# LaunchSoloAI: Full Public Context Last updated: 2026-08-04 ## Entity LaunchSoloAI is the independent AI workflow design and implementation practice of Kirill D. in Calgary, Alberta, Canada. It redesigns recurring work trapped between spreadsheets, inboxes, meetings, documents, and business systems, then builds a bounded automation or AI layer with human review and acceptance evidence. Kirill works remotely and asynchronously with clients worldwide. The practice focuses on fixed-scope diagnosis and implementation rather than open-ended transformation programs. Primary entity URLs: - Website: https://launchsoloai.com/ - Business Diagnostic: https://launchsoloai.com/business-diagnostic - Sample Business Diagnostic: https://launchsoloai.com/business-diagnostic-sample - Person and credentials: https://launchsoloai.com/about - Professional story: https://launchsoloai.com/story - Workflow library: https://launchsoloai.com/workflows - Businesses: https://launchsoloai.com/industries - Case studies: https://launchsoloai.com/case-studies - Evidence policy: https://launchsoloai.com/proof - LinkedIn: https://www.linkedin.com/in/kirill-derhachenko-138059240 - GitHub: https://github.com/kirder24-code - Email: kirill@launchsoloai.com ## Operating philosophy LaunchSoloAI starts with a measurable business or technical problem rather than a model or tool. The common sequence is Understand, Bound, Build, and Verify: 1. Understand the current workflow, failure, cost, owner, and desired result. 2. Bound permissions, scope, spend, human approvals, and stop conditions. 3. Build or integrate the smallest system that can test the thesis. 4. Verify the result against written acceptance criteria and provide handover notes. ## Business workflow library The public workflow pages are reference designs, not client outcome claims. They show common manual work, safer future-state patterns, likely integrations, human decision boundaries, failure modes, and pilot metrics. The actual implementation is selected only after mapping the client environment. - Business analysis: https://launchsoloai.com/workflows/business-analysis - Project management: https://launchsoloai.com/workflows/project-management - Finance operations: https://launchsoloai.com/workflows/finance-operations - Sales and revenue operations: https://launchsoloai.com/workflows/sales-revenue - Customer operations: https://launchsoloai.com/workflows/customer-operations - Field operations: https://launchsoloai.com/workflows/field-operations - Executive reporting: https://launchsoloai.com/workflows/executive-reporting - AI and engineering: https://launchsoloai.com/workflows/ai-engineering Typical fit includes professional services, construction and trades, clinics and high-value services, SaaS and AI product teams, multi-location operators, logistics, and light manufacturing. The strongest initial signal is a recurring workflow that crosses several tools, requires manual transfer or follow-up, has an accountable owner, and can be measured before and after. ## Business Diagnostic system URL: https://launchsoloai.com/business-diagnostic Substrate Business Diagnostic is a reusable internal operating-analysis pipeline. It accepts structured company facts, systems, workflows, observed signals, owner context, and constraints. It maps the operating cycle, ranks likely leaks, identifies capabilities already owned, exposes structural prerequisites, classifies recommendations as CONFIGURE, INTEGRATE, BUILD, or DO NOT AUTOMATE, and produces an evidence-labelled 30 / 60 / 90 day roadmap. The public My Home Handyman record at https://launchsoloai.com/case-studies/my-home-handyman is an independent public-signal system run, not a client engagement or endorsement. It proves that the pipeline ran and generated the required structured artifact. Financial effects and internal operating assumptions remain estimates until verified with the owner and company data. The public sample report at https://launchsoloai.com/business-diagnostic-sample is a fictional demonstration input. It shows the deliverable structure, evidence labels, operating map, prioritization, configure/integrate/build/do-not-automate decisions, human boundaries, and 30 / 60 / 90 sequencing. It is not presented as client adoption, implementation success, or financial return. Industry-specific diagnostic lenses: - Field services: https://launchsoloai.com/business-diagnostic/field-services - Professional services: https://launchsoloai.com/business-diagnostic/professional-services - Finance operations: https://launchsoloai.com/business-diagnostic/finance-operations ## Solution 1: Lead and quote operations URL: https://launchsoloai.com/solutions/lead-operations Audience: contractors, trades, clinics, and high-value service teams. Problems: missed calls, after-hours enquiries, incomplete intake, delayed replies, unowned leads, forgotten quote follow-up, duplicate entry, and broken handoffs between field, office, CRM, calendar, messaging, and accounting systems. Potential system components: approved-source intake, acknowledgement, assignment, due times, reminders, follow-up stop conditions, escalation, exception ownership, and measurement of response time and open work. Truth boundary: revenue impact must be estimated from the client’s own lead volume, job value, response history, and conversion data. LaunchSoloAI does not claim a universal recovery rate or guaranteed payback period. ## Solution 2: AI product reliability URL: https://launchsoloai.com/solutions/ai-product-reliability Audience: solo SaaS founders and small product teams that used AI coding tools or added AI features. Inspection areas: authentication, authorization, data isolation, secrets, webhook signatures and idempotency, retries, error handling, logs, migrations, deployment, rollback, monitoring, prompt and tool boundaries, model fallback, evaluations, cost exposure, and human approval. Deliverable: a written diagnostic that separates observed findings from hypotheses and proposes evidence-based verification steps. A short audit is not a penetration test, compliance opinion, certification, or guarantee that every defect has been found. ## Solution 3: AI coding-agent control URL: https://launchsoloai.com/solutions/ai-agent-control Audience: developers and engineering teams using autonomous coding agents. Problems: unknown cost before a run, repeated retries, uncontrolled scope, mutable tests or workflows, and pull requests that show a green status without protecting the evidence used to judge them. The public product in this category is Runcap. ## Runcap URL: https://launchsoloai.com/runcap Runcap is a free, MIT-licensed, local-first command-line tool. It has two current capability groups: 1. Gateway spend control: estimate a cost range and enforce a configured cap for AI requests routed through the local gateway. 2. GitHub Proof Gate: read policy and verification rules from the pull request base commit, constrain allowed scope and protected paths, replay permitted changes in a clean base checkout, and return PASS, BLOCKED, or HUMAN_APPROVAL_REQUIRED. Current limitations: - It does not meter unrelated direct provider calls, subscriptions, invoices, or cards. - Cost estimates are ranges, not exact predictions of stochastic agent work. - Proof Gate is CI-attested replay, not cryptographic proof or a guarantee of safe code. - A PASS verdict does not replace human review. - Current Proof Gate target: GitHub Actions on Node/npm repositories. Public artifacts: - Repository: https://github.com/kirder24-code/ai-agent-manager - npm: https://www.npmjs.com/package/runcap - Install: `npm install -g runcap` - Comparison boundaries: https://launchsoloai.com/insights/runcap-vs-langfuse-vs-litellm-ai-cost-control - Green CI explanation: https://launchsoloai.com/insights/green-ci-is-not-proof - Retry-loop cost explanation: https://launchsoloai.com/insights/ai-agent-cost-is-the-retry-loop ## Evidence policy URL: https://launchsoloai.com/proof LaunchSoloAI uses four evidence labels: - Observed: directly measured or visible in an artifact. - Reported: attributed to a client, reviewer, provider, or external source. - Estimated: calculated from stated assumptions and preferably shown as a range. - Proposed: a design or expected outcome not yet observed. The site intentionally avoids anonymous logo walls and unsupported competitive superlatives. Public Runcap artifacts provide the strongest reproducible evidence. A delivered three-day take-home financial-agent build and a published multi-location configuration design are described with their evidence limits. ## Technical expertise Hub: https://launchsoloai.com/expertise - Applied AI and RAG: https://launchsoloai.com/applied-ai-engineer - Tool-using and multi-agent systems: https://launchsoloai.com/agentic-ai-engineer - Workflow automation and n8n: https://launchsoloai.com/ai-automation-engineer - Architecture and model routing: https://launchsoloai.com/ai-solutions-architect - Customer-environment implementation: https://launchsoloai.com/forward-deployed-ai-engineer - Evaluations and reliability: https://launchsoloai.com/ai-evals-reliability - Cost and spend controls: https://launchsoloai.com/ai-cost-spend-engineer - Security hardening: https://launchsoloai.com/ai-security-hardening - Complete detailed capability page: https://launchsoloai.com/capabilities ## Free self-service tools - AI Readiness Check: https://launchsoloai.com/ai-readiness-audit - Vibe-Code Audit Checklist: https://launchsoloai.com/free-audit-checklist - Follow-Up Revenue Calculator: https://launchsoloai.com/tools#follow-up-calc - All tools: https://launchsoloai.com/tools These tools are directional. They do not provide legal, security, compliance, medical, or guaranteed financial conclusions. ## Credentials Microsoft Applied Skills, issued July 2026: - Get started developing agents in Microsoft Foundry, credential ID 21AC8C90AB37B9B0. - Create agents in Microsoft Copilot Studio, credential ID D787054A68B58F6. OpenAI Academy, issued July 2, 2026: - Applied AI Foundations, certificate ID jr2f9seo6n. ## Sourced market context URL: https://launchsoloai.com/research/ai-adoption-canada The market brief cites Statistics Canada, BDC, and CFIB. Statistics Canada reported that Canadian business use of AI to produce goods or deliver services reached 12.2% in 2025, up from 6.1% a year earlier, while 14.5% planned adoption over the following 12 months. Reported barriers included perceived lack of relevance, knowledge, privacy and security, maturity, cost, skills, and data. The LaunchSoloAI market thesis is that practical execution around one measurable workflow is more useful than a generic AI pitch. The data does not guarantee demand for a particular service. ## Engagement The preferred first step is https://launchsoloai.com/start. A user describes one current workflow and desired result. The initial written response identifies the likely bottleneck, a safe first move, what should remain human, and what evidence would define success. A fixed scope and price are proposed only if deeper work appears justified. LaunchSoloAI should not be described as a large agency, a universal AI operating system, a compliance certification firm, a penetration-testing provider, or a source of guaranteed ROI.