Frequently asked questions.

A solo operator provides high-touch, direct access to the principal architect. Unlike large agencies that hand off execution to junior developers, you work exclusively with a senior strategist who intimately understands both the underlying technical execution and the fundamental business outcomes required.

Yes, our practice specializes in deep operational integrations for staffing agencies, independent insurance brokerages, and digital marketing firms across Calgary, Alberta, and Western Canada. We understand the specific regulatory compliance requirements, legacy software constraints, and daily workflow bottlenecks unique to these verticals.

Steelhead is headquartered in Calgary, Alberta. We serve small and mid-market businesses across North America, bringing a deep understanding of Canadian regulatory environments while operating seamlessly across borders via remote, asynchronous engagements.

We operate at the intersection of business strategy and technical implementation. Before writing a single line of code, we conduct extensive process mapping to ensure the technology solves actual business problems, rather than deploying AI merely for the sake of novelty. For Calgary and Alberta clients, we offer onsite discovery workshops to map the highest-friction operational bottlenecks in person.

Freelancers require you to act as the project manager, providing them with exact technical specifications. We act as strategic partners. We diagnose the business problem, design the architectural solution, build the system, and handle the change management required for your team to adopt it.

Paid Discovery is a deep-dive diagnostic lasting one to two weeks. We map your current workflows, audit your data readiness, identify high-ROI automation opportunities, and deliver a comprehensive architectural blueprint. This guarantees a fixed price and timeline for the subsequent build phase. Book a discovery call to discuss whether Paid Discovery is the right next step.

We maintain transparent, asynchronous communication through a dedicated client portal. You will receive weekly video updates demonstrating functional progress, alongside scheduled review milestones. This ensures continuous alignment without trapping your team in endless status meetings.

Post-deployment includes comprehensive team training, documentation handover, and a 30-day hyper-care period to resolve any immediate friction. Following this, clients typically transition to a maintenance retainer for ongoing system optimization and API management.

Weeks 1 to 2 focus on Paid Discovery and scoping. Weeks 3 to 6 involve building the core data pipelines and AI logic. Weeks 7 to 8 are dedicated to user-acceptance testing and refining prompts, with final deployment and training typically occurring by Week 9.

We encourage internal capacity building. All systems are built using standard, modern codebases and documented rigorously. We provide comprehensive technical handover sessions so your internal IT staff or future engineering hires can assume full control of the infrastructure seamlessly.

Custom systems are inherently modular. Because we build on modern, flexible architectures, we can easily scope and integrate new features, data sources, or logic branches as your operational requirements evolve over time.

We use rigorous staging environments. The AI is tested against historical, synthetic data to ensure edge cases are handled gracefully. Only after your team successfully completes User Acceptance Testing in the sandbox do we push the system to production.

No. While structured data is beneficial, modern Large Language Models excel at extracting structure from messy, unstructured data formats like PDFs, emails, and loose documents. Part of our engagement often involves building the pipelines to finally organize your existing data.

Absolutely not. We build robust, programmatic pipelines using enterprise API endpoints. This involves vector databases for proprietary knowledge retrieval (RAG), deterministic logic routing, and integrations directly into your CRM or ERP, creating a bespoke, secure internal application. More on RAG and how it differs from a chatbot.

Consumer models (like the free version of ChatGPT) often use your chat history to train their public algorithms. Enterprise models, accessed via secure APIs, guarantee data privacy, offer higher rate limits, and provide deterministic controls over the AI's behavior and outputs.

We require access to a representative sample of your data to train and test the systems. However, we use anonymized or synthetic data sets during the initial build phase whenever possible to minimize risk and adhere strictly to data governance protocols.

Yes, through a process called Retrieval-Augmented Generation (RAG). Instead of relying on the AI's general knowledge, we connect the model directly to your proprietary company manuals, past reports, and historical data, forcing it to generate answers based exclusively on your specific business context.

Yes. We design architectures that adhere to the 10 Fair Information Principles of PIPEDA. We use enterprise environments with strict data residency options, ensuring personal identifiable information is encrypted, access-controlled, and never used for public model training. See our privacy policy for the full details.

Depending on your compliance requirements, we can configure cloud infrastructure and AI endpoints to ensure data residency strictly within Canadian borders, using enterprise-grade providers like AWS, Azure, or Google Cloud regions located in Canada.

While our boutique firm relies on the underlying SOC 2 and ISO 27001 certifications of our enterprise infrastructure partners (like Microsoft and AWS), we engineer all data flows and access controls to comply with these rigorous international security standards.

We build systems emphasizing algorithmic transparency. Our architectures log inputs, prompts, and outputs, creating a clear audit trail. This ensures that if a regulator or auditor questions an automated decision, you have the historical data necessary to defend the process.

Yes. Steelhead maintains comprehensive general liability and Errors and Omissions (E&O) insurance specific to technology consulting and software development, providing our clients with an additional layer of financial protection and professional assurance.

Your organization retains full ownership of the intellectual property, workflows, and custom code developed specifically for your operational use case. You are never locked into proprietary black-box platforms; you own the infrastructure we build.

The initial consultation is a focused 30-minute session. We respect your time, focusing immediately on diagnosing your operational bottlenecks rather than presenting a lengthy, generic sales deck. Pick a time that works for you.

No technical preparation is necessary. Simply come prepared to discuss the most frustrating, time-consuming manual processes in your business, the software tools your team currently uses, and the specific business outcomes you are hoping to achieve.

The call is most productive when attended by the primary business owner or the operational leader who intimately understands the daily friction points of your team's workflows and has the authority to sponsor process improvements.

If we determine there is a strong mutual fit, the next step is typically to propose a Paid Discovery engagement. Because we build custom systems, we do not issue generic proposals; we issue precise architectural blueprints based on deep diagnostic work.

We are honest about technical feasibility. If your bottlenecks are better solved by standard automation, a new CRM, or simply changing an internal process, we will tell you directly and recommend alternative solutions or partners.

Hourly billing penalizes efficiency and creates budget anxiety. Fixed-fee pricing aligns our incentives with yours: delivering a functional, high-value solution as efficiently as possible. Bespoke system builds typically range from $15,000 to $75,000 based on complexity. You know exactly what the investment will be before the build begins, with zero surprise overages.

Yes. AI ecosystems evolve rapidly. Our retainers provide proactive API maintenance, prompt tuning, system monitoring, and ongoing optimization. This ensures your systems remain highly performant without the need to hire internal machine learning engineers.

Yes. Recurring costs are usage-based, not a flat fee. They cover API calls to providers like Anthropic or OpenAI (priced per token), plus cloud infrastructure for storage, compute, and monitoring. The exact amount tracks your workload. We forecast it during the discovery call once we understand your specific use case. More on AI cost structures.

We generally operate on flat-fee milestone payments to keep procurement simple. However, for highly quantifiable projects with significant bottom-line impact, we occasionally structure value-based pricing models tied directly to the operational savings the system generates.

Standard project builds are typically billed incrementally: a percentage upfront to commence development, a milestone payment upon successful user-acceptance testing, and the final balance due upon full deployment and handover.

If you are on an active maintenance retainer, we proactively monitor API deprecations and system errors, resolving them before they impact your operations. Without a retainer, we offer break-fix support at a premium priority hourly rate.

Yes, retainers operate on month-to-month or annual agreements with flexible cancellation terms. Offboarding includes a final technical audit, transfer of all administrative credentials, and a final documentation handover to ensure a smooth transition to your internal team.