How One Person Replaced a $40k/Year Tool With Something They Built in a Day

Marta ran operations for a 30-person logistics company in Guadalajara. Her team used an enterprise project management platform that cost them around $40,000 a year — per-seat licensing, premium tier for the reporting features, plus a consultant they’d hired to set up the initial workflows two years ago.

Here’s the thing nobody on the team wanted to say out loud: they used maybe 15% of it.

Drivers logged their daily routes in the tool. Dispatchers checked a Kanban board to see who was available. Marta pulled a weekly report showing on-time deliveries and open issues. That was it. The Gantt charts, resource leveling, sprint planning, time tracking, and portfolio dashboards? Nobody touched them. They were paying for a Swiss Army knife to open envelopes.

The Moment It Clicked

The contract renewal came up in February. Marta had been telling herself she’d migrate to something cheaper “next quarter” for over a year. But this time, a friend who ran a small e-commerce brand told her something that stuck: “I stopped looking for the right tool and just described what I needed to an AI. It built it.”

Marta wasn’t a developer. She’d taken a Python course once and abandoned it after week three. But she understood her own workflows better than anyone. She could describe exactly what her team needed because she lived inside those processes every day.

She decided to try building a replacement herself before signing the renewal.

What She Actually Built

Marta sat down on a Saturday morning with an AI app builder and started describing what she needed, one piece at a time.

A route logging screen. Drivers needed to check in at the start of their shift, see their assigned stops, and mark each one as completed. No drag-and-drop Gantt nonsense — just a list with checkboxes and a timestamp. She described this in plain Spanish and watched the app builder generate a working interface with a database behind it.

A dispatcher dashboard. A single page showing which drivers were active, how many stops they had left, and a color-coded indicator for whether they were on pace. Marta described the logic: green if they’d completed at least 60% of stops by noon, yellow if between 40-60%, red below that. The builder translated this into conditional formatting and a live-updating view.

A weekly report. The numbers Marta actually pulled every Friday: total deliveries, on-time percentage, issues flagged by drivers (like a bad address or a customer not home), and a comparison to the previous week. She asked the builder to generate a summary table and a simple bar chart. It did.

A simple issue tracker. When a driver flagged something — wrong address, damaged package, customer complaint — it needed to go somewhere Marta could see it and assign it. Not a full ticketing system. Just a list with a status (open / in progress / resolved) and the ability to add a note.

The whole thing took about eight hours spread across Saturday. Not because any single piece was hard, but because Marta kept refining. The first version of the dispatcher dashboard showed too much data. She trimmed it. The route logging screen needed a “skip stop” option she hadn’t thought of initially. She added it by describing the change.

What $40,000 Was Actually Buying

When Marta compared her four-screen app to the enterprise platform, the gap was obvious — but not in the direction she expected.

The enterprise tool had hundreds of features and required a consultant to configure. Marta’s app had four screens that mapped directly to how her team already worked. No training needed. No configuration debt.

But the real cost of the enterprise tool was never the subscription. It was the friction her team worked around every day. Dispatchers coordinated over WhatsApp because the platform’s mobile app took four taps to update a delivery status. Marta maintained a separate Google Sheet for the weekly report because the built-in reporting module required navigating three menus to pull the same five numbers. The drivers had bookmarked a workaround page in the help docs for the route logging screen because the default flow assumed project phases they didn’t use.

Marta’s app didn’t have workarounds because it was built from the workarounds. Every screen existed because someone on the team had been doing that task informally — on WhatsApp, on a spreadsheet, on a whiteboard — and Marta just described the informal version to the builder.

The Parts That Surprised Her

Three things Marta didn’t expect:

Iteration was fast. When a driver suggested adding a notes field to each stop, Marta described the change to the builder during her lunch break and deployed it that afternoon. With the enterprise tool, changes like this went through a support ticket queue and sometimes took weeks.

Her team adopted it immediately. No training sessions. No “please watch this onboarding video.” The dispatchers opened it Monday morning and understood it because it looked like the whiteboard they’d been using informally, just digital.

She kept improving it. Over the next two weeks, she added a fifth screen: a monthly view for her boss showing delivery trends and cost-per-route estimates. With the enterprise tool, this would have been a custom report request. With her own app, it was a 20-minute conversation with the builder.

What This Isn’t

This isn’t a story about enterprise software being bad. If you’re a 500-person company running complex cross-functional projects with dependencies, resource constraints, and compliance requirements, you probably need that Swiss Army knife.

But if you’re a 30-person team using 15% of a tool that costs more than one of your employees, something is misaligned. The tool isn’t the problem — the mismatch is.

And that mismatch used to be unavoidable. Before you could build an app without coding, your choices were: pay for the big tool, cobble something together in spreadsheets, or hire a developer to build custom software (which introduces its own costs and timeline). Now there’s a fourth option: describe what you need and build it yourself.

The Math

Marta’s numbers after one month:

• Enterprise tool: ~$3,300/month ($40k/year)
• AI app builder subscription: under $100/month
• Time to build: 8 hours (one Saturday)
• Time to iterate: 20-30 minutes per change
• Team adoption time: zero — they got it on day one

She didn’t need a CTO or an engineering team. She needed to describe how her team actually works — and an AI app builder that turned that description into working software.

What to Think About

If you recognize your own situation in Marta’s story, here’s a useful exercise before your next software renewal: write down every feature you actually use in your current tool. Not the features you think you should use or plan to use someday. The ones your team touches every week.

If that list fits on a sticky note, you might be overpaying for complexity you don’t need.

You don’t have to build the replacement in a day. You could start with just one screen — the one that matters most — and see how it feels. The cost of trying is a few hours. The cost of not trying is another year of paying for features you’ll never touch.

If you want to see what building your own tool looks like, try Proyecta and start with the thing your team complains about most.