We're turning WhatsApp + Voice notes into structured daily site reports — built for how construction actually runs in emerging markets.
What's your site communication setup look like?
Most construction AI tools are built assuming you have a PMO team, Procore in the budget, and BIM-enabled workflows. That's not the reality for the majority of $2M–$50M projects in East and West Africa — or honestly, a lot of markets globally.
Here's what those sites actually look like:
47 WhatsApp groups per project
Voice notes in Swahili, English, or a mix of both that nobody transcribes
Verbal approvals that vanish into scroll
PMs spending 90 minutes a day on the phone just to reconstruct what happened
The data exists — it's just invisible to the developers and lenders paying for the project.
We've been building an operational intelligence layer that sits on top of that existing WhatsApp workflow. Site team sends voice notes and photos like they already do. A multilingual ASR + structuring layer transcribes, classifies, and extracts activity, issues, and approvals. A human reviewer checks anything uncertain. Developer gets a structured PDF report in their inbox by 6 PM.
No new app. No training. No changing how the site team works.
We're currently running paid pilots in Q2 2026 across East and West Africa — 30-day pilots on one project, we cover the QA cost.
For those of you working in markets where informal comms (WhatsApp, voice notes, SMS) are the de facto site management tool — what's your biggest pain point with extracting usable data from that? Is it the transcription, the classification, the lack of audit trail, or something else entirely? Curious what solutions (if any) you've tried. Welcome any feedback and will engage as much as possible