AI for organizations · Custom agent
A custom AI agent. Built for your process.
When none of our modules fits your process, we build an agent from the ground up. It runs on the same AI Engine as our other agents: it uses your company’s knowledge, rules and connections to your systems. We start with a feasibility study on your data.
We’ll check your agent idea. The study itself costs PLN 8–12k net, and its cost counts toward the implementation.
- 2 weeks
- feasibility study on your data
- Your data
- an anonymized sample, read-only access or your own hardware
- Decision point
- after the prototype, with the target and threshold agreed before signing
- AI Engine
- the same foundation as our modules
When it fits
You know what you want to build. You don’t know if your data can carry it.
You know which process you want to hand over to an agent. None of our modules covers it, and nobody has checked whether the data allows it.
Before deciding, the board wants to know the scope, price and stages of the implementation, calculated on your process.
After Discovery or a process transformation program, you have a chosen idea and want to test it before you build.
Your data cannot leave the company. The idea has to be tested on your hardware or on an anonymized sample.
Every implementation stands on data. That is why we start by checking the data.
First, we check whether one of our modules is enough. See the modules
The same foundation
Whatever the agent, what matters is that it runs on a good engine.
A custom agent runs on the same AI Engine as our modules and uses the same knowledge, procedures, decision rules and system connections of your company. What we prepare for it stays in AI Engine and shortens the path for the next agent.
See how AI Engine worksHow we build
From an idea to an agent at work. With a decision after every stage.
Call
Free. We check your agent idea.
Feasibility study
Two weeks on your data. It ends with an implementation quote with milestones.
Prototype with a decision point
We agree the target and threshold before signing. After the prototype, we check the result against them.
Implementation and maintenance
The agent in everyday work, with support set out in the contract.
Feasibility study
Feasibility study. Know whether the idea works before you commit to the build.
In two weeks, we test on your data whether the agent you want to build makes sense. You get a check of the assumptions, an implementation quote with milestones and a baseline we later use to measure the return. The study is the entry point to every implementation we deliver.
How it works
Two weeks on your data. A quote and a decision at the end.
01Kickoff
Access to data
We agree on what we work with: an anonymized sample or read-only access. If the data cannot leave the company, we work on your hardware.
02Data
Checking the assumptions
We test on your data whether the idea’s assumptions hold. We start with the data because every production agent depends on it.
03Measurement
Baseline
We measure how the process works today. After implementation, we use this point to calculate how many hours and how much money the solution gives back.
04End of the study
Quote with milestones
You get the scope, the implementation price and the stages at which you accept the work. You decide based on numbers from your own process.
Deliverables
Three deliverables to base your decision on.
Assumption check
Whether the idea’s assumptions hold on your data. If the data isn’t ready, we tell you what needs to happen first.
Implementation quote with milestones
Scope, price and the stages at which you accept the work. We give the final price of an agent implementation only after the study.
Baseline
How the process works today, measured before kickoff. We measure the return against it after implementation: the target we aim for and the threshold below which we consider the solution has not worked.
The feasibility study is the entry point to every implementation we deliver: agents, applications for departments and custom-built projects.
After the study
After two weeks, you decide based on numbers.
The study ends with one of three decisions. You make it, based on results from your own data.
Build
Implementation
An agent with the quote and milestones from the study. We write down the target and the threshold before signing. There is a decision point after the prototype.
Data first
Getting the data in order
If the data isn’t ready, we start there: with one product data hub for AI or with a data foundation. The agent goes live only on an organized source.
Data foundationsNot now
No implementation
If the process changes every month, there are only a few dozen cases a month or the documentation has no owner, we tell you plainly that the implementation won’t pay back. You pay only for the study.
Proof
What we have custom-built.
In progress
Financial statements and reports without manual retyping
80-90%*
about 3 min instead of 15-30 min of retyping
In progress
AI that keeps data inside the company
100%*
of production processing on the company’s servers (requirement)
Work in progress. These are results verified at the test stage or the targets we are aiming for. We will confirm them on completion.
That is also how our Quoting agent came about: it started as a custom project.
Explore case studiesScope
Scope and investment.
What we do
- Test the idea’s assumptions on your data
- Measure the process baseline before implementation
- Quote the implementation with milestones
- Work on an anonymized sample, with read-only access or on your hardware
What we don’t do
- We don’t implement during the study. It ends with a quote and a decision
- We don’t look for use cases across the whole company. That is what Discovery is for
- We don’t clean up data during the study. If needed, we quote it as a separate step
Investment
PLN 8–12k net, credited toward implementation
The study costs PLN 8–12k net. Its cost counts toward the implementation when you deliver it with us. The first call, where we check your agent idea, is free.
Book a callQuestions
Frequently asked questions.
How is a custom agent different from a ready-made module?
A module has a defined scope and workflow: quoting, documents, company knowledge, product selection or tenders. We design a custom agent for your process when no module covers it. Both work on the same AI Engine, with the same knowledge, rules and system connections.
How long does the feasibility study take?
Two weeks. At the end, you get an assumption check on your data, an implementation quote with milestones and a baseline to measure the return.
Do we have to hand over our data?
Your data doesn’t have to leave the company. An anonymized sample or read-only access is enough. If needed, we work on your hardware.
How is the study different from Discovery?
Discovery takes 3–5 weeks and answers the question of where AI pays back the most in your company: you get a list of use cases ranked by return and readiness. The feasibility study starts where you already know what you want to build: it tests a specific idea on your data and quotes the implementation. If you know exactly what you want, we go straight to the study.
How much does the study cost?
The study costs PLN 8–12k net, and its cost counts toward the implementation. The first call, where we check your agent idea, is free. We quote the implementation itself after the study, broken down into milestones.
What if the study shows it isn’t worth it?
We tell you plainly. An implementation doesn’t pay back when the process changes every month, there are only a few dozen cases a month or the documentation has no owner. In that case, you pay only for the study, and the build budget stays with you.
Let’s talk about your agent idea.
The first call is free. We’ll check your idea and tell you whether a ready-made module is enough or whether it makes sense to start with a feasibility study.
Book a callOr write to: hello@wearefuture.ai
