Artifical Intelligence AI.
We embed AI models into your business processes and applications, and build the infrastructure AI agents need to operate across your systems. From analysis to first deployment in 6-12 weeks.
Strategy and development from the same team
Most AI projects stall for the same reason: the people advising aren't the ones building. We do both.
✓ We plug into your team as part of the project, not just as a vendor
✓ Full end-to-end process: from analysis to deployment and evolution
✓ The AI logic lives inside your software, with control over how it behaves and scales
✓ If AI isn't the best tool for your case, we tell you in the analysis
✓ We start with a contained pilot before committing to a big investment
✓ Ongoing support and evolution of the solution
How we approach an AI project
Three phases, from "where do I start" to a working, evolving product. Within each one, we build whatever your case needs.
Phase 1 - Analysis and Design
We map your processes, identify where AI adds real value, and hand you a proposal with scope, technology, budget and timelines. This is the phase where, if AI isn't the answer to your problem, we tell you before you spend on it. You usually come out with two or three prioritised use cases.
Phase 2 - Build
We put AI to work inside your systems. Depending on the case, a project combines one or more of these pieces:
- Integration into your software: we connect AI to your ERP, CRM or internal applications to automate specific tasks, with the logic inside your software rather than in an external tool.
- AI-ready software: the APIs, structured data and protocols (such as MCP) that AI needs to operate reliably across your systems.
- Custom AI applications: built-in assistants, copilots or agents with human validation, as a web or mobile app, with the whole process end-to-end.
Phase 3 - Evolution
A model in production doesn't stand still: your data, your operation and what AI can do all change. We measure how it performs, adjust, and stay with you so the solution improves instead of going stale.
Technologies we use.
Our clients.
Companies that trust us to build and evolve their digital products.
We found the perfect partner in Liquid: flexible, reliable and committed to the project.
Guillermo GaspartFounder of BYHOURS
Thanks to Liquid we are more efficient and have gained independence by having our own software. For us it has been a life-changing experience, especially because we avoid quoting errors that could cost us clients.
Marta CaparrósCEO of YouTOO
Frequently asked questions.
n8n and Make connect apps and move data between them; they work well for simple automations. We build the AI logic inside your software, with control over how it behaves, how it scales and how it evolves. In the intro call we tell you what fits your case best, even if that's one of those tools.
Not always, and that's part of the job. Many systems were designed for people (interfaces, forms, manual checks) and AI needs something else: APIs, structured data, protocols like MCP. When that layer doesn't exist, we build it. It's actually a growing part of AI projects.
It can be set up so it doesn't. Depending on the provider and the type of account, data sent through the API can be kept out of their model training; and depending on the case, we can work with anonymised data or with models deployed in your own environment. Privacy and compliance (GDPR included) are defined in the analysis, before moving a single piece of data.
It's a real risk, and it's designed around rather than ignored. Depending on how critical the process is, we keep a human in the loop before an answer reaches a client, limit the AI to answering only from your data, or combine both. That's why we start with a contained pilot: to see how it behaves before giving it more responsibility.
It depends on the scope:
- Analysis and design: 2-4 weeks
- First working version of an automation or integration: 6-12 weeks
- Custom AI application: 3-8 months
We work in phases to validate the result before continuing with the next one.
The cost depends on the scope, complexity and degree of integration with your systems. In general, the price can range between €2,500 and over €50,000:
- Analysis and design: €2,500 to €6,000 (2-4 weeks)
- AI automation or integration: €8,000 to €25,000 (1-3 months)
- Custom AI application: From €25,000 (+3 months)
At Liquid, we offer a personalized quote after an initial consultation.
