Real work, and what it did

What we've built

Processes that now run on their own

Working every day, with the hours measured.

Automation · Collections
960hours/month eliminated

Collections from 6 banks, with nobody in front of the screen

From 12 people doing repetitive work to a process that runs on its own

The problem

The company processed collections from six banks by hand. Two people per bank, four hours a day. The work was repetitive and mistakes got made. And it didn't scale. Every new bank meant hiring more people.

How we did it

We started with the highest-volume bank to prove the model worked. It was running in three weeks. The other five came online over the following month.

What we did

We built one process per bank. It goes into the portal, identifies itself, finds the day's payments and downloads them. Then it matches everything against the company's internal records and sets aside whatever doesn't line up for a person to look at. The rest runs on its own every day, without anyone starting it.

Results

  • 960 hours a month of manual work, eliminated
  • Fewer errors when matching payments
  • Cash visible the same day instead of the next
  • Every run recorded and auditable

How it is kept running

The ops team gets it bank by bank: whether it ran, how many payments went through, and any errors. When one bank changed its portal, we caught it the same day and fixed it before anyone noticed.

These processes predate Mesh, so they were never wired into it. Everything we put to work now reports there from its first run.

Automation · Orders
240hours/month recovered

Order statuses in a supplier's portal

Four people, three hours a day, replaced by a process that runs on its own

The problem

Updating order statuses in a supplier's portal took four people, three hours every day. Log in, find each order, check its status, enter it into their own system, move to the next one. Hundreds of orders, every day.

How we did it

We put it together in a week from a recording of how they did it by hand. Then we tuned it with the operations team for the odd cases — partial shipments, cancelled orders, portal maintenance — and it was in use three weeks later.

What we did

We built something that walks the portal the way a person would, and then loads the updates into the internal system. It sorts out the long result pages and the sessions that drop halfway through on its own.

Results

  • 240 hours a month recovered
  • Full traceability: every update logged with date, time and source
  • Statuses update in minutes, not hours
  • Not a single update missed since it started

How it is kept running

It runs on its own and leaves a record of every execution. The team gets a daily summary with the orders processed and the ones that need a person to look at them. If the portal goes down, it tells both of us at the same time.

These processes predate Mesh, so they were never wired into it. Everything we put to work now reports there from its first run.

Automation · Data
80+hours/month saved

Data that was moved between systems by hand

Dozens of hours a week of copy-paste, gone

The problem

A team was losing dozens of hours a week moving data by hand between two systems that didn't talk to each other. What came out of one platform had to be copied, reformatted and re-entered into the other. Always with the risk of a typo or a field left stale.

How we did it

We mapped every field with the team, built it with full validation and error handling, and ran it alongside the manual work for two weeks before switching over. The team only looked at exceptions.

What we did

We built a process that reads the data from the source system, converts it to the format the other one expects, checks it against the business rules, and loads it. Nobody has to step in unless something unusual comes up.

Results

  • 80+ hours a month of manual data entry, eliminated
  • Cleaner data: the checks catch errors people used to miss
  • Nobody has to supervise it day to day
  • Every run measured: records processed, errors caught, duration

How it is kept running

Every load leaves a record of how much it read, how much it wrote and what failed validation. The weekly report goes out on its own. One time the source system changed a field without warning and we caught it the same day.

These processes predate Mesh, so they were never wired into it. Everything we put to work now reports there from its first run.

Platforms and custom systems

Complete systems we built for clients. See projects →

Government · Blockchain
5mofrom idea to running

A digital wallet for a provincial government

An award the state used to hand out on paper, turned into digital money pegged to the dollar so inflation can't eat it. Built and launched in five months.

View the full case study

The challenge

The Government of the Province of San Luis, Argentina had run a program since 2010 that gave a cash award to every student who finished high school. It worked: provincial graduation rates climbed roughly 20% and moved above the national average.

What we did

We built Billetera Activa, an app where every student receives their award as digital money backed one-to-one by the U.S. dollar. Because it's pegged to the dollar, it doesn't matter how far the peso falls. It's worth the same the day they get it and six months later.

Outcomes at launch

  • 100%Of recipients created and verified an account
  • 84%Of what was issued was actually used
  • 25%Chose to hold rather than cash out right away
  • 100%Security and control targets, met
Large enterprise · Beverages
2,500+distributors buying

The portal 2,500 distributors buy through, rebuilt

For a major beverage distributor: the system that takes orders from 2,500+ distributors across several countries, rebuilt so it stops going down.

View the full case study

The requirement

More than 2,500 distributors across several countries place their orders there. Any outage or delay translated straight into lost sales. New features weren't what was needed. Not going down was.

The Volkom approach

We worked in two stages. The first stabilized and extended the platform the client had built in-house, so orders came in reliably from day one.

Outcomes

  • Customer experience
  • Reliability
  • Business continuity
  • Room to grow

Other things we've built

No page of their own, but all of them shipped. If one of these looks like your situation, we'll walk you through it on a call.

Health · Computer vision
secondsper image, instead of minutes

Reading dental X-rays automatically

Reviewing an X-ray by eye takes minutes an image. We built a system that reads it and produces the diagnostic report on its own, in seconds.

Government · Infrastructure
millionsof citizens served

A province's app for its citizens

An app a whole province uses can't go down. We built the infrastructure, the automated releases and the round-the-clock monitoring that keep it up.

Operations · ERP
3 monthsfrom spreadsheets to a working ERP

A 200-person company that ran on spreadsheets

There was no single place to look up the right number. We implemented Odoo with the modules their industry needed, and then stayed on to keep it running.