FAQ
Direct answers about our studio, scope and limits
Generative AI raises fair questions. Here is how Mind Nexus AI operates as an applied-AI product studio in Vancouver.
Is Mind Nexus AI a mental-wellness or brain-training app, or do you guarantee the AI is always right?
No on both counts. We are a generative and conversational-AI product studio that builds assistants, copilots and knowledge tools for client organizations. “Mind” means machine reasoning and language AI—not the human mind, therapy, wellness or coaching. The .life domain is branding only. Generative AI and LLMs can hallucinate or produce biased output; we build guardrails, evaluation suites and human handoff paths. We do not guarantee accuracy, adoption or business outcomes—those depend on your data, use case and scope.
Do you sell courses or subscription software?
No. We deliver project and retainer engagements as a professional applied-AI studio. Training appears as part of implementation so your team can maintain prompts and knowledge bases—not as a standalone academy product.
Project versus retainer—how do engagements work?
Projects suit scoped builds: a copilot pilot, RAG over a defined corpus, or an evaluation suite. Retainers suit ongoing prompt maintenance, model updates, integration changes and governance reporting. Discovery clarifies which model fits your roadmap and CAD budget.
What budgets should we plan for?
Discovery and roadmapping often starts around C$18,000–45,000. Production conversational AI and RAG builds typically range C$35,000–110,000 depending on integration depth. Retainers run C$5,500–18,000 per month. We quote precisely after discovery—ranges are indicative, not offers.
How long until a pilot is live?
Scoped pilots often ship in twelve to sixteen weeks from signed scope, assuming data access and stakeholder availability. Complex enterprise integrations take longer; we will say so upfront.
Which models and stacks do you use?
We are model-agnostic: OpenAI, Anthropic Claude, Microsoft Copilot, open-weight models on your infrastructure, or hybrid setups. Choice depends on latency, cost, data residency and evaluation results—not vendor religion.
How do you handle data privacy and PIPEDA?
We collect only what engagements require, document purposes in our privacy policy, and configure processing to meet PIPEDA and BC PIPA where applicable. Cross-border subprocessors are disclosed before use. Contact [email protected] with access requests.
Where is data processed and stored?
Depends on architecture: client VPC, Canadian cloud regions, or approved third-party APIs. We document locations in contracts and privacy schedules—no hidden routing.
Who owns IP and custom prompts?
Custom integrations, prompts and evaluation assets created for you are yours under standard contract terms. We retain general methodologies and non-client-specific frameworks.
How do you handle mistakes and hallucinations?
Evaluation before launch, production logging, retrieval quality checks, user feedback loops and defined escalation to human reviewers. We design for visible failure, not silent wrong answers.
What will you not do?
We do not promise guaranteed ROI, job replacement, trading bots, crypto schemes, MLM pitches, wellness apps, or autonomous AI with no human oversight. We do not sell personal data or offer legal, medical or investment advice.