Eating Our Own Dog Food - How and Why We Built PolicyBuilder v2

Could we make an existing service offering more widely available to the nonprofit community by reducing the cost it takes to deliver, and therefore reducing the cost of the service?
Challenge accepted.
Linda Widdop, Tech Impact’s CIO and one of its longest tenured team members, challenged her team to develop a new approach to its Policy Builder service, a low-cost offering that helps nonprofits create a Computer Use policy through a short workshop or a semi-custom consultation. Computer Use Policies are important – it’s the first step to securing an organization’s data, protecting its assets and its employees, and ensuring compliance. Many nonprofits do not have a policy, or have a very outdated one that has not been kept up with the ever-changing technology landscape. Even with its high importance and its relatively low cost at $250, the service wasn’t popular. Nonprofits regularly don’t have a strong IT budget or expertise, and resort to searching the web for templates to use that may not match their specific risks, needs, and environment.
Its importance to nonprofits’ operations wasn’t our only reason for choosing it for this challenge – we also wanted to see if we, Tech Impact, could use automation and AI to replace the human effort needed to deliver this service. Now, we aren’t interested in reducing our workforce by moving our services to computer-led processes – it's quite the opposite. We have been promoting the benefits of automation and AI to expand services and provide more resources for years, not to replace our team members. But in this example, dedicating an employee’s time to doing this manual task takes them away from more intensive and personable that a computer can’t do.
How can we sell automation and AI as technologies that will help others when we haven’t tested the benefits ourselves? It was time to eat our own dogfood, as they say.
Tech Impact’s Data Lab team took the challenge and ran. They diagrammed a process that would use a survey tool (Qualtrics) to capture input from the nonprofit that included standard and variable data required for each section of the policy. Then, they used scripts to import that data in Azure to compose a draft policy and send prompts to Generative AI to customize each section. The output – a customized policy with annotations where we recommend review - is emailed to the nonprofit within two business days.
Did it take a long time to build? Yes. Were there a lot of learning curves? Yes. Is the tool able to be manipulated easily to streamline other policy building services? Yes!
The results were pretty remarkable. We took a service offer, PolicyBuilder, that cost the nonprofit $250 and involved an average of three hours of our staff time to produce, down to a FREE service to the nonprofit and minimal time for our staff to manage through the use of automation. Now, we’re busy building out other offerings, including Disaster Recovery and Incident Response planning, that will be coming soon.
We hope our challenge can give you some inspiration to experiment in your own nonprofit. We even have coaching and training support on platforms like Power Automate to get you started. As always, reach out to us if you have any questions.