The catalog that sells, the engine that charges, the queue that produces, and the dashboard that decides. All in one system.
Built on
Technologies in production today. Not certification badges.
None of these problems get fixed with more people or more hours. They get fixed once, and stay fixed.
The business runs on an Excel file that breaks every time someone touches it, nobody knows which version is the right one, and if that person goes on vacation, the operation goes with them.
Price, stock, sizes, order status. The same five questions every day, tying up someone who should be doing something else.
You commit to a delivery without knowing the workshop's real workload, and then someone has to call and apologize. The cost isn't the apology — it's the customer who doesn't come back.
Someone reconciles payments by hand against the bank statement, writes down who paid, and crosses their fingers that nobody's impersonating you on another channel.
Sales are strong, but at month's end the number doesn't match the feeling. Without a per-unit cost, margin is just a hunch.
The catalog on one side, accounting on another, shipping in a separate spreadsheet. Every handoff between them is a person copying and pasting.
If you recognized two or more, there's something to sort out.
Request a diagnosticA prism doesn't create light: it splits it and puts it in order. That's what we do with a company's operation — the catalog that sells, the engine that charges, the queue that produces, and the dashboard that decides, each piece in its place and all of them talking to each other.
Before writing a line, we look at how you work today and where the hours go. The first piece is chosen for its return, not for taste.
There's no closed product to force-fit. Each piece is designed for the business that will use it, and plugs into the one before it.
Dates calculated from the real queue, margins per product, attribution per campaign. Every number you see here carries its source alongside it.
Your data in your account, your domain, your code. Infrastructure is billed by actual usage, never a fixed number guessed in advance.
We don't sell a closed product with a feature list. We look at how you work, choose the piece with the highest return, and build that first. Below are examples of the kind of system that comes out of that, grouped by area.
Showing 12 examples of All.
Answers price, stock, sizes, and order status around the clock, and hands off to a person when the conversation calls for it.
Replaces the Sunday WhatsApp group. The server assigns by order of arrival, not a person interpreting messages.
Meta and Google campaigns managed from one panel, with every sale attributed to the campaign that brought it in.
Your storefront with prices, real stock, and filters. It's the base everything else plugs into.
Deposits and full payment with automatic confirmation. No more checking transfers one by one.
The phone becomes the register. Scan, record the sale, and in-store and online stock stay in the same number.
The customer picks carrier and branch at checkout, and the shipping order is generated automatically.
Every sale recorded with its per-product cost, so you know what makes money and what just moves.
The work queue with committed dates and an automatic calculation of how far out you can safely commit.
How much material is left and how many units you can still produce, with an alert before the shortage, not after.
One code per order, a public page where the customer sees where it stands, and a verifier against impersonation.
Who can see what, who changed what, and when. Each role sees exactly what it needs to work.
If what you need isn't on this list, it's because the list is an example, not a catalog. Let's talk.
Five stages built on Lean principles. Each one ends with something in your hands, and each one targets a specific waste: time, waiting, rework. Tap a stage to open it.
Use the arrow keys to move through the stages and Enter to open one.
Stage 1 of 5: Understand what hurts, and what it costs
Before proposing anything, we need to know what doing it by hand costs today. Value is defined by your operation, not by our catalog.
A written diagnostic with the estimated cost of staying the same, and its assumptions in plain view.
Overproduction: building features nobody asked for, because nobody asked what actually hurt.
One to two weeks
Move the controls to match your operation's numbers. It's an estimate and it's presented as one: the assumptions are visible, below the result.
For operations in Chile.
Estimated monthly savings updated: CLP 902,026 per month.
Reference estimate built from the assumptions above. It doesn't replace a proposal: it's here to ground the conversation in your own numbers.
Review these numbers with youPrisma 137 is an automation consultancy based in Santiago. Here's how we work.
Whoever listens in the first meeting is the same person who later writes the code. No middle layers translating what you said.
Based in Santiago, Chile. We know Transbank, Khipu, Starken, Chilexpress, the boleta and the SII, because they're part of the day-to-day of the operations we work with.
Before a single line of code: scope, phases, what's in and what's out. No surprises halfway through.
The work doesn't end on delivery day — that's when the system actually meets reality. We stay on for that.
If your question isn't here, write to us and we'll answer it straight.
Yes, and it's actually where the change shows the most. In a small company, two people dedicated to repetitive work are a huge share of the team. Size isn't what determines whether it's worth it: how many hours a month go into tasks a machine can do is.
It depends on the piece, which is why we work in phases. The first delivery is measured in weeks, not months: we pick the problem that hurts most and solve it completely before moving on. A real timeline comes from the diagnosis, not a generic table, and it's put in writing before we start.
Yes. On delivery day the system only just meets reality, and adjustments always come up. Support and maintenance are agreed in writing before starting, with clear scope: what's covered, on what timeline, and what's billed separately.
It's the first thing we look at. If your ERP, your accounting, or your payment gateway works, it gets connected instead of replaced. Replacing what already works makes the project more expensive and slower without giving you anything new.
Yes, and it's the most common case. An Excel file that survived the business's growth is a good specification: it already has the rules, the states, and the calculations. The work is moving it to a database where several people can work at once without stepping on each other, and where nothing breaks when a formula gets dragged.
It depends on scope, and we don't publish prices because a number without knowing your operation doesn't help anyone. Implementation is billed once per phase, and infrastructure by what it actually consumes each month, not a fixed number guessed in advance. The calculator above is here to ground the conversation before talking about money.
Still have a question?
Ask usThe criteria we apply on every project, written down. Write to us and we'll send you the full document.
A fixed monthly fee hides the risk of real usage. When the catalog grows, someone quietly pays the difference. This is the model that avoids that.
Billing by the actual clock punishes using your tools well. Here's the criterion for pricing a piece by what it's worth, not by how long it took to type.
An entire operation is a year-long project nobody can deliver all at once. Starting with what hurts most is more honest, and it also works better.
Our logo is a beam that passes through a prism and comes out in three colors. Here you build it yourself: turn the mirrors so the light reaches all three targets.
Level 1 of 3
Turn the pieces to guide the beam from the left to the targets. Tap a piece to rotate it.
With a keyboard: arrow keys to move around the board, Enter or spacebar to rotate a piece.
The more you tell us here, the more concrete our first response will be. Only name, email, and need are required.