Put AI to workinside your organisation
From readiness to deployment, we bring AI into your day-to-day work: tool selection, custom solutions and MCP builds, all EU AI Act-aware.
Four pillars of AI integration
Comprehensive AI solutions designed for European SMEs, from quick wins to enterprise transformation.
Generative AI integration
Your gateway to AI. Deploy ChatGPT Enterprise, Claude Teams, or Microsoft Copilot with custom configurations for your business.
- Tool selection and licensing optimisation
- Custom GPT agents trained on your data
- CRM and document system integration
Process automation
Eliminate repetitive tasks with intelligent automation. We implement RPA solutions that learn and adapt.
- Process discovery and mining
- RPA bots with AI document processing
- Power Automate, Zapier and Make integration
Data analytics and BI
Turn your data into actionable insights. AI-powered analytics with natural-language querying capabilities.
- Data strategy and governance
- Power BI implementation (from €10/user/mo)
- Custom dashboards and natural-language querying
Custom AI solutions
Bespoke AI applications for your unique challenges. From proof of concept to production-grade systems.
- Document intelligence and contract analysis
- Custom chatbots with RAG architecture
- Lead scoring and demand forecasting
We don't just integrate AI. We train your team to run it.
Most AI help stops at building something for you. We go further and put AI in your team's hands, in the daily work of EU public affairs: monitoring files, drafting positions, mapping stakeholders, and connecting your own data to your assistant.
What we put in your team's hands
Monitor and horizon-scan
Turn eight live monitors and the EU corpus into a morning read your team trusts, in the languages they work in.
Draft positions and responses
Position papers, consultation responses and amendments, drafted from real sources rather than a model's memory.
Map stakeholders and institutions
Who decides, which committee, which rapporteur, which coalition. Mapped, not guessed.
Connect your data to your assistant
Wire your own files and our live EU data into your assistant over MCP, so it answers from the real thing.
Be found by AI
Make sure your positions are cited when an assistant discusses your issue, not absent or misquoted.
Choose tools, stay compliant
Pick the right AI tools for the team and use them safely, EU AI Act-aware, across six languages.
How we work with your team
Pick the format that fits, or combine them. Every one is built around your real dossiers, not a generic curriculum.
Team workshop
A focused half-day on your real files. Your team leaves using AI the same week.
Tailored programme
A multi-session track that builds the workflow around your own dossiers.
Advisory retainer
We stay on call as the tools and the rules keep changing.
Conference talk
A plain-language talk for your event on AI in EU public affairs.
Tell us your team and your files, and we will shape the session around them.
Make your organisation callable by every AI
The Model Context Protocol is the open standard that lets ChatGPT, Claude, Gemini and Mistral call an external service directly. We turn your API, database or register into an MCP server, connect it to every major assistant, and meter it if you want to charge per call.
The third layer of AI discoverability
Being cited is the floor, not the ceiling. Each rung puts you deeper inside the assistant your audience already uses.
Mentioned
Generative engine optimisation gets the models to cite you inside the answers your audience already reads.
Usable
Agent-readiness lets an AI agent crawl your site, understand it and complete a real task on it.
Callable
An MCP server puts your data and your actions inside the assistant itself, as a tool it can invoke. No website in the loop, no scraping, no guesswork.
What we build
From wrapping a single endpoint to a metered, OAuth-secured server listed in the connector galleries.
API to MCP server
We wrap your existing API, database or document set in a JSON-RPC MCP server, and design the tool surface so a model actually picks the right tool for the question.
Working server and tool schemaAuth and hosting
Bearer headers, key-in-URL for the hosts that offer no header field, or a full OAuth 2.1 layer with dynamic client registration when a host demands it. Deployed on your stack or ours.
Deployed and securedConnect every assistant
Claude, ChatGPT, Gemini, Mistral, Cursor and VS Code each accept credentials differently. We connect and test all of them, then write the guide your users follow.
Host-by-host connect guideMetering and monetisation
Per-call billing, key issuance and revocation, usage dashboards. Your data becomes a revenue line instead of a support cost.
Keys, balance and billingEvery host connects differently
This is why most MCP servers work in one assistant and quietly fail in the next. We have already hit every one of these traps on live servers, so your users never meet them.
Two servers we run in production
Our EU Data API exposed as MCP: the most comprehensive and best curated database of EU affairs, fetching data from more than 10,000 sources from the europa.eu digital ecosystem. More than 900 endpoints and an MCP that helps LLMs answer without guesses or hallucinations.
Metered per callOur eight open policy monitors, turned into tools your assistant can use. Ask it what changed this week in EU defence spending or the tariff war and it reads the live monitor instead of guessing.
Free to connectWe do not demo this. We run it.
Connect the Beresol Monitors server to your own assistant in about a minute, and judge the work before you buy any of it.
Five phases, delivered in Scrum sprints
An Agile, five-phase framework that balances thoroughness with SME budgets. Each phase runs in two-week Scrum sprints with reviews, retrospectives, and a backlog you co-own. Tangible value at every iteration.
Discovery and assessment
2-4 weeks
Stakeholder interviews, AI readiness evaluation across 7 dimensions, gap analysis, and a prioritised use-case portfolio.
Strategy and roadmap
2-4 weeks
ROI and complexity scoring, technology architecture recommendations, vendor selection, and a 12-24 month AI roadmap.
Proof of concept
4-8 weeks
Build a minimal viable solution and pilot it with defined success metrics. 56% of SME buyers trust demos most.
Implementation and integration
8-24 weeks
Full-scale development in two-week Scrum sprints, system integration, security implementation, UAT, and comprehensive go-live support.
Training and optimisation
Ongoing
Role-based training, prompt engineering workshops, and adoption monitoring. 38% of AI failures stem from insufficient training.
Turn compliance into competitive advantage
The EU AI Act creates significant obligations for SMEs using AI in employment, healthcare, credit scoring, and more. We help you navigate compliance whilst building trust.
Next deadline: labelling AI content
Wednesday 2 December 2026
Machine-readable marking of AI-generated content, plus the new ban on nudifier and CSAM-generating systems. High-risk duties follow: 2 December 2027 for standalone systems, 2 August 2028 for AI built into regulated products.
SME and small mid-cap provisions
Priority sandbox access, proportionate fees and simplified documentation, extended to small mid-caps by the 2026 AI Omnibus
Risk-based approach
Most AI systems face minimal requirements; we identify your obligations
ISO/IEC 42001:2023 path
40-50% overlap with the AI Act; certification builds stakeholder trust
Assessment and classification
AI system inventory, risk classification, gap analysis, compliance roadmap
Documentation and governance
Technical documentation (Annex IV), QMS development, risk management, EU database registration
Full implementation
Conformity assessment prep, CE marking, AI literacy training, complete compliance package
Ongoing compliance
Quarterly audits, regulatory updates, incident response, documentation maintenance
Penalties for non-compliance
Up to €35M or 7% of global turnover for prohibited AI; €15M or 3% for other violations
The compliance-first AI partner
Unlike generic tech providers, we combine implementation expertise with deep EU regulatory knowledge.
Forward-deployed, not fly-by
Most AI projects arrive as a slide deck and leave as one. Our engineers deploy into your team, in your office, until the tool runs and your people own it. It is the one part of AI adoption that AI cannot do for you.
Regulatory expertise
AI Act, GDPR, sector regulations: we navigate compliance so you do not have to.
SME focus
Right-sized solutions. No enterprise overhead. Practical outcomes for businesses like yours.
EU funding access
Navigate Horizon Europe, GenAI4EU, and EIB programmes to co-finance your AI journey.
Ready to transform your business with AI?
Join the 51% of SME leaders taking the first step. Let's map your AI opportunities together.
Frequently asked questions
Who does EU AI Act-aware AI integration?
Beresol is a Brussels-based software company founded by Victor Solé Ferioli. Its AI integration practice delivers EU AI Act-aware implementation for European organisations and EU institutions, covering AI readiness assessment, tool selection, retrieval-augmented generation (RAG), agent architectures, and custom AI solutions.
What is RAG (retrieval-augmented generation)?
Retrieval-augmented generation (RAG) is an approach that connects an AI model to your own documents and data so its answers are grounded in your sources rather than generic training data. Beresol builds RAG systems as part of its AI integration work, alongside agent architectures and custom AI solutions.
Do you train our team, or only build AI for us?
Both. Beyond building AI for you, Beresol trains and equips your public affairs team to run it in their daily work: monitoring files, drafting positions and consultation responses, mapping stakeholders, and connecting your own data to your assistant. Formats run from a half-day team workshop to a tailored programme, an advisory retainer, or a conference talk.
Can Beresol help an organisation adopt AI?
Yes. Beresol runs the full path from AI readiness assessment and tool selection through to deployment, building EU AI Act-aware, custom AI solutions for European organisations and EU institutions, and training your team to run them. The work spans RAG, agent architectures, and hands-on enablement for public affairs teams.
What is a forward-deployed AI engineer?
A forward-deployed AI engineer embeds with your team to set up an AI tool, gather feedback and fix what breaks, rather than handing over a report and leaving. Beresol delivers AI integration this way for European organisations and EU institutions, pairing hands-on implementation with EU AI Act compliance.
Three ways Beresol works with AI
Each page is dedicated to one job: be found by AI, put AI to work, or feed your AI live EU data. Together they are one system.
AI Integration
Bring AI inside your organisation: readiness, tool selection, custom solutions, MCP builds and team training.
Explore our other services
Our services connect. Public affairs intelligence feeds the GovTech we build, and AI integration adds AI discoverability on top.