Last week I wrote about keeping AI open all day to catch the shovelling tasks. It’s a good habit. If you’re the only one in your organization who has it, though, you end up as one fast-moving person on a slow-moving team.
This post is about the next step, moving from personal productivity to the capacity of your whole organization. That means getting your institutional knowledge out of PDF handbooks and out of your ED’s head, and into a place where everyone on staff can use it.
Why onboarding is a bottleneck
Imagine Canada’s recent report on AI adoption in Canadian nonprofits found that only 12% of nonprofits use AI for staff development and training. Most organizations still onboard people the slow way. A new hire starts, and their manager spends four hours explaining policies. The new hire can’t hold onto it all, then spends the next two weeks searching for the right handbook.
New hires can only learn as fast as their supervisor can teach them, and in a small shop that supervisor has a full job of their own.
An internal knowledge base your staff can ask
Instead of handing people a static document, point a secure, private AI assistant at your staff handbooks and program manuals. A new hire, or someone who’s been there ten years, can ask “How do we handle expense reimbursements for out-of-province volunteers?” and get an answer right away, with a citation to the page it came from.
It’s good for your team, too. Staff can answer their own questions, and it cuts down on the admin drag that leads to burnout.
A mentor for new program leads
You can go further with a new program lead. Give an assistant access to your internal documents, like past successful grants and program reports, and a new hire can:
- Review three years of program history in about 15 minutes.
- Draft their first impact update in your organization’s voice from the start.
- Check their work against your compliance rules as they go.
Onboarding then happens alongside the real work, instead of in front of it. It’s a bit like giving every new hire a mentor who’s always available. That’s how small teams start to work like much bigger ones.
Budgeting as a conversation
The same idea works at the director level. For a lot of directors, budgeting is a yearly headache spread across 15 different spreadsheets, and all of it runs through one person.
With an assistant sitting in on your planning sessions, the budget becomes something you can talk to. Instead of staring at a static spreadsheet, you ask: “If our core grant is cut by 10%, but we raise our individual giving goal by 15%, how does that ripple through program staffing for Q3?”
A good assistant will run the numbers, then suggest options and point out gaps in the plan you might have missed. That’s where AI starts helping directors with the decisions themselves.
How to set up an internal knowledge base
Here’s how I’d start:
- Collect the documents people ask about most. Your staff handbook, expense policy, program manuals and a few strong past grant applications and reports. A small set of current documents works better than everything on the shared drive.
- Clear out old versions. If three versions of the expense policy are floating around, the AI will find all three. Keep the current one and archive the rest.
- Use a private, approved tool. Pick an AI tool your organization has approved for internal documents, with settings that keep your files out of model training. Check who on staff can see what before you load anything sensitive.
- Ask for citations every time. Set it up so every answer points to the document and section it came from. Staff can check the source in seconds, and wrong answers are easy to spot.
- Test it with real questions. Ask your most recent hire for the questions they had in their first month. Run them through the assistant and see what it gets right.
- Name an owner. Someone needs to update the documents when a policy changes. Otherwise the assistant will give confident answers based on last year’s rules.
For budgeting, start with a copy of last year’s budget and one what-if question. Check the math against your own spreadsheet before you lean on it for a real decision.
For leaders: get knowledge out of people’s heads
In a lot of small nonprofits, the most important knowledge lives in the ED’s head. That works until the ED is on holiday, or a new hire needs an answer at 4 p.m. on a Friday.
A shared knowledge base gives everyone the same answer to the same question, which helps with consistency and with compliance. A few questions to ask your team:
- Where do our policies and procedures actually live today?
- How long does it take a new hire to answer a basic policy question?
- Which of our documents are out of date right now?
- Who will keep the knowledge base current when things change?
Once the knowledge lives somewhere everyone can reach, each person’s AI habits start helping the whole team.
