Use cases
41 capabilities in applied AI. Designed, built and running in production.
This is what we do, grouped by technical family so you can find the right solution without needing to know the technology behind it. The examples are representative and anonymised to protect client confidentiality.
- 41capabilities catalogued
- 8technical families
- 6sectors covered
The catalogue
Filter by family, sector or technology.
41 of 41 capabilities
01 · Conversational assistants
Citizen chatbot for public services
Web and WhatsApp assistant that lets citizens deal with procedures, grants and regulations without visiting the council or ringing its helpline.
ConversationalSearch
01 · Conversational assistants
Academic chatbot for universities
Assistant with SSO via the institution's LDAP/SAML and RAG over degree programmes, enrolment and academic regulations. Multilingual.
ConversationalSearch
01 · Conversational assistants
Tourism chatbot for smart destinations
Multilingual assistant (ES/EN/CA/FR/DE/IT) with geolocation for points of interest, events, dining and getting around.
ConversationalSearch
01 · Conversational assistants
CRM-integrated corporate chatbot
Assistant with tool calling over your open-source CRM: looks up customers, opens tickets, drafts emails, runs workflows.
ConversationalAgents
01 · Conversational assistants
WhatsApp AI receptionist for clinics
Assistant that manages appointments in Google Calendar, transcribes voice notes and hands over to a person when it detects an emergency. Pay-per-use pricing.
ConversationalAgents
01 · Conversational assistants
Specialist chatbot for legal and insurance
Assistant with RAG over sector regulations, policy terms and internal procedures. Source citations are mandatory.
ConversationalSearchCompliance
02 · Autonomous agents
AI agents for back-office modules
LangGraph agents that run the client's administrative processes, with predictive observability and an AI Act audit log.
AgentsCompliance
02 · Autonomous agents
Conversational BI with text-to-SQL
Agent that queries your operational databases and returns indicators in natural language, with semantic validation and SQL injection protection.
AgentsData
02 · Autonomous agents
Regulatory query agent
RAG over the Official State Gazette (BOE), the Public Sector Contracts Act and sector regulations. Distinguishes current from repealed provisions and cites the relevant articles.
AgentsSearchCompliance
02 · Autonomous agents
QA and automated testing agents
Agents that generate tests, detect regressions and check legacy code against OWASP patterns.
Agents
02 · Autonomous agents
Predictive observability agent
Monitors logs and metrics, predicts incidents before they happen and proposes corrective measures in natural language.
AgentsData
02 · Autonomous agents
Multichannel sales prospecting agent
Outreach sequences across LinkedIn, email and X with AI personalisation, genuine-reply detection and automatic suppression lists.
AgentsConversational
02 · Autonomous agents
Quiz and exam generation agent
Quizzes from text, PDF, URL or YouTube. Multiple choice, open-ended questions, matching, ordering. Anti-cheating controls and automatic marking.
AgentsModels
03 · Search and knowledge
RAG over corporate documentation
Semantic search with role-based access control over your intranet, manuals and procedures. Verifiable source citations.
Search
03 · Search and knowledge
Multi-framework RAG for EU compliance
GDPR + NIS2 + EAA + AI Act platform that analyses your website and documents, maps obligations and generates the mandatory documentation.
SearchCompliance
03 · Search and knowledge
RAG over audio and video transcripts
Whisper transcription + semantic search with timestamps over council plenaries, lectures or video training courses.
SearchModels
03 · Search and knowledge
Semantic search over case law
RAG over rulings and regulations with legal citation checking and comparison of case law across courts.
SearchCompliance
04 · Process automation
CDC and sync across heterogeneous systems
Log-based change data capture across Oracle, SQL Server, PostgreSQL and Mongo with exactly-once delivery. No impact on production.
Data
04 · Process automation
Open-source ITSM and ticketing
Incident platform with LDAP/SAML SSO, 2FA, condition-action automation and a self-service web portal.
Agents
04 · Process automation
Multichannel citizen communication
WhatsApp + SMS with geographic segmentation, AI-assisted campaign generation and auditable tracking.
ConversationalAgents
04 · Process automation
Automated report generation
Pipeline that extracts operational data and produces DOCX/PDF reports with written analysis and historical comparisons.
AgentsData
04 · Process automation
AI-assisted technical proposals
RAG over the history of winning proposals + analysis of the tender specifications. Cuts bid preparation time from 40-80 hours to 5-10.
AgentsSearch
05 · Governance and compliance
AI Act governance platform
AI system register, risk classification, impact assessment and drift and bias monitoring.
Compliance
05 · Governance and compliance
ALTAI trustworthy AI audit
Assessment against the seven EU ALTAI requirements, with a certifiable report for public tenders.
Compliance
05 · Governance and compliance
Automated multi-framework compliance
Technical scanner and document generator (accessibility statement, privacy policy, AI register) for European SMEs.
ComplianceSearch
06 · AI security
AI security audit of live assistants
Scans the AI chatbots on a website for exposed internal instructions, disclosed commercial terms and data leaks, quoting the exact excerpt that proves each finding.
SecurityConversational
06 · AI security
Adversarial testing of LLMs and agents
A suite of tests against the assistant and its connected tools: prompt injection, jailbreaks and goal hijacking, scored by a judge model against a rubric.
SecurityAgents
06 · AI security
Guardrails and output filters
System prompt rewrite, input and output filters, and explicit limits on what the agent can query or run. Regression tests keep the hole closed.
SecurityConversationalCompliance
06 · AI security
RAG and tool hardening
Role-based access control over the sources, sanitisation of the context that reaches the model and least-privilege permissions for every tool connected to the agent.
SecuritySearch
06 · AI security
Continuous assistant monitoring
Regular rescans, with an alert whenever a prompt change or a new source reopens a previously closed leak.
SecurityCompliance
07 · Vertical platforms and synthetic data
Corporate AI platform for insurance
Internal catalogue of models (classification, fraud, OCR) with centralised governance, costs and observability.
ModelsAgentsCompliance
07 · Vertical platforms and synthetic data
AI assistant for the legal sector
Legal vertical with RAG over case law, assisted drafting and human-in-the-loop validation.
SearchConversationalCompliance
07 · Vertical platforms and synthetic data
AI tooling for a business accelerator
Programme management with automatic startup assessment, mentor matching and reports for European funders.
AgentsData
07 · Vertical platforms and synthetic data
Crowd-capacity digital twin
Digital twin for footfall scenarios and mass events. Tests measures before they are rolled out.
DataModels
07 · Vertical platforms and synthetic data
Synthetic data for health research
Synthetic datasets that preserve statistical properties without personal information. K-anonymity and differential privacy.
ModelsDataCompliance
07 · Vertical platforms and synthetic data
Federated data spaces (Gaia-X)
IDS/Gaia-X-compliant sector data space for sharing data between organisations without giving up data sovereignty.
DataCompliance
08 · Custom models and prediction
Fine-tuning open-weight LLMs for your sector
Mistral or Llama tuned on the client's own corpus. Meets EU sovereignty requirements and lowers inference costs.
Models
08 · Custom models and prediction
Forecasting and anomaly detection
Time-series models (Prophet, ARIMA, deep learning) and anomaly detection on operational metrics.
ModelsData
08 · Custom models and prediction
Supervised classification for clinical triage
Symptom triage model for a remote health assistant, trained on synthetic data. Auditable and reproducible.
ModelsCompliance
08 · Custom models and prediction
Custom OCR for specialist documents
Structured field extraction from policies, invoices, prescriptions or case files. Incremental retraining from user feedback.
ModelsData
08 · Custom models and prediction
Model evaluation and benchmarking
Reproducible harness to compare LLMs and classical models on your specific task, with metrics tailored to your case rather than generic ones.
ModelsCompliance
How we organise it
Eight technical families. Pick the one closest to your need.
Each family groups between three and seven capabilities. You do not need to know which one is yours: the assessment identifies it.
- 01
Conversational assistants
Chatbots and voice assistants for web, WhatsApp and phone, with search over your own information, handover to a person and conversation analytics. Customer service, health, training and internal support.
6 capabilities
- 02
Autonomous agents
Agents that run real business processes: they query internal systems, answer business questions about your data, monitor, test and prospect, with a full, auditable trace of every step.
7 capabilities
- 03
Search and knowledge
Semantic search across your intranet, regulatory documents or recordings. Role-based access control, mandatory source citations in every answer and incremental indexing.
4 capabilities
- 04
Process automation
Connecting systems that do not talk to each other, incident management, multichannel communication, automated reports and AI-assisted technical proposals for public tenders.
5 capabilities
- 05
Governance and compliance
Governance of AI systems under the EU AI Act, ALTAI audits and multi-framework compliance with GDPR, NIS2 and the European Accessibility Act. The mandatory documentation is generated automatically.
3 capabilities
- 06
AI security
Audit and hardening of the assistants you already have live: what information they can reveal, what triggers it and how to fix it. Includes adversarial testing, guardrails and continuous monitoring.
5 capabilities
- 07
Vertical platforms and synthetic data
AI platforms for a specific sector (insurance, legal, healthcare, business acceleration), digital simulation and privacy-preserving synthetic data generation.
6 capabilities
- 08
Custom models and prediction
Fine-tuning open models for a specific sector, forecasting and anomaly detection, supervised classification, custom document recognition and model evaluation.
5 capabilities
Does any of this look like what you need?
Thirty minutes to assess whether it is feasible, how long it would take and what it would cost. If it is not a fit for what we do, we will tell you and point you to the right specialist.
Book a free assessment →