Digital transformation in construction: where the builder should start (without it becoming an endless project)

Digital transformation in construction is, for most of the construction company owners I know, a phrase that gives them a knot in the stomach. Not because they're against technology — it's because every time they sit down to think about it, what shows up in front of them is a giant, expensive project with no end date. Replace the entire system, train the whole company, stop the jobsite midway, and pray it works out twelve or eighteen months from now.

I've been teaching artificial intelligence since 2019, before the generative AI boom, and I spent more than 20 years in technology working with Samsung, Itaú, and Hospital Albert Einstein. I've seen a lot of digital transformation succeed up close, and a lot of it turn into a graveyard of software licenses. The difference was almost never the tool. It was where they started. So let me be direct, engineer to engineer: the construction company that treats digital transformation as a "replace everything at once" project almost never finishes it. The one that treats it as a sequence of small wins, starting with the pain that bleeds the most, finishes — and feels results before spending the entire budget.

The mistake of wanting to digitize everything at once

Let me describe a scene that repeats itself. Friday afternoon, the construction company owner receives a proposal from a management system. Twenty-two modules, integration with everything, "end-to-end digitization of your operation." The price is high, but what really scares him is the implementation timeline: six months of setup, plus training, plus data migration from four jobsites running at full speed. He looks at it, imagines the field team reacting, imagines the jobsite stalling mid-transition, and does what any sensible person would do: he puts the proposal in a drawer. "I'll look at this calmly next year."

The problem is that next year arrives, the proposal is still in the drawer, and the margin keeps leaking every day. While the decision to "transform everything" sits idle, the construction company keeps losing money on the same fronts as always: the emergency purchase that comes in overpriced, information trapped in the WhatsApp of someone who already left the company, the expensive engineer spending the day hunting for data instead of running the jobsite.

That's the core mistake. Wanting to digitize everything at once turns digital transformation into an endless IT project — too big to start, too risky to touch. And here's the part almost nobody says out loud: you don't need that. Digitizing the construction company isn't a button you press once; it's a path you walk in stages. The right question isn't "how do I digitize my entire company?" It's "what's the smallest step that already puts money back in my pocket on the next job?"

Start with the pain that bleeds the most

If I could give one piece of advice to anyone starting out, it would be this: don't start with the technology, start with the pain. Forget the module catalog for a moment and ask yourself an honest question — where is your jobsite bleeding the most money today?

For most construction companies, the answer sits in a handful of well-known places. Materials account for 40% to 60% of a project's budget, and a good chunk of that money escapes through last-minute purchases — the "emergency" buy, which averages 28% more expensive than a planned purchase and disappears into the "materials" line item without anyone noticing. It's no coincidence that market research shows 88% of construction delays are linked to material and supply problems, and that waste is the number one problem in the industry for 73% of professionals. Add to that the hours of skilled people spent hunting for information that should be a click away, and the rework from whoever executed the wrong version of the drawing. That's where the margin disappears.

Starting with the pain that bleeds the most has a double advantage. First, you attack the point where the return shows up fastest — usually purchasing and team hours, not the flashiest new feature in the catalog. Second, you build internal proof. When the first small step returns visible money, the whole company stops seeing "digital transformation" as a cost and starts seeing it as an investment. That's how you win over the field: with results, not with talk.

In practice, this means mapping the jobsite into three loss categories and picking one to start with: team capacity (the expensive professional doing robot work), purchasing (emergency buys and lack of systematic negotiation), and rework (wrong version, lost decision, delayed information). Pick whichever hurts most in your operation. That's your starting point — not the twenty-two modules.

Integrate with the ERP, don't replace it

The second thing that makes construction companies postpone digital transformation is a common confusion: thinking digitizing means replacing the system they already have. "I just paid to implement Sienge, I'm not throwing that away." And they're absolutely right not to throw it away. But digitizing isn't about replacing the ERP — it's about adding a layer the ERP doesn't deliver.

I want to be clear about this because it's where a lot of people get lost. The ERP — Sienge, TOTVS, whatever you use — is excellent at what it does: recording what happened. Invoice logged, contract signed, measurement taken, budgeted versus actual. It's your jobsite's system of record, and you shouldn't touch it. What it doesn't do is act on what's happening right now: it doesn't warn you the mortar will run out before the phase starts, doesn't chase the subcontractor who didn't respond, doesn't gather three quotes when a material needs to be purchased. It's passive data — it waits for you to look.

Digital transformation done right, then, isn't about tearing up what works. It's about integrating — placing a layer of active coordination on top of the record, one that cross-references information and acts at the right moment. You keep the ERP as your financial source of truth and gain the coordination that was missing. I wrote about this difference in detail in ERP in construction isn't enough: the difference between recording and coordinating the jobsite — worth reading before signing any contract. And if your concern is the word "AI" on the label, I covered what's real technology and what's just a badge in AI in construction: what actually changes in managing your jobsite (and what's just hype).

Pilot on one jobsite before rolling out to the portfolio

Here's the step that changes everything, and it's the most underrated one: start with one jobsite, not the whole portfolio. The temptation to roll out on every jobsite at once, to "not waste time," is exactly what derails digital transformation. Too many fronts open at once, too many new people learning at the same time, and when something breaks — and it always breaks — it breaks everywhere.

A pilot on one jobsite does the opposite. You pick a representative site, run the new tool there, and treat the first few weeks for what they are: learning. What works, you keep; what causes friction, you adjust before scaling. The risk stays contained to one jobsite, the cost stays small, and you walk away with something worth its weight in gold when it's time to convince the rest of the company: a real case from your own operation, with your own numbers.

And here's a point I refuse to be anything but honest about. There are plenty of vendors selling digital transformation with promises of spectacular cost reduction from day one. Be skeptical. At TrackBuild, for example, we work with a conservative estimation model — on the order of 3.6% to 7.6% of the project value recovered, combining freed-up team capacity, purchasing savings, and avoided rework. On an $8 million project (in reais), that's somewhere between R$291,000 and R$605,000. But I want to say this plainly: that's our model, a bottom-up estimate, not a measured result from a client. Our pilot runs less than a month. The pilot exists exactly to measure this on your jobsite, with your data — not for you to believe a number I made up on a slide. Serious digital transformation starts by measuring your own reality, not by buying someone else's promise.

Measure results from week 1

The last principle closes the loop: measure from the first week. The digital transformation that turns into an endless project is usually the one nobody can evaluate — it runs for months, burns budget, and in the end nobody can say whether it was worth it. The one that succeeds keeps score from the start.

That means defining, before you turn on any system, what you're going to track. How much time does the team spend per week hunting for information? How many purchases went out as emergencies last quarter, and what was the markup? How many hours does the engineer lose reporting status instead of managing? Write down today's number. Then compare it, week by week, with the number afterward. If the tool doesn't move those needles, it isn't transforming anything — it's just collecting pretty data.

The one number we stand behind, measured in our own operation, is this: about 10 hours a week returned to each project manager with our Communication Hub, the jobsite's living memory that answers questions instantly instead of sending someone to go dig. Ten hours is more than a full workday per week, per manager, given back to the jobsite. It's a number you can verify in your week 1, not a projection for a year from now. And that's how digital transformation stops being an act of faith and becomes management: with a visible scoreboard from day one.

A field warning, because it's decisive: none of these gains happen if the team abandons the tool. A system the team works around is money thrown away — and that's the number one reason digital transformation fails on the jobsite. I covered this in depth in Technology in construction: why the team abandons the system (and how to keep it from becoming shelfware). Worth reading before deciding where to start.

Frequently asked questions

1. How long does digital transformation in construction take? It depends entirely on where you start. If you treat it as "replace everything at once," it becomes a twelve-to-eighteen-month project — which is why most never finish. If you start with a specific pain, on a single jobsite, you feel results within the first few weeks and scale what already works. The right path is a sequence of small, measurable wins, not one giant leap.

2. I already pay for Sienge/TOTVS. Will I have to replace my ERP? No. The ERP records what happened, and it does that well — it's your system of record. Digital transformation, here, means adding a layer of active coordination on top: something that warns, chases, and negotiates about what's happening right now. We integrate with your ERP, we don't replace it. You keep what works and gain what was missing.

3. My field team resists technology. Won't this become just another abandoned system? That's the right objection, and the answer determines whether the project succeeds. The tool needs to remove work from the team, not add to it. That's why the agent asks instead of waiting to be filled out, accepts voice replies, reads the invoice automatically, and works from a phone on the jobsite. And in our case, real engineers join the onboarding — because software alone doesn't change jobsite routines.

4. Isn't R$10,000 a month expensive to start digitizing? Look at it through the lens of value, not cost. It's roughly half a mid-level engineer's salary for a manager who never sleeps and chases every front — around 1% of the project to recover the 3.6% to 7.6% that leaks today (our conservative estimate, to be measured in your pilot). A single avoided emergency purchase, with its average 28% markup, already helps pay for the month. The right math isn't the subscription price; it's what the jobsite loses without it.

5. Where do I start, in practice, next week? Pick a representative jobsite and the pain that bleeds the most there — almost always purchasing or team hours spent hunting for information. Write down today's number for that pain. Run a pilot of a few weeks focused only on that, integrating with the ERP you already have, and compare the number afterward. If it moves the needle, scale it. If it doesn't, you lost little and learned a lot. Starting small and measuring is the digital transformation that actually finishes.

How TrackBuild tackles digital transformation paralysis in practice

Everything in this article is method, not a sales pitch — it applies with any serious tool. What TrackBuild does is shorten the distance between the decision and the first measurable result, tackling exactly the two things that make construction companies postpone: fear of the endless project and fear of replacing what already works.

• AI Communication Hub — the jobsite's living memory, which answers any question instantly, with the source, instead of sending the team to go hunt for information. It's the fastest gain to feel, and the source of the one measured number we stand behind — the project manager hours returned per week, detailed above.
• Purchase alert tied to the schedule — cross-references inventory and schedule, catches a material shortage days in advance, and already gathers quotes, so you buy planned instead of buying in a panic. It's where the margin stops leaking first.
• Integration with your ERP — we come in as a layer of active coordination on top of Sienge/TOTVS, with no traumatic migration and without throwing away what you already paid for.
• Engineers in onboarding, with a single-jobsite pilot — because software alone doesn't change jobsite routines. Our technical team joins in, starts small on one jobsite, and measures results from week 1.

And to avoid repeating the usual mistake — putting everyone in and getting penalized for it — pricing is per module, with unlimited users: construction company, developer, and subcontractors on the same platform, with no per-seat toll, integrating with the ERP instead of replacing it.

Start with the math: get the Leak Diagnosis

Before choosing a tool, do the math on your jobsite. The Leak Diagnosis is a free 30-minute call where we calculate, with your data, how much your jobsite loses per month in emergency purchases, rework, and hours spent hunting for information. You walk away with the number for the pain that bleeds the most — which is exactly where your digital transformation should start. No commitment, no twenty-two-module catalog: the math first, the decision after.

Talk to us

---

Rafael Medeiros is co-founder of technology at TrackBuild. PMP certified, with more than 20 years in technology, he has been teaching artificial intelligence since 2019 — before the generative AI boom — and has developed projects with Samsung, Itaú, and Hospital Albert Einstein. He is a source on AI for CNN Brasil and BBC News Brasil. TrackBuild brings together AI agents and specialized engineers into the jobsite's routine, with unlimited users, so information stops getting stuck and the job stays on schedule.