AI for Nonprofits: Chatbots Copilots Agents and Automation

October 1, 2026/ Mark Becker / ThursdAI / 0 comments

Welcome to ThursdAI by Cathexis Partners, a new series offering practical, responsible AI guidance for nonprofits.

Over the coming weeks, we will move from the basics—what the different types of AI are and what they cost—into the seven-part Cathexis AI Readiness & Enablement Process. Each article will include examples and steps you can take immediately. The goal is not to convince every nonprofit to use more AI. It is to help organizations decide where AI can create value, where it introduces risk, and what must be in place before moving forward.

The series supports the Fundraising.AI Framework for Responsible and Beneficial AI for Fundraising (https://fundraising.ai/framework/). Its tenets—privacy and security, data ethics, inclusiveness, accountability, transparency and explainability, continuous learning, collaboration, legal compliance, social impact, and sustainability—will recur throughout the series. The Framework establishes the principles. The Cathexis process helps nonprofits apply them across people, processes, data, and technology.

Artificial intelligence has become an umbrella term for a growing collection of tools. A chatbot that helps draft an appeal, an AI feature inside a CRM, an automated workflow, and an agent that can complete several steps are all described as AI. They do not have the same capabilities, costs, risks, or requirements.

Before a nonprofit evaluates products, it helps to establish a shared vocabulary.

Chat assistants

ChatGPT, Claude, Gemini, and similar tools let a person ask questions, analyze information, create drafts, summarize material, and explore ideas through a conversational interface. They are general-purpose tools. Their usefulness depends on the information supplied, the instructions given, the model being used, and the person reviewing the response.

These tools can help with early drafts, research plans, meeting preparation, policy outlines, spreadsheet analysis, and many other tasks. They can also produce incorrect or incomplete answers. A confident tone is not evidence that an answer is accurate.

Copilots and embedded AI assistants

A copilot brings AI into software where staff already work. Microsoft has made the term especially familiar through Microsoft Copilot, but “copilot” is also used more broadly for an AI assistant that works alongside a person within an application or workflow. It may help draft an email, summarize a document, analyze a spreadsheet, prepare a presentation, retrieve information, or recommend a next action without requiring the user to leave the underlying software.

The important difference is context. A general chat assistant knows what a user provides during a conversation. A properly configured organizational copilot may be able to work with information already stored in the organization’s environment, subject to permissions and product configuration.

That can make a copilot more useful, but it also makes data governance and access controls more important. If staff can see information they should not be able to see today, AI can make that existing problem easier to exploit.

AI embedded in nonprofit platforms

Nonprofits may already have AI inside their CRM, fundraising, marketing, accounting, meeting, or support platforms. Bloomerang’s Penny can surface recommended contacts, prepare meeting briefs, and draft outreach for review. Fundraise Up uses AI in parts of the giving experience. Blackbaud offers AI chat, predictive insights, and its Development Agent in Raiser’s Edge NXT, which the company says can plan and carry out donor-engagement work with human supervision. Givebutter includes an AI text assistant and a ChatGPT connection to authorized account data. Virtuous Momentum prioritizes contacts and drafts outreach. Confirm current availability and permissions for each product before adopting it.

These products do not all work the same way. Penny and Momentum illustrate embedded assistance. Fundraise Up applies AI within the giving experience. Givebutter’s ChatGPT connection links a general assistant to an authorized account; that connection alone does not make it an autonomous agent. Blackbaud’s Development Agent is an example of an agent embedded in a nonprofit platform.

Embedded AI can be easier to adopt because it is connected to an existing workflow. It may also be overlooked. Before buying another product, determine which capabilities are already included in the technology stack and whether staff understand how to use them.

Workflow automation

Automation follows defined rules: when something happens, perform a specified action. For example, when a form is submitted, create or update a record, notify a staff member, add a task, and send an acknowledgment.

Traditional automation is not necessarily AI. It is still one of the most valuable places for nonprofits to begin because it reduces repetitive work and creates more consistent processes. AI can be added where judgment, summarization, classification, or content generation is useful.

AI agents

An agent can pursue a goal by planning and completing multiple actions, often across applications. Meta describes Muse as a personal agent that can use a browser and connected apps to work on tasks such as filling forms or arranging travel, asking for approval before sensitive actions. Blackbaud describes its Development Agent as a supervised fundraising agent inside Raiser’s Edge NXT. The first is a consumer example; the second illustrates how agentic work may appear in a nonprofit’s existing platform.

The word ‘agent’ is used loosely. Some products follow a prescribed automation with an AI step. Others decide which steps to take, use several tools, and adjust as they work. Ask what the tool can observe, decide, and do independently, what requires approval, and where its actions are recorded. Greater autonomy calls for tighter permissions, testing, logging, and oversight.

Product sources (checked September 29, 2026): Meta, https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/ ; Blackbaud, https://www.blackbaud.com/newsroom/article/introducing-development-agent . Recheck availability and configuration before publication.

A simple way to evaluate the difference

Ask four questions:

1. What information can the tool access?
2. What actions can it take?
3. How much discretion does it have?
4. Who reviews its work before it affects a donor, beneficiary, employee, or financial decision?

Those questions matter more than the label on the product.

A few nonprofit examples

• A development officer uses a general-purpose chat assistant to create three possible outlines for a year-end appeal, then writes and reviews the final message.
• A staff member uses an embedded assistant in Microsoft 365 or Google Workspace to summarize a lengthy internal document the employee already has permission to access.
• A CRM identifies records that may be duplicates or summarizes a donor’s recent history for staff review.
• An automation creates a task and sends an internal notification when a major-gift form is submitted.
• An agent researches funders, checks eligibility, and prepares a draft list for review. A more capable agent might propose CRM updates or outreach, but staff approve consequential changes and messages before they happen.

What you can do today

• Ask staff which AI tools they are already using, including personal accounts and AI features inside existing products.
• Choose one current use and identify what information the tool can access, what action it can take, and who reviews the result.
• Review the Fundraising.AI Framework with the staff members responsible for fundraising, technology, data, and organizational policy.
• Choose one tool your organization uses and determine whether it functions primarily as a chat assistant, embedded copilot, automation, connector, or agent.

Responsible AI questions to consider

• Would a donor, beneficiary, employee, or volunteer understand how this AI is being used?
• Can a person review, correct, and override its output before anything consequential happens?
• Does the use protect privacy, treat people inclusively, and create enough social value to justify its environmental and organizational cost?

Fundraising.AI emphasizes privacy, security, data ethics, inclusiveness, accountability, and transparency in the use of AI for fundraising. Those principles become increasingly important as a tool moves from helping a person create a draft to taking action on an organization’s behalf.

The goal is not to select the most advanced form of AI. It is to choose the least complicated approach that safely solves a worthwhile problem.

Next ThursdAI: What does AI actually cost a nonprofit? We will compare nonprofit pricing and programs from leading providers, examine costs beyond the license, and help you determine what your organization may already be paying for.

Sign up to be notified when the next ThursdAI article goes live!

—

Cathexis Partners helps nonprofits assess their technology environment, improve business processes and select platforms based on documented organizational needs. Contact us to discuss your next step.