I find out what's getting in the way of good work, and fix it.

Mastering Microsoft Copilot

My role

Designer and speaker

Audience

Coaches and consultants

Format

Live virtual conference session

Tool

Microsoft Copilot Notebook

Working with AI

My role

Writer and designer

Audience

CEOs and founders

Format

One-page guides

Built with

Claude

Analyze

Advice on AI adoption usually arrives as long articles. Someone deciding what to hand to AI, or whether to switch on a new tool, needs one page they can check against in the moment.

Design

Each page had to end in a decision the reader could act on: hand a step to AI or keep it, turn a tool on or leave it off, or what a CEO or founder does next.

Develop

Four one-page guides, led by ten common AI workflow mistakes, each paired with what to do instead. The others give five tests for what to hand to AI and four checks to run before turning on an AI add-on.

Implement

Published on LinkedIn between August and September 2026, each as a single page people can save and come back to.

Evaluate

I measured saves, because a saved post is one people plan to come back to and use. These are my most-saved posts on LinkedIn, and the ten mistakes guide reached 664 people, about three times my other posts that month.

• 

  10 AI workflow mistakes to avoid

  Ten common mistakes, each paired with what to do instead.

• 

  Designing a workflow: what to delegate

  Five tests that sort each step into "Claude runs it" or "a person is in charge."

• 

  The trust check for any Skill

  Four checks to run before switching on an AI add-on, ending in enable, escalate, or decline.

• 

  4 ways to build trust during AI change

  A CEO or founder's guide to leading a team through AI change, with the survey data behind it.

LinkedIn Engagement, a lead-scoring app

My role

Product designer and builder

Users

Coaches and consultants who post on LinkedIn

Format

Self-hosted web app, MIT licence

Built with

Claude, Next.js, Anthropic API

Analyze

The people who react to and comment on LinkedIn posts are often warmer leads, and the notification feed loses them within a day. People needed a better way to track and engage with those potential leads.

Design

The score had to be repeatable and explainable. It's built from the user's own description of their ideal client, with their exclusions stated up front. Writing the scoring rubric used the same skill as writing an assessment rubric: observable criteria and stated exclusions.

Develop

A pipeline that pulls engagers from recent posts, enriches their profiles, and scores each person. AI reads each profile and spots the signals, and fixed point tables turn those signals into the score.

Implement

Released as open source on GitHub in July 2026, with setup instructions written for someone who has never run code.

Evaluate

The same scoring engine ran for paying clients as Link2Pipeline before I released this version. Scores come out the same run to run, and low-confidence profile matches are held for manual review instead of being shown as real leads.

Demo data, every lead shown is fictional.

Candlestick Foundations

My role

Designer and developer

Client

CIY Capital

Format

Self-paced e-learning module

Tool

Articulate Storyline

Analyze

CIY Capital wanted to teach candlestick chart analysis, a technical and highly visual skill, as a course learners could take on their own.

Design

I sequenced it from the ground up: what a candlestick is and how to read its body and wick, then bullish and bearish candles step by step, covering open and close, high and low, and the four phases of each.

Develop

I built it in Articulate Storyline with a practice activity after each candle type and an "identify the candle" quiz on dojis and candles with no wick, a top wick, or a bottom wick.

Implement

Hosted online as a live course for CIY Capital, so learners can take it on any device. It's embedded below.

Evaluate

The quiz is scored and shows learners their results. Anyone who misses questions is routed to a remediation path that reteaches the material before they finish.

Online course design

• 30+course builds managed at once, tracked in one Wrike dashboard
• QMevery course built to Quality Matters standards
• Promotedfrom design assistant to designer, then trained incoming designers and built the templates the team worked from

My role

Instructional Designer

Worked with

Faculty, the media team, and new designers

Format

Online courses

Tools

Canvas LMS, Articulate Storyline, Wrike

Analyze

Faculty brought deep expertise and complex academic content that had to work as online instruction.

Design

I lined up each course and module objective with the content, activities, and assessments that would show it was met, for instructional soundness and accessibility, against the Quality Matters rubric.

Develop

I built the courses in Canvas, adding multimedia where it supported the outcomes.

Implement

I prepared faculty to facilitate their own courses, so they felt confident and ready when they launched them for students.

Evaluate

I looked at how well each course worked once it was running, and faculty asked to keep working with me on their next courses.