# ArcanEdge

> ArcanEdge is an AI product and systems studio that turns complex workflows, knowledge, and ideas into working software.

## Reading formats
- [Complete public-site text](https://www.arcanedge.ai/llms-full.txt): All public page content in one generated text document.
- Every canonical page links to its Markdown version through rel="alternate". The homepage version is [index.md](https://www.arcanedge.ai/index.md); other paths append .md, for example [about.md](https://www.arcanedge.ai/about.md).
- The HTML pages remain the canonical source. These reading formats contain the same public content and are generated during the same build.

## Canonical pages
- [What We Build](https://www.arcanedge.ai/what-we-build): AI product development, knowledge systems, agent-assisted workflows, enterprise operations, and AI behavior architecture.
- [Our Approach](https://www.arcanedge.ai/our-approach): Discovery, design, build, validation, deployment, and continued improvement.
- [Selected Work](https://www.arcanedge.ai/work): Public products and approved client work.
- [About](https://www.arcanedge.ai/about): Company and founder information.
- [Insights](https://www.arcanedge.ai/insights): Practical articles about production AI systems.
- [Contact](https://www.arcanedge.ai/contact): Project inquiries.

## Answer-ready guides
- [RAG, agents, or conventional software?](https://www.arcanedge.ai/insights/rag-agents-or-conventional-software): Choose between conventional software, RAG, and agents by matching each pattern to the work, risk, knowledge, and judgment involved.
- [Where human review belongs in an AI workflow](https://www.arcanedge.ai/insights/human-review-in-ai-workflows): Learn where approval, correction, and accountability belong in an AI workflow without slowing every low-risk action.
- [From source to answer: designing evidence-led AI](https://www.arcanedge.ai/insights/designing-evidence-led-ai): Design AI systems that preserve source identity, evidence, citations, confidence, limitations, and human review from input to delivery.
- [What belongs in a product requirements document for AI?](https://www.arcanedge.ai/insights/ai-product-requirements-document): Define AI behavior, knowledge, uncertainty, evaluation, human responsibility, and operating boundaries in a buildable product requirement.
- [Why prompts should be versioned like product behavior](https://www.arcanedge.ai/insights/version-prompts-like-product-behavior): Treat prompts as product behavior by versioning models, tools, schemas, retrieval settings, evaluations, releases, and rollback plans.

## Public work
- [Dream Oracle](https://www.arcanedge.ai/work/dream-oracle): Guided AI reflection, private journaling, and pattern awareness across web, Android, and iOS. Contribution: Founder, designer & full-stack developer.
- [United Tradesmen](https://www.arcanedge.ai/work/united-tradesmen): The public website is live. The workforce and operations platform is moving into beta with real testers. Contribution: Product & technology lead.
- [Nordic Road](https://www.arcanedge.ai/work/nordic-road): Multilingual, location-specific intelligence organized around sources, evidence, analysis, and human review. Contribution: Backend design, implementation & AI architecture.
- [ClientIQ](https://www.arcanedge.ai/work/clientiq): Prompt and conversation design, product planning and delivery coordination for Kora, an assistant connected to firm information, case data and staff review. Contribution: Prompt engineer, product & workflow lead.
- [Adina Cosmetic Ingredients](https://www.arcanedge.ai/work/adina): An internal document assistant for support and sales teams, with source references and highlighted details across technical material, tables, images and scans. Contribution: Prompt engineer & project manager.
- [Digital Starr](https://www.arcanedge.ai/work/digital-starr): GPT instructions and integration requirements for a customer-data assistant using SQL queries, validation, record matching and human review of uncertain matches. Contribution: Prompt engineer & project manager.
- [InstaLetter](https://www.arcanedge.ai/work/instaletter): Questionnaire and prompt workflows connecting a person's background with a target role to produce tailored resumes and cover letters. Contribution: Prompt engineer & project manager.
- [Salesbod.ai](https://www.arcanedge.ai/work/salesbod): Custom GPT instructions, sales conversation flows and search-based business context, refined with the development team across product updates. Contribution: Prompt engineer & project manager.
- [Adamo](https://www.arcanedge.ai/work/adamo): Earlier client work by Bruce Meek on prompts, persona and conversation flows, with repeated revisions to keep assistant behavior consistent across day-to-day tasks. Contribution: Prompt engineer & conversational AI project manager.

The portfolio includes released products, work in development and earlier client contributions. Each case study distinguishes the contributor's role and scope. A public website or portfolio entry does not mean the full platform is generally available.

## Contact
- Email: info@arcanedge.ai

Last updated: 2026-10-26