Custom software development, process consulting and technology outsourcing
Four ways to work with us. They are contracted separately, closed at a fixed price per phase, and they all start the same way: understanding how your company operates before touching a line of code.
The process is redesigned before it is automated
A process carrying inherited steps does not improve by being digitised: it just gets faster at doing what is unnecessary. First we walk the circuit with the people who run it, measure where the data breaks and remove what adds nothing.
It is the step that sets us most apart and the one that costs us most time: instead of arriving with a closed proposal, we put on the table how companies like yours have solved it and what your sector is doing. You decide with two points of view instead of one.
Four project types, one way of building
It is built by DSA Codex, our software factory, on standard, documented technology. Pick a type: you will see how we intervene, what is delivered and how long it takes.
The code is 100 % yours, from the first commit. We start with a one-hour diagnostic at no cost: we tell you what to build, what not to, and in what order.
We build in the technology your team already knows:
Standard and documented, nothing proprietary to tie you in. If you change supplier tomorrow, any competent team picks the work up without rewriting the system.
Already have an RFP, an RFQ or a defined scope? You do not need the diagnostic: ask us for the range and we will give it to you in the first email.
An application that fits your operation, not the other way round
Four layers, each with its owner and its test. B2B platforms, B2C portals, client extranets and a mobile field app. Code and environments are in the client’s name from day one.
Data stops travelling as an email attachment
A central exchange point with error handling, retries and a log of every transfer. No system needs to know the others, and the current ERP stays.
We automate the procedure, never the decision
The path gets shorter and the judgement points become explicit, with a record of who approves what. The exception stays human.
Replaced module by module, with the service running
The old system is never switched off in one go. We document what nobody documented, isolate it behind an integration layer and replace the pieces one at a time, with rollback available at every phase.
AI agents inside our guardrails, not instead of our engineers
We are not AI opportunists. We are software engineers with more than twenty years of craft who now multiply their output with agents. The difference is what sits between the agent and production: a control circuit nobody skips.
The agent writes. Automated checks and a senior engineer decide what reaches production. What the client gains is not cheaper code: it is the same quality standard, delivered sooner.
More capacity, same standard
Assisted generation of code, tests and documentation inside the control circuit. No commit reaches production without passing the guardrails and human review. It is what lets us commit to timelines we would not have accepted three years ago.
Only where it actually solves something
Conversational assistants that handle queries out of hours, automatic classification and prioritisation of incidents, data extraction from documents and predictive models on your own history. All of this is already running on our platforms; it is not a brochure promise.
A team with names, not a bucket of hours
A fixed technical lead, stable profiles and real time-zone overlap between Spain and Latin America. Development teams, systems engineers, ERP consultants, a service desk and external technical leadership as Interim CTO.
between Madrid and the team in Latin America
Why we do not publish rates
Because a profile is not a rate: it depends on the technology, the seniority, the years of experience and the responsibilities. What we can tell you is what actually deserves comparing.
Your IT department, without the headcount
Hardware and software, from the desk to the network. Ticketing with a signed SLA, remote support for every site and an on-site engineer when the machine has to be touched. With our own lab in Madrid: we repair and prepare here, we do not ship it out.
on critical incidents, with a monthly report
When the process is common to a whole sector, it is already solved
Two products of our own, in production with clients and maintained by us. They are deployed and configured, not built from scratch, so the timeline and the cost are in a different league.
DSA Core
Operational platform for field, construction, logistics and services companies: work orders, control tower, subcontractors, warehouse, transport, assets and finance.
See DSA Core →
DSA Compliance
Regulatory compliance in Spain and Latin America: whistleblowing channel with statutory deadlines tracked, criminal risk map, data protection, time tracking and reporting to the board.
See DSA Compliance →
When we do NOT recommend building custom software
If an off-the-shelf product covers the process, we say so and help you deploy it. Building what already exists is the most expensive way to waste time.
How we invoice, said before the first meeting
Nobody should find out how a project is paid for in the third meeting. This is the full path: each phase is contracted separately at a fixed price, signed before starting. The amount comes out of the assessment, and until then we do not quote blind.
Free diagnostic
One hour to understand your operation. We come out with a recommended order of action.
Assessment
Two to four weeks. Process map, functional specification, roadmap and investment range per phase.
Fully credited back if the project goes ahead.
First phase in production
Eight to twelve weeks, with a verifiable deliverable every two. Scope and amount come from the assessment and are signed before any code is written.
No open line items, no hours to justify.
Additional module
On a platform already built. The architecture and data model are already paid for: every new piece costs a fraction of the first.
Outsourcing and dedicated teams are contracted on a monthly fee fixed by quarter. We do not publish a rate because it depends on the profile, but the honest comparison is not against another consultancy: it is against what that person costs you in-house.
See the comparison →
Three ways to contract. None of them is billed by the hour.
Fixed-price project
Scope, timeline and price defined before starting. The overrun risk is ours.
Dedicated team
Stable capacity by quarter, with a technical lead and a delivery commitment each cycle.
Support and evolution
Corrective and evolutionary maintenance with an agreed improvement allowance and a service level agreement.
Two answers and we tell you the exact sequence
It is the same logic we apply in the first meeting. There is no right service without knowing the starting point.
Standard, documented technology. Nothing proprietary to tie you in.
If tomorrow you decide to change supplier, any competent team can carry on without rewriting the system. And we integrate with whatever ERP you already have.
What people ask before contracting
Questions about timelines, code ownership and geographic coverage are answered in the general questions. These are the contracting ones.
Ask a question →How much does custom software development cost?
We do not publish rates because a project is not a price list: it depends on the process, the systems you already have and the business logic. What we do publish is how it is contracted. Each phase goes separately at a fixed price, signed before any code is written, with no open line items and no hours to justify. The assessment is credited back in full if the project goes ahead, and the prior valuation session costs nothing. Ask us for the range and we give it to you in the first email.
Can I contract only the assessment?
Yes, and it is the most common route. It is delivered complete — process map, functional specification, roadmap and investment range per phase — and it is yours to execute with whoever you want, including another supplier. If we do the project, it is credited back in full.
Do you use artificial intelligence to write code?
Yes, with agents working inside our guardrails for quality, security, readability and maintainability. No code reaches production without passing the automated checks and review by a senior engineer. AI increases output capacity; it does not replace engineering judgement.
Is it cheaper than hiring in-house?
Generally yes, because the cost of an in-house engineer is not their salary: it is social security contributions, recruitment, equipment, training and the months it takes to fill the vacancy, plus the severance risk. With a dedicated team there is a monthly fee fixed by quarter, no employment liability, and if a profile does not fit it is replaced with no transition cost.
What happens if the scope changes mid-project?
Estimation overruns are ours to absorb: that is the point of a fixed price. What does get re-quoted is a scope change requested in writing, and it is done before executing it, never in the final invoice. Since each phase is contracted separately, the change usually lands in the next phase without renegotiating the whole contract.
What warranty is there once the system is in production?
Every development comes with a warranty period in which we fix, at no cost, any defect against what was specified. After that you can contract support and evolution, with a response within 4 business hours for critical incidents, or take it away: the exit is a documented handover in 30 days and it costs nothing.
Bring the process that works worst. One hour is enough to know whether there is a solution.
It is led by the founder together with the engineering lead who would run the project. Free, no sales deck, and with a written conclusion. And before proposing anything, we show you what we have done in cases like yours.
Write to us through whichever channel you prefer. We reply within 24 working hours.
Request diagnostic Message us on WhatsApp