Apple Podcasts ¦ Amazon Music ¦ Spotify
Hi, and welcome back to Try AI for Growth, the podcast where I share the small, surprising and sometimes slightly chaotic stories of how I use AI to make life easier — at home, at work, and inside the organisations I care about. I’m your host, Sara Vicente Barreto, and today I want to take you behind the scenes of something very real: how AI helped me turn a huge, overwhelming field report into structure, clarity, strategy and more importantly, actions.
Now, before you think this is going to be a dry analytics episode, let me reassure you — this is actually a story about decision-making, procrastination, mental load, and how AI can help you take a mountain of information and turn it into something functional.
The Background
A few weeks ago, I came back from a field trip to Mozambique — nine full days of visiting projects, schools, communities, partners, logistics teams, kitchens and classrooms. It was intense, emotional, and full of insights.
As usual, I wrote everything down: pages and pages of notes, observations, photos, conversations, and follow-up actions. And if you’ve ever done fieldwork, you know the real work often starts after you come back — when you sit in front of your laptop thinking, “Where on earth do I begin?”. Whether you are the one doing the site visit and writing the report or if you are just the one consuming it, either way, it is not easy to start.
Additionally, I knew that my team needed clarity and structure to come from me. I can write for hours, but what we needed was to convert the experience into actionable insights.
The Problem
Every time I thought of opening the document, I felt overwhelmed. “I don’t have time for this right now” would be my first reaction.
There were hundreds of details, dozens of projects, and all sorts of follow-ups scattered through different days. Some were trivial, some were strategic, some required decisions — and all of them mattered.
I wanted a consolidated action list, a project analysis, strategic insights, and even templates that could be reused in the future. But the honest truth? It felt like too much for one brain to organise alone. And I had too many other things to do.
So, in classic Try AI for Growth fashion, I did what I often do when I feel completely stuck:
I opened ChatGPT and told it the truth —
“I am drowning in information. Help me organise it and give me a headstart.”
The Experiment
This turned into a long and interactive AI conversation, not always smooth, as I have learnt is the synonym of a better outcome at the end of the day. As a bonus, the process itself taught me as much as the output.
The first thing I did was upload the raw field notes and ask ChatGPT to read through them and let me know when it was ready for questions. This surpasses the optimisation for speed that can come with these models.
Only after I asked it to extract every single action from the entire trip. I had been careful enough to mark all actions as action points when I wrote the report, so I hoped this would be a straightforward task. I listed out which fields I wanted in my output table, and after ChatGPT queried a few of my restrictions, we were ready to go
The Hits and Misses
Funny enough, it was more complicated than I thought, partly due to overthinking and partly due to a Python error. The overthinking came from the fact that ChatGPT read through the entire text to identify more actions than those I marked clearly as such. That had a potential benefit, so I decided to go along with it and see what the outcome was.
The second part was the issue was more complicated. First, the document was deemed too big (which I thought was strange, as it was barely 50 pages), and it could not extract files into Excel. Luckily, when ChatGPT detected the problem, it also provided me with a solution, which was to ask for extracts every 2 days. I thought this was a bit silly, but hey, it solved the problem. So I moved in batches of 2 days, as ChatGPT smoothly extracted the actions from my document in table format, and I only had to copy and paste them into Excel. As I eyeballed them, I could tell they were mostly correct. It came back with over 120 clean, structured actions, written consistently and labelled by topic, project, owner and location.
A Step Further
Once I had my list of actions, it was like a big weight was off my shoulders. I now had something workable, even if 90% ready, that I could go through and assign to the local team, alongside using it for our regular catch-up meetings. It is not unusual that some actions agreed in one supervision trip wait another 6 months until someone remembers them.
Then, I wanted to take it one step further.
In my meetings on the field, we did a deep activity analysis for each project. As we have the same project across different locations, my first step was to bring it all together, while also picking up the details by pillar in order to facilitate updating our 2025 reporting when the time comes. I tested it out for our Feeding Program, and as I iterated and clarified what did and did not belong there, I got a detailed overview of the activities by project, alongside a summary of the challenges and highlights for the project area overall and perspectives and priorities for 2026.
Now, I was no longer extracting information and categorising; I was onto new fields, with a strategic analyst on my side.
Get Ready for the Future
This last part is moving a little bit ahead of my own speed, so instead of doing this for all project areas, I asked ChatGPT to provide me with a prompt that I could use in the future when I wanted to request this analysis again for the remaining project areas.
And because I am always curious for more, I finished off by asking it
“What else could you do for me with this report?”
I got over 10 suggestions. I liked them all, but I do not think they were all feasible on the basis of my single report. I still picked one out for the fun of it. I chose an analysis of the 5 pillars of the NGO’s action across all projects, and how we were doing in each of them. It was a first draft, but not bad at all, and also something I had not done before.
In other suggestions, ChatGPT proposed doing prep for fundraising grants, external communications, a summary of the trip for donors, creating KPIs and a dashboard, analysing the competencies of the field officers and improvement needs, and even preparing the next trip. I am definitely coming back to this chat in the future.
The Unexpected Lessons
As always, I like to extract lessons from these experiences and share those with you. As impressive as ChatGPT’s answers may be sometimes, the magic wasn’t just in the answers. It was in what the process revealed.
- AI can lighten cognitive load – this was something hanging over my head that I knew I had to start, but knowing the size of the task was making it hard to tackle it. In fact, the help of AI did not get it done, but it gave me a head start that I can now work with and, importantly, share with the team and get their help. It is a valuable tool for extracting and labelling information (if you set it up right)
- Iterations get better answers – I liked that ChatGPT asked me a lot of questions, more than it used to. Whilst I sometimes get annoyed thinking – why does it not just give me the table I asked for – the 10 questions on structure, labels and how to capture information made it better
- Don’t be disheartened by bugs – the increased usage probably means that you get slower responses and some bugs. At first, I thought of giving up when I started getting the Excel errors, as I had faced some frustrations like this before. However, because ChatGPT gave me a solution each time it gave me a problem, I decided to give it a go, and we got to a good outcome
Why This Matters
This episode isn’t about Mozambique, or project management, or even NGOs. It’s about something bigger. AI can help you move from overwhelm to clarity, from stuck to structured.
I have spoken about procrastination often in my blog. More often than not, my procrastination happens when I am faced with a big task that I need to break down into smaller bits. This was a clear example of one, and with the help of AI (and 30 minutes of my time), I ended up getting quite a few tools that I can work with.
The AI did not replace me, but it certainly replaced how I will use my time on this task next. Rather than copy-pasting and looking for actions, I will now dedicate my brain power to analysing and directing the next steps.
AI can free your brain to focus on judgment, leadership, and creativity — while it handles the heavy lifting of organisation, formatting, and connecting dots.
And in a world where our to-do lists keep growing and our heads keep filling, that kind of support can be a game-changer.
If there’s something you’ve been postponing because it feels too big — a messy report, a complicated plan, a huge backlog, or even a strategic document you don’t know how to start — try this:
- Put the raw material in front of AI
- Ask for structure, ideally sharing your thoughts on how you want to use it
- Iterate and provide information as you would to someone in your team helping you
- Refine and provide feedback
- Use it as your tool to move forward
It won’t magically solve everything. But it will make the first step feel lighter — and sometimes, that’s all you need.
Thanks for listening to Try AI for Growth. If this episode resonates with you, or if you’ve used AI to untangle something complex, I’d love to hear your story. And if you’re enjoying the show, don’t forget to subscribe and leave a review.
Until next time — keep experimenting, and keep having fun.