# AI & Automation

> Canonical site: https://simnetiq.store/en/services/ai-automation
> Last generated: 2026-05-10

## AI & Automation (C-03)
URL: https://simnetiq.store/en/services/ai-automation
Badge: AI · PIPELINES · AGENTS
Tagline: Production AI, wired to real pipelines.

Any model — Anthropic, OpenAI, open-weight, or fine-tuned. Any pipeline — n8n, Make.com, custom Node or Python on a dedicated VPS. Any data source — Postgres, S3, Notion, Stripe, Telegram, your own SaaS. We connect them into automations that survive production: scheduled, observable, retry-safe.

Most AI work stops at a chatbot demo. We build the operational layer around it — the triggers, the data pipes, the retries, the observability — so the model becomes a system, not a feature flag.

### Sub-services
- **S-01 · AI Feature Engineering** — Prompt architecture, caching, streaming, structured outputs, tool-use. Built against evaluations, not opinions.
- **S-02 · Retrieval & RAG** — Embedding stores (pgvector, Pinecone, Weaviate), hybrid search, re-ranking, chunking strategies tuned per corpus.
- **S-03 · Agentic Workflows** — Multi-step agents with verifiable side effects. Tool catalogs, guardrails, cost ceilings, replay-safe execution.
- **S-04 · n8n & Make Pipelines** — Self-hosted n8n on your VPS or ours. Webhooks, schedules, retries, dead-letter queues, full audit trail.
- **S-05 · Custom Pipelines on VPS** — When n8n isn't enough — TypeScript or Python services on a dedicated VPS, queued via BullMQ or Redis Streams.
- **S-06 · Data & Vendor Integrations** — Postgres, Supabase, Stripe, HubSpot, Telegram, Slack, Notion, S3, Sheets, custom REST or GraphQL — wired in cleanly.

### Tech stack
- **AI Providers**: Anthropic Claude, OpenAI, Google Gemini, Mistral, Together AI, Replicate
- **Open-Weight & Local**: Llama 3, Qwen, DeepSeek, Ollama, vLLM, LM Studio
- **Vector & Retrieval**: pgvector, Pinecone, Weaviate, Qdrant, Chroma, Voyage AI
- **Pipelines**: n8n, Make.com, Temporal, BullMQ, Trigger.dev, Inngest
- **Runtimes**: Node.js, Python, Bun, Deno, Docker, Caddy / nginx
- **Infra**: Hetzner VPS, Railway, Fly.io, Supabase, Cloudflare Workers, Vercel

### Pricing tiers
### Pilot — FROM £1.5K
One AI feature, end-to-end, in two weeks.
  - Single workflow scoped against a real use case
  - Anthropic or OpenAI integration with caching
  - Basic eval harness
  - Deployed to your stack or ours

### Production — FROM £4K
AI feature plus the pipeline that feeds it.
  - LLM feature + retrieval / RAG layer
  - n8n or custom pipeline on a dedicated VPS
  - Vendor integrations (Stripe, Postgres, Telegram, etc.)
  - Observability, retries, cost ceilings
  - Eval harness + first month of monitoring

### Platform — CUSTOM
Full AI + automation platform built on your data.
  - Multi-feature AI surface
  - Self-hosted models or hybrid
  - Custom agentic workflows with guardrails
  - Full pipeline + data integration suite
  - Ongoing operational ownership

### Meta
- **Models**: Anthropic · OpenAI · Local
- **Pipelines**: n8n · Make · Custom
- **Typical engagement**: 2–10 weeks