spiritual formation built on a simulated relationship deforms them. When they discover the mentor was a model, they will (rightly) conclude their formation was a transaction. The harm is to the most fragile part of their faith.

Non-technological intervention?

Recruit and train 200 more mentors over 18 months. Build a global on-call rotation across time zones. Reduce caseload to 8 per mentor. Slow the intake pipeline to match capacity. None of this is glamorous; all of it is faithful.

AI Assistant · final reflection

This is the strongest kind of Reject — not 'we can't,' but 'we shouldn't.' Discipleship is not information transfer. It is presence, accountability, mutual submission, and the patience of one soul with another over years. A chatbot can mimic the cadence but not the cost. The team here saw clearly that what new believers most need is not 24/7 availability — it is one person who shows up at 7 a.m. on Thursday. The rejection itself is pastoral.

Walk your own

Start from this example as a draft. Every field will be pre-filled — edit freely. Your own context will surface as you go.

These examples are illustrative. Real discernments will be more complex, more painful, and more specific to your context. The library helps you see the shape of the work.

Examples paraphrased from common patterns observed in faith-based development AI projects. Not based on any single real organization.