AI, Machine Learning and Philanthropy
By Erik Rubadeau
Don’t panic, artificial intelligence isn’t here. Yet.
There is no denying that artificial intelligence has arrived as a topic of conversation in our sector. But will AI and Machine Learning be the next evolution of the nonprofit organization or will it simply be a shiny new object that will distract us from more important (but less shiny) priorities? In my opinion, the answer is both.
Attend any conference these days and you will no doubt see sessions on artificial intelligence, machine learning, chatbots and more. I myself held a session on the potential of chatbots at the DMAW Digital Day earlier this year.
From keynotes to masterclasses, many thought leaders in our sector have started to pontificate on the arrival of artificial intelligence, some going as far as to say that organizations that fail to harness artificial intelligence soon will be left behind.
Personally, I’m not convinced about this doomsday prophecy, and frankly I think it only goes to further destabilize the priorities of organizations who are already trying furiously to keep up with the near instantaneous expectations of the modern consumer.
Yet I do believe that organizations need to start opening internal dialogue about artificial intelligence sooner rather than later. In order to properly leverage any new technology you need to have a plan for it, and a well considered technology plan that is tailored to your organizational needs is not crafted overnight.
Just like the recent technology waves that have come before it (responsive websites, mobile payments etc.), AI will begin as an opportunity for those organizations who are in a position to experiment with early adoption. Some of this is happening already. And from the groundwork of these early adopters will come the learnings of potential applications for AI for the rest of the sector. Yet, if the integration of AI into our existing daily life is any cue, it seems unavoidable that AI will weave itself into the fabric of fundraising over the coming decade.
As someone who believes very passionately in the power of technology to move the nonprofit sector forward, I can also acknowledge that technology has the capacity to be an active hindrance to this goal.
When technology is implemented poorly, be it a CRM, an email marketing platform or an artificial intelligence toolset, it will cause more harm than good.
Even more disappointing is when organizations don’t invest in training or hiring the right staff to manage a new technology. Gone should be the days of tossing your digital keys to the newest hire or youngest person in the office and expecting them to sort out all your technology problems.
If you follow Avinash Kaushik (analytics genius and all around smart guy), you will be familiar with his concept of 10/90 when it comes to technology and people. Technology costs (subscriptions, licenses, etc) should be 10percent of your budget, and with the other 90percent you should be prepared to spend it on the people required to effectively use and maximize your technology investment. Tools, after all, are just tools without hands to use them.
Unless you invest in training your staff, hiring new staff, or partnering with an agency to support your technology needs, your new technology investment can end up slowing your organization down, or worse, sending it backwards a few steps.
For these reasons, you have to have a well considered plan for new technology adoption, and with artificial intelligence, the gap between those who prepare and those who don’t will be even more pronounced.
So how do we go about crafting a plan for a technology we may not fully understand? The first step is to get away from all of the surface level generalizations that are being tossed around and to begin to deepen our understanding about the basics of AI.
Let’s take a minute to more clearly define what we mean when we say artificial intelligence. According to a 2018 paper by the Brookings Institution, artificial intelligence systems contain three characteristics: intentionality, intelligence and adaptability. And you may also recognize these three characteristics as somewhat human. This is what artificial intelligence is aiming for. One definition defines AI as the simulation of human intelligence processes by machines.
Any way you cut it, AI is a computer program trying to make a decision about the question it has been asked based on the data that is has to work with. As the program gets positive and negative feedback, it will learn from these results and from the context in which the feedback was provided. Over the course of millions of simulations it can get pretty good at learning all about the specific task or theme. Sounds impressive, and it is. Yet it’s important to know that AI has clear limitations: just because the program has become an expert on riding a bike doesn’t mean that it will ever know anything about skateboarding or driving a car.
In its current format, artificial intelligence is not a replacement for human fundraisers and won’t be any time soon. The robots are not coming for our jobs. AI is not an overarching software overlord, or a massive and expensive technology initiative that needs to be rolled out across the entire organization in order for your nonprofit to survive the impending evolution. It is also not a silver bullet to solve your existing technology or data issues.
If it’s not a silver bullet, what is it and what benefit can it actually be for your organization?
Artificial intelligence can be used to solve repetitive or data intensive problems at the individual level. Every role in your organization is filled with tasks that would benefit from automation, predictive insights and computer assisted analysis.
Essentially by helping to solve for the problems of the individuals inside the organization, AI will help to further the overall efficiency of your organization by freeing people’s time from repetitive tasks, or improving their efficiency by mining their data sets and providing predictive analysis or promoted next actions based on identified patterns.
On the flip side, AI can also improve your customer service and donor experience interactions by reducing friction and increasing the speed of response to a host of regular actions and questions.
Perhaps now we can begin to see how AI will help improve our organizational efficiency. There is however one very important barrier to entry to these benefits: good, clean data.
Data is the lifeblood of artificial intelligence. AI, and more specifically machine learning, rely on large sets of data to help find patterns, opportunities and predictive next steps from your fundraising database.
If your data is no good, then your predictive outcomes will also be no good. Google’s Avinash Kaushik has a simple phrase to describe this phenomenon: ‘garbage in equals garbage out’.
Data hygiene is the key reason why the AI conversations need to start happening in your organization sooner rather than later. If your data house is not in order, and you’d like to be in a position to benefit from AI advances over the coming years — there is no greater short term priority than getting your data in order.
The key to your AI success will hinge on the quality and quantity of your data.
In a world where we are now able to capture so many data points, the primary barrier to action is no longer acquiring the data itself, but rather the analysis paralysis that comes from being so overwhelmed by the size and depth of these massive data sets that you will otherwise have to sift through by hand to try to identify a new opportunity.
This is where artificial intelligence in its current format can really shine.
You only have to look at the nonprofit tools like Wisely, boodleAI, Gravyty, Accessible Intelligence and others to see some of the interesting artificial intelligence products and tools that are currently pointed in the nonprofit sector.
Here’s a few of the tools I’m watching with interest:
Wisely is a prospecting tool that makes it easy to find and assign compelling donors to a gift officer’s portfolio. Full disclosure, I know these guys and I think they are doing really cool work. Wisely’s Prediction Engine sifts through your database looking for your next opportunity and prompts you with who you should be speaking with and what giving level they might best be suited for. This is the kind of insight that can really drive new value for mid-level and major giving.
boodleAI leverages proven AI and machine learning to rapidly model the untapped data sitting in your nonprofit database along with billions of third-party data points to help you achieve significant lifts in donor acquisition, retention, and engagement rates. Essentially, they drill into your prospect audiences and look for potential donors based on a set of interests and ideal donor traits. They help you identify prospects who stand the best chance of being your next donor.
Accessible Intelligence is an AI spin-off from the team that brought you Engaging Networks. I’ve seen Graham speak on the work they are doing a few times now and am quite excited by the potential. Their tools will look to help you write the right kind of copy and provide the right offers to the audience segments at the right times in your lists.
Gravyty has a suite of AI tools for non profit organizations. The one that caught my eye was the ‘First Draft’ tool which identifies the best donors at the right time for discovery, cultivation, solicitation, and stewardship. It then provides a draft email for you to edit and send to the donor. It learns your tone and style, adjusting to match your unique organizational voice. The tool sends new suggestions daily directly to your email, no matter which service you use.
Each of these tools focus on a specific pain point of the fundraiser’s role. Yet, the thing to realize is the very singular focus of each of these tools. They each solve for a very specific task. Additionally, these applications are only the tip of the iceberg. This list doesn’t even begin to touch the realm of chatbots, guided digital assistants or even what can be done to improve workflows in other areas of your organization.
Over the coming decade I have no doubt that artificial intelligence will slowly work its way into the fabric of everything that we do as fundraisers. Yet until then, there are opportunities to begin to leverage now. Perhaps the most important initiative to undertake in the near term is to enhance the quality of your data.
Do I believe that artificial intelligence can solve all of our problems?
But I do believe that the application of artificial intelligence tools in well considered scenarios can lead to positive outcomes that help to solve for the perennial pain points of capacity, money and time.
Erik Rubadeau is CEO of Yeeboo Digital, a data driven technology and full service online fundraising and technology agency.