Getting started with AI: What non-technical staff need to know

AI is moving faster than any of us can keep up with. If you feel overwhelmed, under-qualified, or just too busy to figure out where to start, you’re not alone.

As non-technical staff at Candid, we were in a similar situation when we applied to the Decoded Futures cohort program, a three-month in-person learning cohort for nonprofits, hosted by Tech:NYC and supported by Robin Hood, Salesforce, Google.org, the Altman Foundation, OpenAI, and AWS. Our goal was to commit dedicated time to exploring the capabilities of AI and how to use it in our work.

We’d like to share some of what we learned to help fellow non-technical nonprofit staff get started on their AI journey.

Start with a basic AI policy and one tool

To get started, choose one generative AI tool aligned with your organization’s AI policy. If your organization doesn’t have a policy, begin by creating a personal AI use policy. A policy ensures you’ve thought through the risks of using AI, established clear boundaries around data sharing, and defined which tools to use.

Being part of the Decoded Futures cohort narrowed our options from the start: We knew we’d be using ChatGPT, one of the most widely used generative AI tools available today. Other options include Claude, Copilot, and Gemini. Focusing on one tool first helped us build foundational knowledge before expanding to test others.

AI needs to fit your workflow

A common mistake when using AI is to force your workflow to fit the AI. But really, it should be the other way around. AI will only be useful to you if it’s solving a real pain point in your work or for your organization.

Outlining the steps in your workflow makes it much easier to determine where AI can bring value. Only then can you decide whether you need AI to help with answering questions, automating tasks, drafting communications, sorting through information, or providing deep research.

So, we went through the process of documenting and identifying the areas of our work where AI might be helpful (even if we didn’t fully know how to use AI). We decided on two projects that involved analyzing and synthesizing large amounts of information and one project to automate data entry. Once we pinpointed our workflow bottlenecks, it was much easier to experiment to see whether AI could be a solution to that problem.

Test to understand how AI works

In our work, we are required to monitor a high volume of legislative and regulatory developments. Manually reviewing and synthesizing this information is time-intensive, and it’s difficult to review them in real time to surface emerging trends. We wanted to test whether AI could help us with this review process.

Using a prompt, we fed ChatGPT 70 URLs, requested a summary using the “deep research” option, and within minutes we had 14 pages of synthesized content. What we didn’t expect was that the back-and-forth of refining our prompts to get the final product would itself become the lesson. Each adjustment taught us something new, and within a few hours of testing, we didn’t just have a document but a working knowledge of how to prompt effectively (for example, providing context and being specific with instructions), and a better understanding of the capabilities of the tool.

Learn with others

One of the aspects we valued most in the Decoded Futures cohort was the built-in learning community. Each nonprofit team was assigned a volunteer technologist, and learning was done in-person, allowing for plenty of hands-on experimentation. At the end of the cohort program, several of the nonprofits demonstrated their prototyped AI solutions.

Here’s the takeaway: As your organization embraces AI tools, make sure to create a learning space where non-technical staff can benefit from the collective knowledge of others. This can mean finding opportunities for in-person trainings like those offered by Decoded Futures, or online trainings like the AI Fluency for nonprofits course from Anthropic. It could be creating an internal platform, such as a community of practice, a Slack channel dedicated to AI, or opportunities for AI-knowledgeable staff to mentor colleagues who are still new to AI.

AI technology is constantly evolving, and staff who are new AI for nonprofits can sometimes feel like they’re falling behind. Making use of and finding inspiration in what others have done can help make the learning curve less steep.

Nobody has AI fully figured out, not even the experts. You don’t need to have it all figured out before you start, and you don’t need a technical background to make meaningful progress. What you need is a clear problem worth solving, the willingness to experiment, and people to learn alongside. The nonprofit sector has advanced through collaboration and shared knowledge; AI is no different. Start small, stay curious, and bring others with you.

Photo credit: Clint Bush