Chatbot trained on your docs
Web, WhatsApp or intranet. Answers from your data, never makes things up, links the source and escalates to a human when needed.
SERVICE · AI AGENTS
Four production-ready products: a chatbot trained on your documentation, a 24/7 phone assistant, a sales copilot that replies to leads in under 5 minutes, and semantic search. Integrated with your CRM, website and voice channels.
Service · What's included
Web, WhatsApp or intranet. Answers from your data, never makes things up, links the source and escalates to a human when needed.
Handles calls, qualifies, books and resolves across any time zone. Natural voice, transcription and handoff to your team.
Responds to leads in under 5 minutes, qualifies them and pushes them into your CRM with full context. Your reps arrive warm.
Catalogue, knowledge base or tickets. Search by intent, not exact keywords. Results that cite their source.
Service · Impact
Service · Process
We map your operations, volume and stack. You leave with a prioritised plan and a locked scope for the POC.
We build the agent on real data over one concrete flow. Working POC in 5 days or we don't sign.
Production, monitoring, prompt tuning and integrations. We measure real savings every month.
Service · FAQ
We never give catalogue prices. We quote after the diagnostic, once we've seen your operations, volume and integrations. Book the call and you leave with a concrete proposal.
Working POC in 5 business days or we don't sign. Full rollout with stable integrations: 3 to 6 weeks, depending on the complexity of your systems.
Yes. We're based in Barcelona but operate remotely worldwide: Spain, EU, LATAM, US, UK and Asia. Working languages ES and EN. If meeting in person works for you, Barcelona is also an option.
We work through enterprise APIs (Claude, Azure OpenAI) that don't train on your data. Your data stays in your systems. We sign NDA and DPA at no cost.
Yes. We use the official WhatsApp Cloud API or integrators like Twilio or 360dialog depending on volume.
Anthropic's Claude by default for conversation and reasoning, GPT-4.1 or local Llama 3 depending on the case, ElevenLabs for voice. The best model for each piece.
We'll tell you whether it makes sense, the stack we'd use, realistic timelines and a ballpark budget. If it fits, we move on. If not, you save months.