🗺️ Sidequest #2: The Sync Loop of Doom 🗺️
Until you know the terrain, you don't know what belongs in your inventory.
You're rolling out a tech-based data collection system in a social program—tablets in the field, cloud sync, real-time dashboards. It even has a chat bot that can guide you through unexpected challenges. You've vetted the tool. You've mapped the data. Some of the team were reluctant about this new technology but you managed to get everyone on board and everyone has been trained.
Now we can watch the data just flow in seamlessly. Right?
You start to get emails. Some data are missing. Other data aren't quite right. Data collection staff are frustrated because they entered the data correctly, they swear. But now its awry.
You're getting passive-aggressive emails with phrases like “doing our best” and “per our last conversation.”
You decide to go out with the data collection team for a couple of days on a quest to some of the sites to find the problem. It must be something they're doing right? Data from a couple of the staff never seem to have problems so perhaps it's just that more training is needed.
Village1: The first round of surveys go fine. Survey data is relayed through the tablet from the respondent's mouths into the cloud.
Village 2: This site is more remote. You start to see what some of the staff are talking about. Syncs are failing mid-survey. The devices are trying to sync over weak mobile networks, dropping packets mid-transfer. Data is getting corrupted and lost.
Hmm. But what about the data collectors whose data are coming in just fine?
Village 3: You accompany one of the numerators whose data seem to always be in order. And there it is.
You realize the he has resorted to printing surveys and entering the data manually back at the office. Then you realize…
What you bought: A digital system where you enter data once, sync in real time, skip the clipboard, and get AI-assisted help via chat. The data are seamlessly stored in the database and a monitoring dashboard shows in real time your progress against indicators.
What you got: a $15,000 paper-based workflow. Layered over a non-functioning digital system. Plus missing data.
Unless the tablets were manually flipped into airplane mode, they’d try to push data over unstable 3G. That unstable connection scrambled everything—like trying to fax a document mid-earthquake.
What went wrong?
You forgot to check the infrastructure.
Your Quest
The problem is clear. Now you're in the middle of it.
Do you patch the workflow? Replace the tools? Retrain the team? Or rethink the design entirely?
Your quest is to figure out how to keep the system usable, reliable, and trusted—with the infrastructure you actually have, not the one you hoped for.
You must choose:
📖 Double down on training and enforce airplane mode?
🪓 Redesign the tool to sync on demand and fail gracefully offline?
🗃️ Switch to a fully offline workflow with batch uploads at the end of the day?
🧑💻 Bring in a local developer to build a lighter, context-appropriate version?
📜
Admit it’s not working—yet—and revert to paper until the system can catch up?
This isn’t a pop quiz. It’s field reality. And every day you wait, the dashboard keeps lying.
Takeaway:
Infrastructure isn’t a backdrop. It’s the whole map.
Your tools will only perform as well as the terrain allows. Before you roll out the next great tech solution, test it under the actual conditions it’s meant to survive—low signal, dying batteries, distracted staff, and broken workflows.
Build for the map you’ve got, not the one in your pitch deck.
Flip the switch—just make sure you know what it’s connected to.
Bonus XP:
Look up which undersea fiber optic cable connects your servers to your field site. (Seriously—there’s a map.)
Then ask yourself: how fragile is that connection?
AI might live in the cloud, but the cloud runs on wires. And some of those wires are one backhoe or earthquake away from breaking your whole system.
Next in this series: How AI Breaks in Social Impact: Infrastructure Gaps
Anthralytic helps mission-driven teams make smarter decisions by combining human insight with AI-powered tools—safely. We build strategy, evaluation, and accountability systems that work—without the jargon. Learn more at anthralytic.ai.
