Working AI inside the perimeter
We deploy inference, index your documents, set up answers with sources and roles. The pilot checks quality on real staff questions — with acceptance criteria in the contract.
The pain we close
You need a “corporate ChatGPT” without the cloud
Documents are scattered — AI without RAG is useless
No clear pilot with metrics
After launch, nobody is there to support it
What’s included
Model stack
Llama / Qwen / Mistral and others — for Russian and the job.
RAG
Indexing, updates, answers with sources.
Pilot
4–8 weeks, quality metrics, handoff.
SLA
Monitoring, incidents, additional training by plan.
Related experience
ChatNeuron
ChatNeuron — cloud AI agent and widget · CIS-ready
We built NeuronChat as SaaS: portal with dashboard, agents, dialogs, leads, analytics, and knowledge base. Site widget, RAG training, multiple agents for different domains in one account. CIS-ready: multilingual AI and local CRM/1C integrations.
B2C NDA
AI Tutor: smart learning for kids
We built an MVP with AI for kids’ learning: assignment control, performance, question bank. The client did not develop the product further — the case stands as edtech-MVP launch experience.
FAVORIT
Costbl — SaaS cost estimation for metal parts from drawings
More than 5 months, two stages. Stage 1 — prototype: AI chat and PDF parsing (screenshots in the stage 1 block). Stage 2 — current state: full calculation (video in the stage 2 block). Product: https://costbl.ru/
Guides
Aligned with the overall On-Premise AI estimate on the hub and /prices.
| Package | What’s included | Price | Timeline |
|---|---|---|---|
| RAG / chat pilot | Model, document corpus, roles, report | from $5,682 | 4–8 weeks |
| Prod + SLA | Monitoring, knowledge-base updates, changes | by plan | monthly |
The full quote drivers block is on On-Premise AI and in pricing.
FAQ
How is this different from ChatNeuron?
ChatNeuron is a fast product contour for typical conversations. Here — local models and data strictly inside the client’s perimeter.
What do you measure on the pilot?
Share of answers with a correct source, human escalations, response time, pilot-group satisfaction — we lock these in advance.
Launch a RAG pilot
A sample document corpus and who will use it — enough for a 4–8 week plan.
