Communication on the job site: what information trapped in WhatsApp costs your construction company

Communication on the job site at your construction company today lives almost entirely inside WhatsApp — and that's exactly why it costs money you can't see on any invoice. It's 4 PM on a Tuesday. The designer sent a revised formwork drawing, changing a column's elevation. He sent it to the "Jardins Site - Engineering" group. But the structural contractor, who's pouring concrete on Thursday, is in the "Jardins Site - Structure" group. Nobody forwarded it. On Thursday, the column goes up at the old elevation. And nobody is going to call this "a communication failure." It'll become "rework," "an unforeseen issue," "that column problem" — and the cost disappears into the middle of the job, exactly the way an emergency purchase disappears into the middle of the "materials" line.

If you run a mid-size construction company, this scene isn't hypothetical. It's Tuesday. And it repeats at different scales every week, because the communication system you're using to coordinate the job wasn't built to coordinate a job. It was built to exchange messages.

I spent my career inside industrial pulp and paper projects — developments in the billions of reais — coordinating dozens of crews and companies at once. And I learned one thing that holds just as true for a Suzano plant as for your R$30 million building: the job doesn't fall behind because it lacks competent people. It falls behind because the right information doesn't reach the right person at the right time — it stays trapped on someone's phone, in the wrong group, in the head of someone who already left. In this article I want to show why WhatsApp-based communication breaks down structurally, what it really costs, and what to do to bring everyone in without paying a toll for every single person.

200 messages a day across 15 groups: the root of the chaos

Let's start with the simplest number: volume.

A mid-size construction company running three or four jobs doesn't have "a WhatsApp group." It has fifteen. There's the general group for each job, the engineering-only group, the administrative group with finance, the group with the electrical contractor, the group with the plumbing contractor, the direct chat with the foreman, the direct chat with the mortar supplier. Add it all up and you comfortably have 200 messages a day flowing past you — a photo of a crack, a three-minute voice note from the foreman, "can you confirm the measurement?", a meme, a good-night message, and somewhere in the middle of it, the one message that actually mattered that day.

Now look at what this format does to information:

• It has no owner and no address. The revised drawing was "sent to the group." Which group? Received by whom? Read by whom needed it? Nobody knows. Responsibility dissolves into "I sent it" versus "it never reached me."
• It doesn't stay put. The screen scrolls, and it's gone. The detail agreed on Monday is buried under 600 messages by Friday. Nobody scrolls back two weeks of chat to find what was decided.
• It lives in silos. What's discussed in the structural group, the installation crew never sees. Each crew has its own WhatsApp, and information that needs to cross from one to another simply doesn't cross.

WhatsApp is a brilliant tool for what it was built for: chatting. The problem isn't WhatsApp — it's that you're asking it to do the job of coordinating a construction site, and coordinating a site requires information to have an address, a memory, and traceability. Three things a message group, by nature, doesn't have. This is the same point I made in Project management: why spreadsheets and WhatsApp groups no longer control your construction company: the tool isn't bad, it just wasn't built for the task you pushed onto it.

Trapped information vs. information that flows

Here's the distinction that changes everything, and I want you to hold onto it: there's information that flows and information that stays trapped. The entire cost difference between a job that runs smoothly and one that bleeds money lives in that word.

Trapped information is the design revision sitting in the inbox of an engineer who's on vacation. It's the decision to switch finishes that was agreed on in a voice note nobody can find anymore. It's the price the supplier quoted last week that only the buyer remembers now — roughly, from memory. The information exists, it didn't vanish. It's just locked somewhere the person who needs it can't reach when they need it.

And the cost of this isn't abstract. Industry surveys point to waste as the number-one problem in construction today — something around 73% of companies recognize it as their biggest pain. A large part of that waste isn't cement thrown away: it's people standing idle waiting on an answer that's trapped, it's material bought wrong because the correct spec never circulated, it's a stage redone because the valid version of the drawing never arrived. Waste, on the modern job site, is largely a communication problem wearing a different costume.

Information that flows is the opposite: the design revision automatically reaches whoever's executing that scope, and it's marked as received. The decision about the finish is logged with date, author, and reason — and anyone can ask and get the answer with its source. The supplier's price is recorded, not held in anyone's memory. The reason for a delay is documented on the day it happened.

The enemy, notice, is never your team. Your engineer isn't careless, your foreman isn't disorganized. The enemy is trapped information — and the system that keeps it trapped is simply the way the job communicates today. Switching communication systems isn't switching apps. It's moving information from "trapped" to "flowing."

The contractor nobody told

Let me tell you the scene that, in my experience, costs the most — and that is communication, pure and simple.

The average job isn't built by your construction company alone. It has four, five, six companies inside it at once: the developer who contracted the work, your construction company, the structural contractor, the electrical contractor, the window/door contractor, the waterproofing subcontractor. Each with its own team, its own WhatsApp, its own way of receiving (or not receiving) information.

Now think about how many times a job's delay started like this: a stage's date changed, or a drawing's elevation changed, or a material got delayed — and the information never crossed from your construction company to the contractor who depended on it. The window/door contractor showed up with the crew on the agreed date and the masonry wasn't ready, because the sequence change stayed in the internal engineering group and never reached him. You paid his crew to sit idle, or paid for remobilization, or swallowed the friction. And at the root of it was just this: a message that never crossed a boundary between companies.

The industry feels this hard. Estimates suggest the vast majority of job delays — something around 88% of projects — run into problems tied to materials and to the sequencing of supplies and crews. And what jams that sequence is almost never the material itself: it's the information about the material, the date, the sequence, that didn't reach whoever needed to act in time. Coordinating the companies on a job is, at bottom, a communication problem between organizations that don't share the same WhatsApp group — and shouldn't.

Everyone in, with no per-user toll

And this is where most attempts to solve the problem trip up — including attempts with software.

You hire a system to organize communication. Great. Then you discover it charges per user. And what happens in practice? You put your internal team in, because you're paying for that. But the electrical contractor, the window/door contractor, the subcontractor — those you leave out "to save on licenses." Result: your construction company's information looks great inside the system, and communication with whoever is executing half the job stays on WhatsApp. You paid to organize half the problem and left the other half exactly where it was. The pricing model sabotaged the outcome.

This isn't a pricing detail. It's the difference between solving and not solving. Job-site communication only flows when everyone is inside the same place: the construction company, the developer, every contractor, every subcontractor. If any important link is left out because it costs one more license, information goes right back to being trapped exactly at the boundary between companies — which is precisely where it costs the most, as in the case of the contractor nobody told.

That's why our pricing is per module, with unlimited users. You bring the entire job in without thinking about seats — because bringing everyone in isn't a luxury, it's the condition for communication to stop leaking. Charging per user, on a construction site, is charging a toll precisely for doing the right thing.

The daily standup that documents communication on its own

There's one last point, and it's what separates recording from coordinating.

You can organize all of a job's communication in one place, with everyone inside, and still depend on people logging what happened. And we already know how that ends: at the end of the day, under pressure, logging is the first thing to go. The engineer spends the whole day putting out fires and doesn't stop to document that a crew stalled because of rain. The information, again, stays only in his head — trapped.

The fix isn't demanding more discipline from people who are already overloaded. It's making communication document itself as a byproduct of the routine that already exists. In practice: instead of a form nobody fills out, an agent that asks — that checks the status of each crew during the daily standup, accepts the answer by voice or by message, and structures what was said on its own. The engineer doesn't get one more administrative task. He answers what he'd answer anyway, and the record is ready: who said what, when, which blocker, which mitigation.

That's the point where communication stops being just message exchange and becomes coordination. And that's where the one number I stand fully behind comes from: the job's intelligent memory, which answers any question with its source instead of sending everyone hunting through a PDF, gives back around 10 hours a week to each project manager — more than a full working day, every week, of qualified people going back to running the job instead of digging for information across fifteen groups. That's the one we measure. The rest of the savings — the rework avoided, the contractor who doesn't sit idle — goes into our estimation model, which points to recovery in the range of 3.6% to 7.6% of the project's value adding up all the leaks. And I want to be upfront: that larger number is a conservative, bottom-up model, not a result measured at a client — our pilot is less than a month old. What I stand behind fully is the 10 hours; the rest is a defensible estimate, but still an estimate.

Frequently asked questions

1. WhatsApp is free and everyone already uses it. Why switch to something paid? It's not free — it just doesn't come with an invoice. You pay for WhatsApp in the revision that never crossed and became rework, in the contractor's crew standing idle because the date change stayed in the wrong group, in your engineer's hours spent hunting for information among 200 messages. That bill arrives every month, diluted into "unforeseen." The question isn't WhatsApp versus paid system; it's trapped communication versus flowing communication — and trapped costs more, it just doesn't show up itemized.

2. I already have Sienge/TOTVS. Doesn't the ERP solve job-site communication? No, because they're different layers. The ERP records what happened — invoices, measurements, balances — and it does that well. But it doesn't make information flow between the people on the job at the right time: it doesn't chase the daily standup, doesn't notify the contractor of a change, doesn't answer "what's the latest revision of the drawing?" It's a good archive, not a coordinator. That's why we integrate with the ERP instead of replacing it — the active communication layer sits on top of what you already have.

3. My field team isn't going to drop WhatsApp to fill out another system. That's the right objection, and the honest answer is: if it depends on filling out a form, they really won't — and they're right not to trust anyone who promises otherwise. That's why the model is an agent that asks instead of waiting for input, that accepts answers by voice and by message, that works on the job-site phone. The team doesn't get new work; they get a question at the right time and stop being chased for status all day. And that's why we bring in engineers during onboarding: field adoption isn't luck, it's design.

4. Won't bringing contractors and subcontractors in blow up the per-user cost? Not in our model — because pricing is per module, with unlimited users. You bring the construction company, the developer, and every contractor in without paying per seat. That's by design: communication only flows with everyone inside, and no system that punishes you for including people is going to solve the problem it was hired to solve.

5. Doesn't this just become another tool for watching over the team? The goal is the opposite. The record documents what your team already said and protects whoever raised the flag: when a crew stops because of rain or because a supplier's material was late, that gets logged the same day — and it's your defense when someone asks why the job fell behind. Organized communication isn't a camera pointed at anyone; it's proof that the information was given, so the delay doesn't unfairly land on whoever raised the flag in time.

How TrackBuild attacks trapped communication on the job site in practice

Everything this article described — the revision that never crossed, the contractor nobody told, the decision trapped in a voice note, the engineer hunting for information across fifteen groups — is exactly what TrackBuild was built to attack. Not with one more message group, but by moving information from "trapped" to "flowing":

• AI Communication Hub — the job's living memory. Instead of scrolling through PDFs and chats to find what was decided, you ask and get the answer with its source. This is where the hours given back to each project manager every week come from — the one gain we measure and stand fully behind.
• AI Project Manager — checks the status of each crew during the daily standup, accepts answers by voice, and structures what was said on its own, logging blockers and mitigation. Communication documents itself as a byproduct of the routine, not as extra work at night.
• Engineers in onboarding — software alone doesn't change a job site's routine. Rollout comes with method and engineering people designing alongside your team which conversations enter the system and how, so the tool doesn't become one more abandoned app.

And because pricing is per module, with unlimited users, you bring the construction company, the developer, and every contractor into the same place — with no toll per seat, which is the only way job-site communication actually flows. If you want to go deeper into how information stops getting lost, it's worth reading Job-site document management: where's the latest version of the drawing (and why nobody can find it) and, on why an agent that acts solves what a passive dashboard doesn't, AI in project management: what AI agents do that dashboards don't.

Find out what trapped communication costs on your job site

Disorganized communication is one of the job site's silent leaks — the kind that never shows up under its own name in the report, but adds up to more than any overrun you can actually see. The rework from the revision that never crossed, the crew standing idle waiting for an answer, your engineer's hours spent digging for information: each one disappears into "unforeseen," and that's exactly where the margin lives.

If you want your job's actual number before deciding anything, the Leak Diagnosis is a free 30-minute call where we calculate, with your own data, how much your construction company loses per month to trapped information and miscommunication — no commitment.

Talk to us

---

Felipe Barreto is co-founder of engineering at TrackBuild. A civil engineer with a master's degree in civil engineering from UFMG, an MBA from IESE Business School, and PMP certification, he managed the construction of plants for Suzano, Klabin, and other pulp and paper industry leaders — projects in the billions of reais. TrackBuild puts AI agents and specialized engineers into the job site's routine, with unlimited users, so information stops getting trapped and margin stops leaking.