AI in your product,not in a demo.
Assistants, agents, search over your documents, content generation: we integrate AI where it saves time or money, with guardrails, quality measurement and controlled costs.
Who it is for
For companies moving from trying ChatGPT to a reliable AI feature, integrated with their tools and data.
At a glance
- What you get
- AI features built into your product, evaluated before launch, with costs tracked per use.
- How long
- 2 to 4 months → 9 to 12 months
- Ranges published by the market
- How much
- $25,000 – $60,000 → $120,000 – $300,000
- Ranges published by the market
01What we build
The scope,plainly.
- aAssistants and chatbotsConnected to your documents, knowledge base or customer support.
- bAI agentsMulti-step tasks, with human approval where the risk calls for it.
- cSearch over your documents (RAG)Answers that cite your internal sources instead of making things up.
- dAI features in a SaaSAnalysis, drafting, recommendations, right inside your product.
- eQuality evaluationA proof of concept on your own data first, then test sets that measure answers before and after every change.
- fData and privacyWhat the model sees, what it keeps, and the assessment Quebec's Law 25 requires.
02Proof
Already built,and you can check it.
Rehearse has AI play the person across the table, then scores the conversation. Resistaa uses AI to help start a blocked task. Both are in production: Rehearse in three languages, Resistaa in ten.
03Frequently asked questions
Before youwrite to us.
01How much does an AI project cost?
Apptunix publishes $25,000 to $60,000 for an AI chatbot and $30,000 to $300,000+ depending on complexity in Canada. Our costs page shows these figures and their sources.
02Are our data used to train the models?
It depends on the provider and its contract. That choice is made at scoping and documented in the privacy impact assessment Quebec's Law 25 requires.
03What is RAG?
Retrieval-augmented generation: the AI first searches your documents, then writes its answer from what it found, with the sources.
04Can AI be wrong?
Yes. That is why we measure quality on test sets, limit what the AI can do alone and keep human approval on sensitive decisions.
→Maison GR
AI that serves your business, not your slide deck.
Describe your project in a few lines: we reply with a first direction and an order of magnitude.