Preparing to Build with AI: Foundations for Small Teams

Written By Jonathan Coburn

What you need to know before creating your first AI-powered tool.

So far in this series, we've encouraged you to explore AI, try it out on small tasks, and start thinking about how it can support your work. You’ve probably seen how useful these tools can be for writing, summarising, or analysing text.

Now you might be asking:
Could we actually build something of our own using AI?

The short answer: yes.
And it’s likely simpler and more achievable than you think.

This blog is your practical guide to getting ready — what building with AI really involves, how to decide if your team is ready, and the essential ingredients you’ll need in place.

What Does “Building with AI” Really Mean?

When we say “building with AI,” we don’t mean designing your own model or writing lines of complex code. That’s already done for you by companies like OpenAI, Anthropic, Google, and Microsoft.

What you’ll be doing is using their models (like ChatGPT, Claude, Gemini, or Copilot) to power small, purpose-built tools that support your team, services, or beneficiaries. That could mean:

  • Creating a tool that screens and tags incoming grant applications

  • Building a simple assistant that answers staff questions about HR policies

  • Automating the analysis of open-text survey responses

  • Generating first drafts of reports based on project data

You're not inventing the AI — you're connecting your content and needs to it, often using no-code or low-code platforms.

What You Can Realistically Build

Small organisations are already building tools like:

  • A feedback analyser that categorises responses and flags themes

  • An internal chatbot trained on organisational policies and FAQs

  • A grant summary generator using past reports and application data

  • A simple form that auto-scores project submissions based on AI review

These tools can be built using easy-to-learn platforms like Tally (for forms), Glide (for mobile/web apps), Airtable (for storing structured data), and Zapier or Make (for connecting everything together). Most connect directly to AI models from OpenAI or Claude, often via an API — but the platforms themselves handle that complexity for you.

What You Need in Place Before You Build

You don’t need a dedicated tech team. But you do need a few things lined up before you get started.

1. A Clear, Useful Idea

Start with a single task or process that:

  • Takes a lot of time

  • Is repetitive or admin-heavy

  • Has a clear goal (e.g., triage, summarise, tag, draft)

  • Doesn’t rely on sensitive or confidential data in early versions

Try something you've already tested manually using ChatGPT or Claude. If it worked once, there's a good chance it could be automated or scaled.

2. Data or Content to Work With

You’ll need something for the AI to work on. That could be:

  • A spreadsheet of past funding applications

  • Word or PDF reports

  • Staff handbooks, policies, or training guides

  • Survey data or free-text responses

The data doesn’t need to be perfect — just accessible and relevant to the task you're solving.

3. The Right Tools for the Job

You don’t need to build everything from scratch. These platforms are especially useful for small teams:

You can start testing most of these for free or at very low cost.

4. A Person to Set It Up (or a Freelancer to Help)

This doesn’t need to be a technical hire. Many charities work with:

  • A digitally confident team member

  • A volunteer with low-code experience

  • A freelancer (via various platforms or services)

Even if you do use external help, it's important that someone in your team understands the basics of how the tool works and what it’s for.

Use No-Code, Low-Code, or Hire Help?

Here’s how to think about it:

  • If your tool is simple, internal, and based on existing data — you can likely build it yourself using no-code tools.

  • If the logic is more complex or you want to build something public-facing — you might need a bit of freelance help.

  • If you need a fully custom system or integration — you’ll need a developer or digital partner.

Start by building something small on your own. That way, if you later decide to invest in a more polished version, you already know what you want.

Common Traps to Avoid

You don’t need to learn everything.

You just need to log in, try something, and see what happens.

Choose one tool from the list above. Set aside 20 minutes. Pick one task you’d love to do faster or better. And give it a go.

Your organisation is already using AI. Now’s the time to use it on purpose.

Don’t Start Big — Start Tiny

Your first AI build should be the simplest version of something useful.

Instead of building a full job application platform, start with a tool that scores one question from an application form for your vacacies. Instead of an organisation-wide chatbot, create one that answers basic onboarding questions for new staff.

Start with a working idea. Test it. Share it. Improve it.

The goal is not to get it perfect — it’s to learn what’s possible, and what’s worth doing.

Are You Ready to Build?

If you can answer “yes” to most of the following, you’re in a good place to start:

  • We’ve identified a single process or task we’d like to improve

  • We’ve tested AI manually on that task with good results

  • We have some clean data or content to work with

  • We know which platforms or tools we want to use

  • Someone on our team (or a trusted partner) is ready to build a basic version

If you’re missing a piece, that’s okay. Fill the gaps, not the silence — and take one small step forward.

Final Thought: You Don’t Need to Be a Tech Organisation to Build with AI

You just need to be mission-driven, open to learning, and ready to test ideas quickly. The tools are more accessible than ever. The value lies in starting with real problems and building small, useful solutions.



Want to stay up-to-date with resources, tool guides, and access to others on the same journey. Sign up to our priority list today.

Jonathan Coburn
Previous

Building with AI: A Beginner’s Guide for Charities and Social Enterprises
Next

Making AI Work for Your Mission: Choosing the Right Tools and Tasks