“Artificial intelligence is a machine’s ability to perform some cognitive functions we usually associate with human minds.” (Mckinsey) 

Want a better understanding of AI? Imagine talking with a toddler. They sometimes say silly and confusing things that reflect how they perceive the world with their limited experience. The same could be said for early adoption of AI. 

AI solutions iterate and get better as they learn, helping organizations become more efficient and more effective. But if you don’t thoughtfully create a process and properly implement it, AI could do more harm than good. 

Copying and pasting AI-generated text in an email to your donors could be disastrous. But, if you train a generative AI product with past donor communications and responses, you’ll get better results that serve as quality first drafts.

AI readiness

It’s important to first define success for any initiative; that’s why strategic planning is critical. You need to understand what you want to achieve and create milestones for success. Similarly, do you have the right people with the right skills and the right tools to achieve those goals? Here are some questions to answer as you consider your own AI readiness. 

#1. What are your goals?

Lewis Carroll wrote, “If you don’t know where you are going, any road will get you there.”

Using AI just to use AI likely won’t meaningfully advance your fundraising. Goal setting will help you choose which road you take to achieve AI success at your organization.   

Your current fundraising/advancement plan is a great place to start. The goals and strategies in the plan should highlight potential uses for AI and desired outcomes. Most organizations have existing goals to increase or maintain donor retention as part of their fundraising plans. This might represent an opportunity to use AI in some of the following ways:

• Segment which lapsed or almost lapsed donors would respond best to different channels such as email or text messages.
• Identify the best time of day or day of the week to send messages to different donors.
• Analyze past messages and generate communications that are most likely to prove successful at engaging your donors.

But don’t focus on specific tactics as described above early on. Start with goals such as, “We will use AI to help grow donor retention by 3% this year.” Those more sophisticated ideas can be used later to help identify which AI solutions could be useful for your organization.

You want to make sure stakeholders and people who will likely implement or use AI solutions have an opportunity to provide input and share their concerns. The details of how you use AI to fundraise don’t need to be the topic of conversation at a board meeting, but you should gain enough internal feedback that the director of development or CEO could have a meaningful discussion with an AI enthusiast on the board.

#2. Is your data ready?

If your organization has and adheres to a data usage manual, your data should be in good shape. If you don’t have clear policies and procedures for using different technology systems, start there before focusing on AI. 

You don’t have to have a data warehouse to effectively tackle AI, and even top-notch organizations have holes in their data. Many people-focused development officers forget to enter all their activities in the CRM, for example.

Although your data doesn’t have to be perfect to be effective, there a few things that should be in place:

1. Do you know what data you have and where you store it? Data is the basis for AI to make decisions and recommendations. If you don’t know what data exists or where it is stored, it’s going to be hard to provide any AI solutions – and it could be hard for you to make decisions even without AI. If your organization has a data usage manual and security/privacy policies, this will be easy. If not, this is a vital step in understanding your organization’s data vulnerabilities and making effective use of any AI solution you wish to implement.
2. Are there standard codes used across the organization? For example, if you haven’t always used appeals codes or have changed your appeals codes, make sure your older appeals follow the new code system. If your appeal codes (or any codes for that matter) are consistent, AI will make better decisions because it can more easily pick up on patterns. If AI sees an unusually high open rate for your annual tax receipt email and that message is improperly coded, the recommendations are going to be counterproductive. 
3. Have you checked for duplicates? At some organizations every time someone moves and makes a new gift, they are entered in as a new record. This is problematic not only for your use of AI but also because you can’t effectively thank your donors for their lifetime of gifts. Most modern technology solutions have tools that make it easy to check for duplicates so you can review before merging data. This will help AI understand donor giving patterns and might help your donor retention rate. 
4. How do you track pledge payments? It’s important you differentiate pledge payments from distinct gifts. When someone agrees to a $30,000 gift and pays $10,000 a year for three years, that is different from a donor making three separate $10,000 gifts. A repeat $10,000 donor might be distinctly different from someone who gives $50,000 once. A human can easily look at a record and see the difference in contact notes, but most AI solutions are not sophisticated enough to identify and correct this difference when making decisions. 

#3. Who will implement this?

Just like it’s important to have clear goals for anything, it’s important to have responsible parties. When it comes to AI, there often are multiple responsible parties. A chief data officer, chief technology officer or chief privacy officer needs to be in the middle of the process, but they may not be the ones to implement the solution. In fact, someone on their team may be responsible for part of the implementation, and there will likely be a different end user or set of end users responsible for achieving the goals with that solution. 

If you are a large organization, you may partner with a nonprofit technology services firm or a solution implementor to help you consider internal processes, pinpoints, opportunities, etc. Working with a third party helps to mitigate risks and ensure there are fewer stones left unturned, but it isn’t a requirement. You can do many of these things on your own.

If you are a smaller organization, you might not have the luxury of a software implementation firm or even much support. At smaller organizations, there are rarely people responsible for organization-wide technology. If that is the case, you will need to rely more heavily on your AI fundraising technology provider for a successful implementation. Remember, you’re rarely alone and many of your peers in/at AFP, AASP, CASE, AHP and other professional associations likely have already been there and are willing and able to share their experiences to help you. 

#4. Are your people ready?

Are you ready to pay six figures for an AI scientist? Probably not, but the good news is you don’t have to. Bringing on an AI scientist makes sense for research companies and large sophisticated organizations with well-defined AI goals, but there are enough AI fundraising solutions that you don’t need someone with those refined skills. 

Anyone involved in the process should know some basic elements of AI, and those more involved should take a deeper dive. If nothing else, as a buyer of technology, you want to be able to ask enough questions to understand what you are getting, how your data will be used, product limitations and a fair price. 

If you’re unfamiliar with terms like “Turing test” and “unsupervised learning” or can’t differentiate machine hallucinations from someone’s account of Woodstock, Fundraising.AI has a great glossary of AI terms two-thirds down the page. 

Remember too that AI is constantly evolving, as the landscape of available technology in our sector and at large is in flux. This means anyone tasked with AI will need to spend time continually learning about these changes. That doesn’t mean spending hours every week, but scheduled monthly professional development time and even professional development budget dollars will go a long way to ensure your team is properly equipped.

#5. What are your ethical guidelines?

This is one of the most important questions to ask in the entire process. Put simply, ethics is doing the right thing, and without ethics it is difficult for anyone to have trust. If your community can’t trust you, you don’t have a nonprofit future. It’s important to put guidelines in place that protect your donors and other key constituencies. The  Fundraising.AI framework is a great place to start.

Ethical guidelines should protect all your constituencies and, in doing so, protect the organization itself. A data usage manual and data security and privacy policy are foundational, and all technology used at your organization – be it fundraising technology, volunteer management software, email marketing solutions, etc. – should be covered. AI is no exception. 

There are specific concerns associated with AI, and an AI usage policy will help address them. There are obvious concerns such as, “What will a vendor do with my data and anything learned from my data?” that should fall under broader policies but may have specific implications within the context of AI. For example, many organizations may collect and store age, race, gender, religion and even shirt size. It’s important to consider how, if at all, that information might be used for AI. 

This all might seem like a lot to take in, but these are important decisions. If you can answer these questions, you’re ready to dive into AI solutions at your nonprofit. In many ways, AI is like other nonprofit fundraising technology: As long as you have a plan, set goals and understand they will need to change as you implement and learn, you’ll be well suited to use AI to help your organization succeed.