A leader at an established fiscal sponsor recently asked me: ‘How should we be using AI?’ It’s November 2025, and questions like this are coming up for executive directors and CEOs more and more frequently.
There are news stories every day about how AI is going to be helpful – or harmful – to individuals, companies, employers or society as a whole. In general, there’s a lot of hype and not a lot of substance.
Pair that with the fact that the technology is rapidly evolving, with new and updated tools being released weekly, or even daily – and there’s a lot of confusion right now about what these tools can do, how secure they are, what they cost, and how to implement them effectively (either on an individual, or organizational, basis).
Within the nonprofit sector, there are already individuals and organizations trying to stake out their place as “AI for nonprofits” experts, ranging from individuals and small businesses to large corporations. So far, most of them seem to focus on using AI to generate content of some form or another – but none of them have focused on fiscal sponsorship.
With that in mind, Schulman Consulting has put together the following “Field Guide” for late 2025. We anticipate the content of this guide will change dramatically over time, but wanted to share our current best thinking.
From our perspective, the main use cases in fiscal sponsorship for the current AI tools (as of early November 2025) are:
- Content Generation/Research – Generating/editing text, images, videos, slides, even websites, apps, code or strategies/plans of action
- Custom Chatbots – Creating specific “bots” that “know” your organization and can converse with internal and external audiences
- AI Agents/Assistants – Using AI to assist on tasks that are tedious, mundane or just difficult for humans; these “assistants” can accomplish specific tasks as instructed and can be linked together and overseen by a single assistant to produce more complex workflows
As mentioned above, this is very much still an emerging technology, but there are already a number of areas of potential value for fiscal sponsors and their staff. Based on conversations with FS staff, others in the sector, and those in the AI field, we’ve laid out some potential examples specifically for fiscal sponsors within each of the three categories above:
Content Generation/Research – Using AI to Create Content and Surface Information
| Example 1: | Using publicly available chatbots for research and learning (Spoiler alert: Individuals within your organization are almost certainly already using AI for this purpose today) |
| Level of complexity: | Low; just ask a question and get a response |
| Improvement over previous tools: | Significant |
| Pros: | Get deeper answers than from standard web searches; available “off-the-shelf,” some are free to use or relatively inexpensive (usually about $20/month for individual plans) |
| Cons: | These chatbots aren’t always correct – often need to double-check facts and info elsewhere; if using a free plan, these services will likely train their models on your data/information |
| Current Providers: | ChatGPT, Claude, Google Gemini, Perplexity, many others |
| Non-AI Alternatives: | Web search, other research |
| Example 2: | Reviewing incoming emails to 1) automatically categorize them, 2) prioritize that need a response, and 3) generate a draft response |
| Level of complexity: | Low-ish; requires some setup and connection to email provider (i.e. Google, Outlook, etc.) |
| Improvement over previous tools: | Significant |
| Pros: | May allow your human staff to get through more requests in a given time period; available “off-the-shelf,” some are free to use or relatively inexpensive (usually about $20/month for individual plans) |
| Cons: | Every reply still needs to be reviewed by a human; if using a free plan, these services will likely train their models on your data |
| Current Providers: | Superhuman, Cora, Fyxer |
| Non-AI Alternatives: | Canned responses (in Gmail), copy and paste template responses from a document |
| Example 3: | Developing marketing content (text, images, slides, even videos) |
| Level of complexity: | Low-ish; upload information and instructions and get back a response |
| Improvement over previous tools: | Significant |
| Pros: | May speed up the process and it often helps to start with something other than a blank document; may provide new ideas and perspectives; available “off-the-shelf,” some are free to use or relatively inexpensive (usually about $20/month for individual plans) |
| Cons: | For best results, needs to be trained on your organization’s voice and goals; if using a free plan, these services will likely train their models on your data |
| Current Providers: | ChatGPT, Claude, Perplexity, Google Gemini |
| Non-AI Alternatives: | Have staff do it and/or hire a marketing resource |
| Example 4: | Thought Partner/Strategist |
| Level of complexity: | Low-ish; upload information and instructions and get back a response |
| Improvement over previous tools: | Significant |
| Pros: | A partner with access to most of the internet; available “off-the-shelf,” some are free to use or relatively inexpensive; may provide new ideas and perspectives |
| Cons: | Needs to be trained on any relevant info/data about your organization (if not, you get fairly generic responses — which can be helpful…sometimes); if using a free plan, these services will likely train their models on your data |
| Current Providers: | ChatGPT, Claude, Perplexity, Google Gemini |
| Non-AI Alternatives: | Start with a blank page; spend more time brainstorming; have a more limited pool of resources to draw from |
Custom Chatbots – Creating an AI to answer questions and/or provide specific information
| Example 1: | Answer basic questions from prospective projects who visit your website and/or are filling out your online application form – external facing |
| Level of complexity: | Medium; requires some set-up and expense |
| Improvement over previous tools: | Meaningful improvement over earlier, non-AI chatbots |
| Pros: | Allows you to answer questions from prospects without having to use as much staff time |
| Cons: | Requires upfront time, effort and budget to get things set up correctly; may still need to check the bot for accuracy |
| Current Providers: | Google NotebookLLM, Tidio, Collect.chat, custom development |
| Non-AI Alternatives: | FAQs on the website, answering emails from prospective projects |
| Example 2: | Talk to core staff to answer policy and procedure questions as needed – internal facing |
| Level of complexity: | Medium; requires some set-up and expense |
| Improvement over previous tools: | Meaningful improvement over earlier, non-AI chatbots |
| Pros: | Now everyone doesn’t need to reference the policy manual (or memorize it) |
| Cons: | Requires upfront time, effort and budget to get things set up correctly; may still need to check the bot for accuracy |
| Current Providers: | Google NotebookLLM, n8n, custom development |
| Non-AI Alternatives: | Referring to the Policy Manual |
| Example 3: | Talk to project staff to answer basic Policy & Procedure questions but also “where is this payment/contract/etc. in the process?” questions (would need to be connected to other systems like accounting and CRMs for it to work in certain cases) |
| Level of complexity: | Medium-High; requires more extensive set-up and possibly custom development |
| Improvement over previous tools: | Significant |
| Pros: | Allows you to answer simple/easy questions from projects without having to use staff time |
| Cons: | Requires upfront time, effort and budget to get things set up correctly; may still need to check the bot for accuracy |
| Current Providers: | Custom development |
| Non-AI Alternatives: | Staff time/emails; upgrading other systems so projects have more direct visibility into current workflows |
AI Agents/Assistants – Using AI to Fully Automate Activities
| Example: | Automated “team members” that are developed to accomplish certain tasks or groups of tasks. An example might be an AI that reviews and summarizes all incoming project applications, then assigns a risk score to each one based on predetermined risk tolerances embedded in the system by the sponsor. Another might be sending contracts or grant agreements to an AI to have it pull out the pertinent details and put them into a structured format (spreadsheet or database) for later use or reference. |
| Level of complexity: | High; requires very extensive set-up – likely custom development |
| Improvement over previous tools: | Very significant |
| Pros: | Allows you automate workflows in just about any area of your organization — think onboarding/offboarding projects, onboarding/offboarding employees, reviewing contracts/grants to ensure funds are being spent correctly, gathering information for monthly financials, or annual reports/reporting – freeing your team from mundane work and allowing them to focus on higher-leverage activities |
| Cons: | These likely require some custom development to setup correctly and maintain; also AI almost always requires a “human-in-the-loop” (i.e. human review) |
| Current Providers: | Custom development (initial expected budget: $5,000-$50,000) |
| Non-AI Alternatives: | Zapier, Make, higher level “robotic process automation” tools, manual workflows |
Please note: In addition to the above listed “cons” for each use case, significant environmental impacts of continued (and expanding) use of AI tools have been identified and will continue to grow as the use of these technologies becomes more prevalent. This is definitely something to take into consideration as you think about your organization’s adoption of AI technologies.
There are many, many more ways to use AI, many other tools out there than those mentioned above, and many improvements coming down the pike that will make a lot of this easier to use and produce more effective results.
Getting Started // Taking the Next Step
As mentioned earlier, people in your organization are already using standalone AI tools – and/or those embedded in your existing systems – so the train has already left the station to some degree. That being said, it’s in the best interest of most organizations to set up AI-related policies (and training) so everyone is clear on how your organization wants AI to be implemented and to try and mitigate risk around AI usage. (If you’re not sure where to start on that, the short, four-part policy listed in this article is a good place to start.)
After that, each organization has to decide for itself how/if it will utilize AI for use cases beyond individual employee usage. When considering to what extent your organization will adopt AI in the near term, I recommend answering a few questions:
- If we haven’t already, what policies should we set around current AI usage for our team?
- What level/amount of usage are we comfortable with?
- How secure are the tools our team is using? Are we comfortable with the risks/impacts involved (including environmental) and/or how can we put in safeguards to mitigate those risks?
- Where do we think this could be most helpful?
- How much of an investment in time and money is the organization willing to make at this point? (Keep in mind: even the best tools are only as useful if people are trained on them and willing to use them.)
- How might this all change in the next 6-12 months?
If you want to share about how your organization is using AI today and how you are approaching this rapidly evolving technology, please leave a comment below.
Finally, we have begun working with clients in this area, so if we can be of any assistance to your organization, please don’t hesitate to reach out.
