Skip to main content
TIEMAN.IT

AI that works for your business, not the other way around

n8n automation, custom AI bots, agents and RAG over internal knowledge bases. Built from Borne for businesses with ops pain: repetitive manual work, poor data flow between tools or internal knowledge nobody can find.

Book a callView portfolio

Tieman IT bouwt Web + Shop + AI . Dit is de AI-tak (automatisering & agents). Meer weten over websites & Next.js of webshops & Magento?

-- Diensten

Wat ik voor je bouw

-- Prijzen

Transparante tarieven

Alle prijzen zijn vanaf-bedragen. Exacte offerte volgt na een kennismakingsgesprek.

AI Quickscan

Opstart vanaf
Op aanvraag

Businesses that know there is something to automate but do not know where to start.

  • Intake of your current processes via questionnaire and call
  • Analysis of three to five concrete automatable steps
  • Advice on which models and tools fit your data and privacy requirements
  • Written report with priority order: impact versus implementation complexity
  • Honest assessment of what does not make sense for your situation
  • Optional: verbal walkthrough after delivery
Plan kennismaking →
Meest gekozen

n8n / Automation

Prijs
Op aanvraag

Businesses with clear recurring workflows they do not want to manage themselves.

  • Self-hosted n8n on your own VPS or infrastructure
  • Connections to existing tools: CRM, email, ERP, webhooks
  • Error handling and notifications built in
  • Documentation of every workflow for internal management
  • Optional retainer: monthly adjustments and monitoring
  • No per-task fees, fully yours
  • Handover with explanation so your team can make changes
Plan kennismaking →

Custom AI bot or agent

Prijs
Op aanvraag

Businesses with context-dependent tasks where a simple script falls short.

  • Task analysis and selection of the right model (Claude, GPT or local)
  • Prompt design and structured output tailored to your use case
  • RAG setup if the bot needs to reason over your own documents
  • Integration with existing systems via API or webhook
  • Logging and fallback so errors are visible
  • On-premise option for data sovereignty
  • Full source code in your ownership
Plan kennismaking →

Prijzen excl. btw · Hosting apart gefactureerd

-- Veelgestelde vragen

Heb je een vraag?

That depends on the task. For reasoning and instruction-following I often use Claude. For broader integrations or specific needs I choose OpenAI OpenAI's latest model or a local model via Ollama. I am not tied to one provider: I pick what fits your data, privacy requirements and budget.
I always discuss this at intake. If you process customer data or personal data, I first look at whether a local model (on-premise, no cloud API) is the safest option. For less sensitive data, Claude or OpenAI are fine as long as you document the processing properly. I do not provide legal GDPR advice, but I will not build a system that unnecessarily exposes you to risk.
Both. n8n is my first choice for connecting existing tools: it saves build time and your team can adjust it later. If the logic is too complex for a workflow tool, or if performance matters, I build a custom service in Python or Node. I always present the choice with the trade-offs laid out.
I do not build marketing chatbots purely for show without a measurable business case. I do not produce twenty-page AI strategy reports without implementation. I also do not deliver out-of-the-box SaaS integrations where there is nothing to build. If your request is a better fit for a different type of provider, I will tell you directly.
I start with an intake: which processes cost time, where does data go wrong, what do you want to change. I then deliver a short analysis: what is feasible, what it delivers, which order has the most impact. Once agreed, the build happens in iterations. Each iteration I deliver something working, not interim deliverables that only exist in my head.
That depends entirely on scope. A focused n8n workflow is considerably cheaper than a fully autonomous agent with RAG and its own infrastructure. After intake I provide an honest quote based on actual complexity. No fixed packages that do not fit, no surprises afterwards.
Yes. Via Ollama I run local models on a VPS or dedicated server. That is relevant when you have sensitive internal documents you do not want to send to a cloud API. The quality of local models is lower than the large cloud providers, and I discuss this honestly so you can make the trade-off clearly.

Ready to eliminate manual work?

Tell me which process costs time, which tools you use and where data gets stuck. I will come back with an honest proposal. No sales pitch, no bullshit.

Book a callView my cases