Projects in AI automation, operations, and systems design.

The Problem

Norrita was manually tracking hundreds of property tax protest cases across spreadsheets, email threads, and paper files. During peak season, admin work was consuming time that should have gone to taking on new cases and serving existing clients.

What Was Built

Built a full case management system from scratch — a web application with a case intake portal, document checklists, case timelines, an invoice calculator, and a payment ledger. Every active case, document, and payment in one place, accessible from anywhere.

The Problem

Tosha was launching a solo accounting practice with the technical skills but no operational model. Without a structured intake process, pricing framework, or automation plan, scaling beyond a handful of clients would have required proportional hour increases — a ceiling with no solution.

What Was Built

Delivered a comprehensive AI Business Efficiency Audit: a 16-section operational blueprint covering service architecture, flat-rate pricing, a full client lifecycle workflow, a 9-automation AI implementation map built on the Solution Sphere framework, month-by-month revenue projections, a 30-day launch roadmap, a lean tech stack under $200/month, and a compliance and risk framework.

The Problem

A 16-section AI Business Efficiency Audit is the core deliverable of every client engagement. Writing one manually — synthesizing interview notes, mapping automations, building revenue projections, and formatting a professional document — takes 8 to 12 hours per client. At that pace, delivery speed becomes a constraint on how many clients can be served.

What Was Built

Built an AI Audit Report Generator that ingests the Fireflies transcript from the client interview and outputs a fully structured, client-ready audit report following the Solution Sphere framework. Covers all 9 sections: executive summary, service architecture, workflow design, automation opportunity map, revenue model, 30-day roadmap, tech stack, success metrics, and compliance framework. Also generates a companion presentation deck via NotebookLM with slide-by-slide speaker notes for the client delivery call.

The Problem

Every client engagement follows a 12-step process from discovery call to delivery. The proposal sits at Step 4 — after the audit is complete and before the contract is signed. Writing a custom proposal manually after every audit added hours of work between a high-value deliverable and a signed contract, slowing the sales cycle at its most critical moment.

What Was Built

Built an AI Proposal Generator that pulls data from both the pre-audit discovery interview and the completed AI Audit Report to produce a tailored proposal. Covers engagement scope, deliverables, timeline with milestones, investment, and next steps. Structured to under-promise and overdeliver on timeline so every engagement starts with a client who feels ahead of schedule. Output is ready to present at Step 4 of the 12-step client process — same day the audit is delivered.

The Problem

Running an AI consulting practice without its own operating infrastructure created a credibility gap. The business that builds AI systems for clients needed to demonstrate those same systems internally.

What Was Built

Built the full Circa AI Group website, including brand system, hero copy, case study metrics bar, and an end-to-end GitHub → Vercel → Cloudflare deployment pipeline.

The Problem

Institutional knowledge about clients, frameworks, and systems lived across browser tabs, chat histories, and memory. Every new engagement started from zero instead of compounding on what already existed.

What Was Built

Built a structured Obsidian-based second brain wiki — a private knowledge base pre-loaded with Circa AI Group's full business context: the Solution Sphere framework, client notes, ICP profiles, offer architecture, and operational playbooks. Functions as a living internal wiki that grows with every engagement.

The Problem

Identifying and reaching qualified prospects at scale requires either significant manual effort or expensive third-party tools. Neither option made sense for a lean solo consulting practice that needed a replicable outbound system.

What Was Built

Built a three-agent podcaster lead scraper pipeline: Agent 1 scrapes show directories for qualifying criteria, Agent 2 enriches each lead with contact data and context, Agent 3 drafts personalized outreach sequences. The full pipeline runs end-to-end with minimal manual input.

The Problem

Solo practitioners often struggle to articulate what makes their practice distinct — not for lack of self-awareness, but because they have never had a structured way to surface it. Generic brand exercises produce generic answers.

What Was Built

Built an interactive Ikigai-based brand assessment tool as a React component. Prospects and clients work through four intersecting dimensions — what they're good at, what they love, what the market needs, and what they can be paid for — to surface a positioning statement grounded in their actual strengths rather than what sounds good in a bio.

The Problem

LinkedIn content for consultants defaults to either generic advice or self-promotional announcements. Neither builds authority with a specific ICP. Publishing without a repeatable content system means inconsistency — the fastest way to lose traction on the platform.

What Was Built

Developed a 10-angle thought leadership content framework scored against Circa AI Group's ICP targeting CPAs, property tax consultants, and insurance brokers. Each angle has a defined purpose, example structure, and scoring rationale. Functions as an editorial calendar backbone that produces relevant, non-repetitive content across any posting cadence.

The Problem

Virtual and executive assistants are being displaced by AI tools they don't know how to use. Agencies offering traditional VA services face a positioning problem: the market is shifting toward operators who can configure and manage AI workflows, not just execute tasks.

What Was Built

Designed the CPA Executive Assistant workshop product — a training program for VA/EA agencies that repositions their staff as AI-powered operators. Covers workflow automation, prompt engineering for admin tasks, tool configuration, and client communication systems. Delivered via Pathwright with a structured curriculum and certification pathway.