Algorithms, automation, and AI (artificial intelligence). What do you think of when you hear those terms? What images come to mind? How do they feel in your body?
If you’re like me, you may feel equal parts excited and concerned. Excited about the possibilities for new discoveries and insights, and concerned about what this means for the future of evaluation consulting, data privacy, and human connection. I can enter the rabbit hole of fear pretty quickly when I don’t fully understand something, feeling the worry tighten my shoulders and shorten my breath.
I feel this deeply on a personal level every time new diabetes management technology is released. The most recent being automated insulin delivery systems. Relinquish control of my health to an algorithm and AI? That sounded terrifying. What if the technology glitched? We’re talking about literal life and death outcomes.
But the reality is that the newer diabetes technology has significantly increased the quality of my life, my glucose control, and my physical and mental health. I wear an insulin pump that syncs with a continuous glucose monitor. Every 5 minutes, these two pieces of technology (see image below) talk to each other to make micro-adjustments to determine what amount of insulin to dose based on my glucose level. I can view everything via an app on my phone, which alerts me when things go a bit (or a lot) wonky. The results also transfer over to my physicians, so they can view data in real time. It’s bionic and futuristic in the best way possible.
What I was originally fearful of has demonstrated its value. Is it perfect? Absolutely not. The algorithm doesn’t know everything.
For example, I still have to tell the insulin pump how many carbohydrates I’m consuming and when. It doesn’t know if this was a regular meal or a holiday feast, or if I’m planning to go hike up a mountain right after. There’s still a level of human interaction and feedback that requires my expertise and input. But when we work together, the outcome is better than before.
So I wondered, how can we leverage AI to support our data and evaluation work?
What might this partnership with AI look like for consultants, clients, and colleagues? Well, beyond asking them to write love letters to evaluation. Could it support evaluation capacity building? Assist with coding qualitative interviews? What are the ethical, financial, and logistical boundaries and considerations?
First, let’s consider general business and organizational operations. Below are some free or low-cost automation and AI tools that I have incorporated into my consulting practice:
- Asana: Ok, I’ll admit, I was super against onboarding a project management tool. My thought was that if I need a tool to guide me, then I’m probably doing too much. Seeing a trend here? Turns out, tools are there to support us, not judge us. With Asana, I can input all the big tasks for projects, including reminders for when things are due. It also allows me to collaborate with colleagues and subcontractors in real time. Other folks use tools like Notion or Todoist.
- Canva: I’m a long-time user of Canva to create social media and report graphics. Recently the tool has entered the AI space with their Text to Image App and Magic Write function. The text to image app creates an AI-generated picture based on words or sentences that you provide, which can be a helpful alternative for report images.
- ConvertKit: ConvertKit is an email marketing platform that allows me to send out one-time newsletters as well as email sequences based on certain criteria like downloading a freebie or attending one of my evaluation trainings.
- Google Analytics: Google Analytics is a free web-based service that offers insights about my website traffic, including how visitors are interacting with my site and looking at website conversions. This helps me to understand what is working the way I expected and what content visitors are most interested in overtime.
- Grammerly: Grammerly is an AI writing tool that gives real-time suggestions about grammar, spelling, and punctuation. I have this linked to my web browser, where it provides helpful feedback on my emails and blog posts. I especially appreciate its love for the Oxford comma.
A few tools that I am planning to explore further include:
- Descript: Descript is an AI-powered transcription tool that can help make your content more accessible by providing transcripts and captions for podcasts, audio recordings, virtual trainings, etc. Many of my colleagues use other AI-based transcription apps like Otter.ai.
- Excelformulabot: This is an add-on for Excel and Google Sheets that helps users generate and understand formulas, SQL queries, and VBA code.
- Polymer: Polymer is a no-code required platform that transforms, analyzes, and visualizes data, similar to tools like Tableau, PowerBI, and Datawrapper.
- Zapier: Using a bunch of different tools is great but can quickly become overwhelming if they don’t work together. Zapier helps link siloed systems so that your tools can work together towards your desired outcome.
In the next post, I talk more about how we might leverage ChatGPT when planning and implementing program evaluations.
Before you head over there, comment and let me know what tools and automations you’re using in your work now and which ones are on your list to try.
Whatever you choose, remember to review the level of data privacy and transparency around decision-making so that you understand the security and ethics of the tools.