AI advisory and training for organizations without an IT department

AI Productivity without Risk

Using AI safely and effectively

Practical AI • Responsible Results • Real Impact

Large companies have IT teams and security officers to tell staff what is safe to type into a chatbot. Small businesses, congregations and nonprofits have the same tools, the same risks, and nobody to ask. Towpath AI closes that gap: the real benefits of AI for a small team, and the limitations, risks and ethical questions that come with it, understood before they become problems.

See the four modules Bring it to your team

Illustration of a lightbulb, a thinking head, speech bubbles, an open book, and a compass

Who this is for

Organizations where the person writing the newsletter is also the person running payroll, and there is no help desk to call.

Small businesses

Owners and small staffs handling quotes, proposals, customer email, scheduling and marketing without a marketing department behind them.

Churches, synagogues and faith communities

Clergy, administrators and volunteers producing bulletins, newsletters, teaching material and member communications every single week.

Nonprofits and community organizations

Small teams writing grant applications, donor letters, board reports and volunteer materials with more work than hours in the day.

What AI changes for a small team

Where the time actually comes back.

The busy work goes first

Newsletters, meeting minutes, thank-you letters, grant boilerplate, event announcements. The writing that has to happen but never gets easier.
One person can cover more ground

A solid first draft in two minutes instead of an afternoon means the same small staff can take on work there was never room for.
Plain language replaces the manual

You describe what you need the way you would describe it to a colleague. Nothing to install, nothing to configure.
Expertise without a consultant

Complex material — a contract, a regulation, a grant guideline — explained at whatever depth you need, as many times as you need to ask.
The guardrails become part of the job

With no IT department to set the rules, knowing what is safe to share and what to verify becomes a skill each person carries.

“AI cannot replace human qualities like creativity, empathy, and judgment. Instead, AI will amplify our human capabilities and help cultivate our creative spirit.”
Satya Nadella, CEO of Microsoft, 2023

Benefits first. Risks understood. Growth under control.

Towpath AI exists so that small organizations can gain what AI offers while understanding its limits, its risks and the ethical questions it raises. The training is the centerpiece, and it sits inside a larger approach.

Assess

Advisory

Where could AI help your organization, and where should it stay out? We look at the work your people do, the data you hold and the tools you already have, and identify the uses that save real time without exposing members, donors or clients.
Equip

Training

AI Productivity without Risk gets your staff, board and volunteers started: what AI is, how to prompt it well, the guardrails that do most of the protecting, and how to use it as a thinking partner rather than a shortcut.
Mature

Program

A practical plan for growing your use of AI in a controlled way: what to try next, how to review the results, which policies to write down, and when it is worth bringing in more capable tools.

No IT department? Then the guardrails have to be yours.

Four habits that do most of the protecting, taught in plain language and practiced during the session.

Don't paste what you can't take back

Member and donor records, personnel files, financials, contracts, client health information, passwords. Once it is typed into a public chatbot it lives on someone else's system under someone else's terms. Practice data minimization: share the least the task actually requires.

Assume it can be confidently wrong

AI is built to sound right, not to be right. It will invent a statistic, a citation or a quote in the same steady tone it uses for the truth. Before anything reaches a congregation, a board or a funder, someone verifies the claims.

You own what you send

Nobody accepts “the AI wrote it” as an explanation. Whoever signs the letter is responsible for what is in it. Ethan Mollick names this as one of his four rules for working with AI: be the human in the loop, questioning and cross-checking what the AI gives you instead of passing it along untouched.

Free tools are not private by default

Every major provider may use your conversations to operate and improve its service, and each offers settings that limit this. We walk through where those settings live and what each company says it does with your inputs.

What you'll walk away with

Confidence in knowing what AI is, and what it is not
The ability to write effective prompts that save you time on everyday tasks
Clear guardrails so you can use AI safely and responsibly
Advanced techniques to harness the true power of AI in your work and life

The four modules

Each module builds on the last. Sessions are hands-on: you work in a live chatbot on your own real tasks, not on made-up examples. No technical background is assumed or required.

AI Fundamentals

What AI is and what AI is not.

What generative AI actually is, in plain language, and why it is not a search engine.
Why it predicts the next word rather than looking things up, and what that means for trusting it.
The difference between how a model is built and what happens when you send a prompt.
A tour of the major tools, including which are free and what the paid tiers add.

Essential Prompting

The basic prompts that everyone needs to know.

Why a good prompt makes all the difference, shown side by side with a bad one.
The ABCDs of every great prompt.
Ten prompt types worth knowing: explain, summarize, draft, rewrite, ideate, plan, structure, compare, critique and evaluate.
How to bring your own files, documents and web sources into a prompt, so the answer is about your organization rather than the average one.

AI Risks & Limitations

The module that matters most when there is no security officer down the hall.

Recognize hallucinations: why they happen, the red flags, and how to verify before you publish or send.
Prompt safely to reduce guessing and fabrication, and take responsibility for AI-assisted work.
Protect member, donor and client data: what never goes into a public chatbot, and why.
What OpenAI, Anthropic, Google and Microsoft say they do with your inputs, and the settings you control.
Where AI regulation stands today, and what small organizations should be ready for.

Thinking with AI

Beyond drafting email — AI as a thought partner.

Ethan Mollick's four rules for co-intelligence, and what each one looks like in a small organization.
The interview pattern: let the AI ask you questions before it answers, so the output reflects what only you know.
Keeping context alive across sessions instead of starting over every time.
Five ways to use it: recovering institutional memory, stress-testing a decision, coaching your learning, sharpening your writing, and planning your week.

A taste of the method

Two tools from the course you can use tomorrow morning.

The ABCDs of prompting

A chatbot predicts what a good answer looks like. Ask a generic question and you get the most average answer possible. The more direction you give it, the more the answer becomes yours.

A Ask: The clearest, most actionable task. "Rewrite this donor thank-you letter so it sounds warm rather than corporate."
B Background: Who you are, who the AI is, and the context. "I'm the administrator of a 300-family congregation writing to first-time givers."
C Constraints: Guardrails, sources to use and avoid, what not to do. "Use only the facts I've given you. Do not guess; if you're not sure, say so."
D Deliverable: Length, structure, tone and format. "Under 150 words, plain language, no jargon, ready to send."

From AI response to trusted answer

Five questions before anything goes out under your organization's name.

Pause. Does this answer matter enough to verify?
Identify the claims. What facts, numbers, names, dates, quotes or recommendations is the AI asking me to believe?
Check the source. Where did this come from, and is that source reliable?
Compare against reality. Does it match trusted documents, official sources, current data, or what I already know to be true?
Decide how to use it. Use as-is, revise, verify further, or discard?

Start with a conversation

Whether you want an assessment of where AI fits in your organization, the course for your staff, board or volunteers, or help building a program around it, the first step is the same. Sessions run in person or online and are sized to your group, from a single staff meeting to the full four-module course.

Tell us about your organization and what you would like people to walk away with.

Prefer email? chipkalfaian@towpathai.com

Credits and sources

Course material draws on the work of others, credited here and on the slides where it appears.

Ethan Mollick, Co-Intelligence: Living and Working with AI (Portfolio/Penguin, 2024). The four rules for co-intelligence taught in Module 4, including “be the human in the loop,” are his. The spectrum of human, collaborative and AI work used in that module is also inspired by his framing.
Li et al., “Large Language Models Understand and Can Be Enhanced by Emotional Stimuli” (2023), arXiv:2307.11760. Source for the material on emotional stimuli in prompting.
Pascal Bornet et al., Agentic Artificial Intelligence (2025). Concepts adapted for the discussion of AI augmenting human work across industries.
Quotations from Satya Nadella, Eric Schmidt, Mo Gawdat, Geoff Woods, Dario Amodei and Arthur C. Clarke are attributed to their speakers wherever they appear in the course.