You don’t ask good questions if you consistently think you’re right about everything. - paraphrase of Jordan Peterson

Scott Meyer - Co-Founder of Chipp.AI

AI is your chance to have a teammate.

AI is great at what you hate.

Three types of knowledge:

• World
• Personal
• Professional

AI Models can be

• Linear: For conversations and PDFs… takes less power/time
• Reasoning: For research… takes more power/time

When interacting with AI, think of a sandwich, you are the bread on the top and bottom.

When Prompting use the RIPE

• ROLE
• INSTRUCTION
• PARAMETERS
• EXAMPLES

Step Prompting is a way to break big tasks into smaller tasks.

• give instructions one at a time, and the AI completes each step before moving to the next step

Branch Prompting is creating a decision tree for the AI interactions.

• The AI can then handle different scenarios based on user input.

The Privacy Pyramid (see image)

Moving into the build phase.

Scott led us through a demo using Chipp.AI

Kevin Peters is from Avid AI - Avid AI sits on top of the data to support marketing and fundraising

AI definitions for this session. There are four times of AI in the context of nonprofits:

• Machine Learning (ML)
• Large Language Models (LLMs)
• AI Agents
• Very Scary Robots

He won’t be talking about the last two types (he’ll only discuss ML and LLMs)

ML

ML takes 3 key forms in today’s nonprofit software

• segmentation
• projections
• personalization

The more donors share the more they give. This includes if they share to multiple organizations. If they share to multiple organizations, they’ll give more to all organizations.

There is a potential problem with the current ML models. ROI is 80 cents on the dollar.

• For nonprofits nationally (globally?) fewer and fewer donors are providing the majority of the revenue.

ML Models should be made if they’ll

• reduce cost
• improve personalization –> key area to future growth

Realistically, you should have 5K+ donors in order to create a good model (rough estimate). Also, are you collecting enough and the right data to make a good model?

What are segments worth modeling?

• those who are likely to respond (direct mail, text, calling)
• those who are likely to upgrade
• those who are likely to give a recurring gifts
• those who are in affinity groups

Technology to consider:

• Fundraise Up
  • Smart Donation Amount Adjustments
  • Overcoming cart abandonment
• Wiland
• Donor Bureau
• Dataro

LLMs

There is a definite problem with scraping the internet for data.

LLMs didn’t know what good looked like.

Remember you market to your actual donor pool.

LLMs are good for:

• Summarizing donor reports
• Write fundraising copy
• Scale personalization
• Let the LLM become the fundraiser

Technology to consider:

• Zapier
• ChatGPT
• Avid (presenter self-promotion)

General Advice and Pitfalls with Marketing

• Do they like text or images more?
• How would they like to get their communication?
• Most major donors give a first gift of <$250.
• Ask an AI client for help overcoming behavioral psychology issues.
• Grok is better for marketing copy

People give to people!!!

• Is the LLM removing “the people”?

A Technology to Know About

• Version2.ai - AI bot that can build relationships with donors. Basically acts like

Next up is Barbara Villasenor telling the story of how the Inter-Belief Network (IBN) formed at Google.

Disclaimer - She is not speaking on behalf of Google.

I’ll let you listen to her story through the recap.

Four lessons she’s learned through her work at Google and building IBN

• Fear has to go. Pray.
• Things will/must happen on God’s timeline. Slow down. Pray.
• Love others with the love of Jesus.
• Find a mentor and be a mentor.

Now Gareth Russell from Jersey Road is discussing that the tide is turning for Christian organizations.

The US is going post-Christian. Europe is already post-post-Christian.

Christians need to:

1. tell positive stories of what Christianity is doing.
2. [missed this one]
3. get really good at saying I’m sorry for Christian abuses of power.

BUT… things are shifting!

He thinks the church is in a much better place than we often describe or discuss.

• [This aligns with Kinnaman’s description of you people and men increasing their belief.)

Here’s why Christian organizations will thrive in the near future:

• Spiritual purpose will soon be thought of as better than doing good.
  • Don’t tone down your Christianness
• Authenticity is a priority. It resonates with the media. The world wants authenticity.
  • Authenticity is a superpower.
• Identify fosters loyalty, especially our Christian identity.
  • We should be humble.
  • We should be less bothered about building our brand and more bothered by building trust.

The next keynote is a panel discussion with:

• Jacob Bullock
• Daniel Wilson
• Damien Daspit
• Caleb Bruce (panel facilitator)

This session is all about bible translation, which I find interesting, but I won’t be taking notes on this one.

Be back during the next keynote.

To start of the afternoon sessions, Richard Zhang opened us in prayer. He’ll be discussing how the technologies we develop should and can point to Jesus.

Founder of Christians in AI

Industry 5.0 is currently happening (started in 2020)

Autoraters are/have replaced humans as the feedback/raters. They rate how good a summary was.

Reasoning has empowered agentic workflows

Reasoning should

Based on the upside-down triangle photo, there are ways we can train our models. He will focus on prompt engineering.

• Rubric design
  • large prompt design because models are becoming more resilient to prompting
• Automatic Prompt Optimization (see photo for the paper that drives his work)

Agentic workflows aren’t taking value away from humans.

• They are showing the ways that we place value on things that aren’t of biblical importance.
  • His example was AI Agents being able to review academic papers just like human reviewers.

Some say

• AI will make us Gods (or like God)
• AI will make us useless

But if we use it well, it will enable us to do the redemptive work of God

James Cuthbert from Biblica

Don from the International Misson Board

Dan from Prediction Guard

Scott Meyer from Chipp.ai

See how easy it is to create AI Agents

We’re going to build an AI Agent.

Why would we use Chipp.AI?

• because it’s better for team development
• it allows us to build CONTROLLED or GOVERNED ChatBots

A sales pitch for Chipp Academy and their 4:30 session.

Chipp Academy

• RIPE framework

From a collaboration standpoint

• you can create much faster than usual

Venture AI: AI-Powered Spiritual Formation for GenZ and Beyond

People are coming to faith online, but one-to-one discipleship isn’t possible.

Venture AI is trying to fill the gap in discipleship and mentorship. Making a powerful and scalable digital discipleship experience.

GenZ in a digital-first world

• Digital Natives
• Lonely
• “I’m unique” with the expectation of personalization
• Authenticity
• Peers > Experts
• Zealous (about whatever it is they’re excited about)

Two types of “new” Christians

1. BRAND new
2. RE-new (used to be Christians or in the church, but left for some reason - like deconstruction or church hurt)

Key Tech Shifts

• big data: can we understand people distinctly and uniquely
• LLMs: allows hyper-personalization to every user
• new GenAI app stack: enables custom, kingdom-minded, redemptive AI
• AI Agents: helps create “teams” for everyone

How does Venture AI serve GenZ well in a digital-first environment?

Venture AI is “Your coach for the Christian journey.”

• don’t automate or outsource, but support and guide
• the person drives their own journey with the support of the AI

micro-discipleship

• personalized (addresses uniqueness)
• bite-sized (meet them where they are and support each step)
• transformative (not just content consumption)
  • see the image of their transformation model

way of Jesus framework

• see image for the 5 components
• How is the new Christian progressing in each category?

Are we offering things and answering questions that people aren’t asking?

Venture AI is looking for:

• partners to pilot the tool
• talent (engineers, designers, developers, etc.)
• funding

Ben is the founder of servant.io

Ben was a marketing professional who has moved into thinking about and developing discipleship tech.

The presentation was pretty marketing focused.

Discipleship is hard.

• reach people
• get them into the funnel
• keep at arms distance while you work through the funnel
• reengage to get people to buy

What is the difference between discipleship and teaching?

• We are currently operating in a box driven by market forces (maximizing profitability).
• We optimize around a high-reach, low-touch systems
• Teaching and Preaching operate in this model
• Discipleship does not operate in this model
  • Discipleship operates in high-presence, low-reach system

The marketing solution to this is AI.

• removes the technical constraint (now I can quickly create specialized content for each individual)
• removes the cost constraint (I no longer need a huge team to do this.)
• reduces adoption resistance (people are becoming more comfortable with these approaches)

Discipleship is very conversational, not like a funnel.

• Think conversationally, not structured (How does this impact recruitment?)
• Multimodal Engagement (website, AI chat, human conversation)
• Transform the organization to have rich, adaptive experiences while not losing the connection
  • more capacity with the same resources and technology

Technology has a “Shadow Side”

• Tech will do what you tell it to do.
• So take the time and thoughtfulness to tell it what to do.

Kingdom-Centered Design Principles

1. Enhanced human connection (human in the loop)
2. Co-laboring Together
3. Augment Don’t Automate the Human Condition (are we replacing the human or are we helping them flourish)
4. Faith is a Market Only You Can Serve
5. Governance Yes, Silos No
6. Focus on Felt Needs

Steps for Moving Foward (see image)