Most of the AI software you see for the trades has the same problem early automobiles had.

The first “cars” weren’t really cars. They were carriages with engines bolted on; the proverbial horseless carriage.

Problem was, the whole design still assumed a horse: the shape, the layout, the way you “drove” the thing. It took time (and a few broken axles) before car makers stopped copying the old model and started building something that actually fit the new reality.

That’s where a lot of HVAC, plumbing, and electrical software is today.

The tools aren’t useless. But they often feel like they were designed by someone who has never sat next to constantly ringing phones, or a dispatcher trying to patch a schedule back together, or a GM looking at five branches’ worth of half-complete job notes and trying to sort out which numbers are real.

How bad is this problem? Well, Autodesk and FMI estimated that “bad data” may have cost the global construction industry $1.85 trillion (yes that’s “trillion”) in 2020. And their definition of bad data is stuff we all recognize: data that’s inaccurate, incomplete, inaccessible, inconsistent, or untimely. When your systems don’t match how your work actually happens, bad data isn’t an edge case, it’s the default.

Now apply that situation to the trades. When your office is slammed, your schedule is shifting, and your team is doing triage all day, “perfect data entry” will never be the priority. If your software requires perfect behavior to produce usable data, it will fail—because it’s built for an ideal world, not your reality.

Why trades software feels outdated (even when it’s “modern”)

A lot of off-the-shelf software fails trades businesses because it’s built like a product demo, not like an operation.

Most vendors start with a generic model of how businesses “should” run. By contrast, trades businesses run on how work actually flows.

That difference shows up everywhere:

In the trades, speed beats elegance.

Your customers don’t care how sleek your dashboard looks. They’re comparing how fast you answered the phone, how quickly you scheduled, and whether you showed up when you said you would.

In the trades, the work is distributed.

Your work is split up over the field, the office, the warehouse, the customer, and sometimes a subcontractor. That means information has to move cleanly—without forcing everyone into a new routine.

In the trades, exceptions are normal.

The schedule changes. Parts are missing. A tech calls in sick. A job that should’ve been one hour becomes three. You need a system that can flex without breaking or creating ten new steps.

And yet a lot of software is still built on a take-it-or-leave-it model. “Here’s the system. Your team needs to adapt to it.”

That’s a horseless carriage. It’s taking the old operational reality and just bolting on features—without accounting for the way trades really work.

How do you know you’re living in a “horseless carriage” stack?

See if any of this sounds familiar:

Your real process lives outside the system.

The CRM says one thing, the spreadsheet says another, and your dispatch lead has the truth in their head. If someone goes on vacation, the wheels wobble.

Your team spends more time moving information than using it.

They’re bogged down copying notes, rewriting customer context and hunting down job history. Or they’re following up because the system didn’t. 

Buying more tools just makes things more complex.

The new tool requires yet another login; another set of notifications—yet another feature you pay for, but your team avoids because it doesn’t fit the day-to-day flow.

Everything works until you try to scale.

The “good enough” patchwork holds at $5M–$10M. Then you add locations, volume, lead sources, techs, and acquisitions—and suddenly your stack can’t keep up.

Owners feel the pain as “software that’s annoying.”

Aggregators feel it as “software that prevents scaling.”

When you buy a portfolio company, you inherit their stack, their habits, and their workarounds. Multiply that across five, ten, twenty companies and you get a mess that’s expensive in two ways:

1) It slows down what you’re trying to standardize.

If every branch has a different way of handling missed calls, cancellations, job closeouts, and reviews, you can’t reliably measure performance, because the underlying process isn’t consistent.

2) It makes adoption harder, not easier.

If your tech strategy becomes “everyone needs to use the same software the same way,” you’ll burn political capital fast. Trades teams don’t resist tech because they hate change. They resist tech because they’ve been burned by tools that add friction.

The future isn’t “one platform to rule them all.” Instead, it’s repeatable workflows that across different companies, embedded into the tools teams already use.

That’s an actual car, not a horseless carriage.

So what does “building the car” look like for you? 

Cars evolved away from carriages because designers rethought the entire experience: steering, suspension, ergonomics, safety, speed. Cars were designed around new capabilities, not old habits.

In trades tech, “the car” means systems built around today’s demands:

Speed-to-lead is table stakes. 

If you can’t follow up fast and consistently, you’re buying leads just to donate them to the competitor who responds quicker.

Customer communication has to be consistent.

Not perfect, but consistent. The customer experience is the sum of a hundred small moments: confirmations, updates, reminders, follow-ups, review responses.

Your data has to be usable without perfect behavior.

Your teams are busy. They don’t have time to sit down and “do admin.” The system has to capture information that matters as part of the flow, preferably by automation.

Your workflows need to bend to your business.

The best trades businesses are good at operations. Tech should amplify that, not force everyone into some generic operating system.

This is where AI and automation actually do what you need them to do—when they’re installed like tools in a toolbox, not sold like buzzwords.

At The Graphite Lab, that’s our entire approach.

We inject automation and AI directly into the software stack you already run—so you get “extra hands” without a disruptive changeover.

Tools. Runs. Assemblies. Simple building blocks.

If you’ve heard us talk about tools and assemblies, it’s not because we’re trying to rename automation.

It’s because it’s the cleanest way to build tech that your team can actually live with.

A tool does one job well. It takes an input and produces an output.

A run is when that tool executes its given task.

An assembly stacks tools together into a workflow that solves a real operational problem from beginning to end.

It’s a flexible system that lets you build the car one system at a time—without ripping out the engine while driving.

What the “car” does differently in real life

It closes revenue leaks automatically.

Missed-call follow-up. Estimate follow-up. Canceled-job win-back. These are operational realities that stop money from disappearing between cracks.

It reduces admin load on your staff.

Instead of your best people doing repetitive copy/paste work all day, automations handle the predictable steps. That frees up your team to focus on tasks that require human judgment: customer conversations, schedule decisions, escalations.

It makes outcomes more consistent across branches.

For aggregators and multi-location operators, this is huge. You can standardize what matters (speed-to-lead, review handling, closeout discipline) without forcing every location to change their procedures.

It improves the customer experience in ways that customers notice.

Faster response. Clearer updates. Fewer “wait, what’s going on?” moments. The kind of polish that earns trust, referrals, and reviews.

And the best part is: it doesn’t require your team to “become tech people.”

If the solution depends on your entire staff changing their behavior, it won’t stick.

But if it acts like a quiet, reliable crew member in the background, it will.

So here’s a quick reality check: are you buying features or fixing workflows?

Owners and operators get sold features all the time:

“AI summaries.”

“Automated messages.”

“Smart dispatch.”

“New dashboard.”

Here’s the test that cuts through the hype:

Can you describe the workflow in one sentence? For example: “When a missed call happens, the customer gets a fast, human-sounding follow-up and replies get routed to the right place.”

That’s a workflow you can actually measure.  If the vendor can’t explain how it fits into the day-to-day flow, you’re probably looking at another horseless carriage feature.

Where The Graphite Lab fits into all this

We’re not adding bells and whistles to clunky tools. We’re building the car, by:

• embedding AI and automation into your existing software stack (ServiceTitan, Jobber, Housecall Pro, and more)

• delivering modular tools and assemblies that match how your trades business actually works

• designing around adoption and speed, not “more platforms”

We’ve delivered 1,400+ automations for trades teams, with measurable wins like:

• less time spent on calls

• fewer cancellations

• major reductions in payroll and admin task time

• productivity lifts in vendor bill handling

• improved online reputation through faster review responses

It’s not magic, it’s just AI that’s designed for your reality.

If you want to see whether your operation is dealing with a horseless carriage problem—or just a few fixable workflow gaps—start with the bottlenecks you feel every week: missed calls, follow-up speed, cancellations, closeouts, reviews, vendor bills, payroll admin. 

And let’s see what the best way is to fix them.