FAQ
Questions we hear before the first prototype
Candid answers about CogniForge AI — a Vancouver applied-AI studio that builds retrieval-augmented generation systems, AI assistants and workflow automation for Canadian organizations.
We prefer direct questions. The list below covers engagement models, indicative CAD budgets, timelines, tooling choices, PIPEDA data handling, intellectual property and the limits of generative AI — because those limits matter as much as the capabilities.
Is CogniForge AI a course or AI-income scheme, and is the .life about wellness or metalworking?
No. We are a knowledge-driven applied-AI studio that designs and builds custom AI systems for client organizations — not a course, not an "AI income" or passive-income scheme, and not a training academy. The .life domain is branding only; we are not a life-coaching or wellness business. "Cogni" refers to cognition and knowledge — not neuroscience or brain-training products. "Forge" means building AI systems with craft — not blacksmithing, metalworking or gaming mod culture.
We deliver AI strategy, generative-AI applications, large language model assistants, AI agents, retrieval-augmented generation, workflow automation, machine learning models, data pipelines, model evaluation and MLOps as professional services. Humans stay in the loop. We do not guarantee accuracy, cost savings, revenue or ROI — outcomes depend on your data quality, project scope, budget and adoption.
How do engagements work — project vs retainer?
Most new client relationships begin with a fixed-scope discovery sprint or proof of concept project. You receive a written roadmap, prototype or production-ready build depending on scope. When systems are live, many clients move to a retainer for MLOps, guardrail updates, evaluation harness maintenance and support. Retainers are structured monthly with defined hours and response expectations — not open-ended feature factories.
We quote CAD project fees before work starts. Change requests that alter scope are discussed transparently. We are an AI consultancy focused on custom build delivery, not staff augmentation by default.
What are typical CAD budgets?
Indicative ranges (not quotes): discovery sprints C$18,000–C$35,000; RAG and knowledge-system builds C$85,000–C$220,000; document intelligence pipelines from C$60,000; agent workflows vary widely with integration depth. Retainers often start around C$6,500/month for evaluation and MLOps support.
Budget depends on corpus size, security requirements, number of integrations and how clean your source documents are. We will tell you if preprocessing will dominate the budget — that honesty saves everyone time.
How long does a prototype take?
A focused prototype loop — retrieval pipeline, basic assistant UI, evaluation on your question set — often lands in six to ten weeks after kickoff, assuming document access and stakeholder availability. Production deployment with full guardrails, API integration and MLOps takes longer. Discovery-only engagements may finish in two to four weeks.
Timelines are agreed in the statement of work. We do not promise instant delivery or guaranteed measurable outcomes by a fixed date.
Which models and tools do you use?
We are model-agnostic pragmatists. Engagements may use commercial LLM APIs, open-weight models, vector databases, orchestration frameworks and custom ML components. Cloud, hybrid and on-prem patterns are all in scope when data privacy requires them. Tool choices are documented in architecture decisions — we do not lock you into a vendor because of our preferences.
Fine-tuning, prompt engineering and retrieval design are selected based on evidence from prototypes, not hype cycles.
How do you handle data privacy and PIPEDA?
Canadian client data is handled under PIPEDA-aligned practices: purpose limitation, consent where required, access controls, encryption in transit and at rest, retention schedules and breach notification procedures. We minimize what enters model training versus inference-only use. Cross-border processing is disclosed and contractually governed when it occurs.
Contact [email protected] for access or correction requests. See our Privacy Policy for full detail.
Who owns the code, models and IP?
Custom code, prompts, retrieval configurations and integration work product specified as deliverables are assigned to the client upon payment unless otherwise agreed in writing. Pre-existing studio tools, libraries and frameworks remain ours, licensed for your use as part of the deliverable. Third-party model weights and API terms follow their providers.
We do not claim ownership of your documents or business data. Client corpora are processed only for agreed purposes.
How accurate are the assistants you build?
Accuracy depends on source quality, coverage, chunking strategy and how questions are phrased. Retrieval-augmented generation reduces but does not eliminate hallucinations. We implement evaluation harnesses, confidence thresholds, citation requirements and human-in-the-loop escalation — and we report limits honestly.
We do not guarantee zero errors, full autonomy or replacement of your team. AI outputs require human review for consequential decisions.
What do you explicitly not do?
We do not sell courses or certifications. We do not promote AI-income, crypto-trading or forex bots. We do not claim AGI, sentience or guaranteed ROI. We do not fabricate client logos or partner badges. We do not build systems designed to run with zero human oversight for high-stakes outcomes. We are a Vancouver AI studio focused on dependable knowledge automation — interesting problems, honest limits.
Answers here describe how we scope knowledge builds — not promises of perfect retrieval, fixed ROI or hands-off automation. Every production system keeps human review in the loop.