How to find smart, low-cost AI models
A short method for choosing API models by capability, total task cost, provider terms, and performance on your organization’s real work.
Every guide, how-to, case study, prompt, tool overview, talk, and FAQ in one place. Filter it down, then share the link.
37 resources
A short method for choosing API models by capability, total task cost, provider terms, and performance on your organization’s real work.
AI tools give you two separate dials: which model answers, and how hard it works before answering. A professor/student analogy shows which one actually helps for the task in front of you.
Major AI vendors and meeting recorders will hand over conversation content when compelled by valid legal process, and both OpenAI and Anthropic report doing exactly that — but most reported cases so far involve a phone or laptop seized from a suspect, not chat data pulled from the vendor.
Use ChatGPT Codex or Claude Cowork to understand footage, find strong moments, and create a social edit. This guide shares what we tested, where Premiere may fit, and what still needs human review.
Before you commit to a campaign or project plan, ask AI to imagine it has already failed — and tell you why.
OpenAI's general AI assistant for drafting, summarizing, analyzing files, and multi-step research. The most widely recognized tool of its kind, and often the one your staff have already tried.
A free AI research assistant that answers questions using Faunalytics' library of 6,000+ human-vetted animal advocacy research summaries, with citations to the underlying sources.
Google's general AI assistant, built into Gmail, Docs, and Drive. Eligible nonprofits get the premium features at no cost through Google Workspace for Nonprofits, which makes it the cheapest serious starting point for most teams.
How to divide an AI budget across a team: a paid assistant seat for everyone plus a meeting recorder, specialist tools for specialist roles, the AI features already inside your existing software, and heavier plans for the few people who will use them.
AI note-takers can turn a meeting into a summary, decisions, and action items. Use them to capture the call, then check the details before sharing the notes as the record.
Vegan Hacktivists worked with a small team to design and build a Google Apps Script monitor for its public resource hub. AI helped assess the problem and write the code; the finished script checks links with deterministic rules and leaves decisions to staff.
Use n8n for flexible AI-heavy multi-step workflows, Make for fast visual connections across familiar SaaS tools, Apps Script for lightweight Google Workspace tasks, and custom code when logic outgrows no-code.
All three run the same Claude models. What changes is how much of your computer each one can reach. Cowork is where we suggest most people start.
Deep research searches many sources and produces a detailed, cited report. Use it for complex questions, then verify the sources and claims.
You need to understand a research paper quickly and report on it without overstating what it found.
A bill is moving and you need to mobilize supporters to contact their legislators this week.
A board meeting is approaching and you need to turn scattered program updates and metrics into a coherent report.
You want to send donors a no-ask update that keeps them connected between appeals.
You have a grant deadline, a mountain of messy program notes, and no draft.
You have messy notes or a transcript from a meeting and need clean minutes with decisions, action items, and owners.
You need a clean, newsroom-ready press release for an announcement, campaign launch, report release, or rescue/event, and you're starting from scattered notes.
A general AI assistant from Anthropic for drafting, summarizing, analyzing files, and thinking through problems. For most staff, this is the right first tool to learn.
A section-by-section worksheet for writing an AI policy your staff will use, with starter questions for each section and a skeleton outline you can copy.
A working framework for organizational AI adoption: six areas — narrative, tools, training, guardrails, workflows, and rhythm — that need attention together for experiments to build lasting capability.
Our AVA Summit talk on fluidity and agency in a changing world: six lanes every organization's AI strategy needs to cover — with the full slide deck to use with your team.
An AI tool that turns prompts, outlines, and existing documents into first-draft presentations, documents, and lightweight websites.
A practical walkthrough of setting up a Claude Team account for a nonprofit: verification and discounts, admin choices, connectors and shared skills, privacy settings, and recommended defaults for advocacy orgs.
Four organizations serving the animal advocacy movement publish full AI policies covering values, tools, and day-to-day rules. One is notably more permissive than the rest. Compare their choices, then adapt the parts that fit your organization.
An AI agent pursues a goal across multiple steps, using tools to read files, search, update systems, and check its work.
Nonprofit plans can reduce the bill sharply. Standard paid seats usually cost tens of dollars per person each month; automation platforms can be inexpensive at small scale; APIs charge separately by usage.
A capacity-building services program uses an n8n workflow to research each new applicant, compile a source-linked brief, and post a summary to Slack — so the team starts every review with consistent context.
A program team answering a high-volume support inbox designed a service that matches incoming questions to its FAQ and drafts replies in its voice — as Gmail drafts for staff to review. Currently at the concept stage.
It depends on the vendor, plan, and settings — business and API tiers usually do not train on your data by default, consumer plans vary, and policies change, so check and document your settings.
A team that runs a recurring conference piloted an n8n workflow that researches speaker applications with AI and web search, then assembles a source-linked workbook and per-submission summaries for organizers to review.
The objections staff often raise about AI, taken seriously rather than dismissed — with what is true in each, the data behind it, and language options for the conversations leaders are already having.
A communications team running recurring events uses Make.com and an AI drafting step to turn event details into ready-to-review Gmail drafts, complete with calendar links. Nothing ever sends automatically.
A simple four-tier model for deciding what data can go into AI tools, what to check before you trust a vendor, and practical defaults a small organization can adopt this week.
Vegan Hacktivists offers free AI and automation support to animal protection organizations — from a first conversation to a full build.