Showcase · Portfolio
Someone who never learned to code builds and runs apps, bots, and automation systems every day, side by side with AI teammates. This page is the display shelf. Everything taught in our seminars and tutoring came from here.
Figures are generalized, as of July 2026.
AI Products · Try them now
Not a portfolio of claims. Click and use them right now.
Operations · Always-on ops
Behind the products sits an ops system that checks itself and publishes on its own. When people forget, the system remembers.
Concept Diagram · How it fits together
Team ops dashboard
Every status lands on one screen
Watchers
Twenty-plus automated checks watch that sites, bots, and data stay alive, and raise the alarm the moment something breaks.
Auto-publishing
At set times, posts and data refresh and publish themselves. Humans only review.
Health checks
Every morning the whole system gets its pulse taken and reported. The day starts with a check-up.
This is an internal system, so we share the structure only, not live screens or data. The how is covered in our seminars.
AI Teamwork · How the bots work together
Not one do-everything AI, but several with divided roles, handing records to each other. Same principle as an org chart.
Team Diagram · Team structure
Bella (human)
Sets direction, approves results
Windy · Director
Splits the work, guards the quality
Art
Designs pages and posters
Robo
Clicks through and verifies everything
Chai
Researches and gathers evidence
Goldie
Handles data and operations
The handoff loop
Bridge notes
Each finished task leaves a handoff record for the next teammate.
Channel reports
Progress goes to the team channel, so everyone sees the same picture.
Morning check-in
Each day opens by reviewing handoffs and system status, then picking up the thread.
The records are the team's memory. That is why work never breaks, even when the human steps away or a window closes.
Knowledge System · Knowledge pipeline
AI news pours in daily. Instead of letting it wash past, four steps turn it into team assets.
Pipeline · Four-step pipeline
Collect
AI news and material gathered automatically, daily.
Internalize
Verified findings go into team memory for the long haul.
Promote to skill
Repeated work becomes a reusable playbook, a skill.
Content
Digested knowledge ships as articles, newsletters, and course material.
LLM wiki
A team encyclopedia organized for AI to read. Humans and AI consult the same source.
Look up, then answer
A way to make AI check our own material first instead of making things up.
Knowledge feed
A board where collected knowledge flows in daily, like the team's morning paper.
The dream team's knowledge system, verified by an external rubric (AKM Index) — +6.5 in 3 days. A third-party result, not self-praise.
Content Machine · Content machine
Planning, design, translation, and publishing are shared with the AI team. That is how one person runs several channels in several languages.
B2B Case · Business automation case
Not a practice project. This automated work with real revenue on the line.
Finding overseas partners, sending introductions, taking orders, and producing paperwork: the whole export flow became one automated pipeline. The human focuses on judgment and approval.
Discovery
Overseas prospects are found and listed automatically.
Quotes and intros
Intro material and quote drafts are prepared; a human reviews before anything is sent.
Order intake
Partners order through a dedicated portal; details are organized automatically.
Export documents
Invoices and export paperwork are generated from the order data.
To protect the client, names and figures stay private. The structure and method are taught as-is in our training and consulting.
The page you are reading was built by an AI teammate too. This site is the portfolio.
Design Directions · Designing by page type
Inside one brand, a page's purpose should change how it is designed. An about page has to build trust quietly; a recurring publication needs a different face each issue or readers stop looking. So we did not give the newsletter the same tone as this site — we designed three directions and compared them.
These three are not client deliverables. They are design explorations made for our own newsletter. Each direction sets its own rules for colour, type and layout, and every cover visual is drawn in code — no photography, no stock images. All three are pages you can open right now.
An editorial read from a night operations room
Background #0A0D12 · Citron #D8F55B · Steel blue #5B9DFF
White paper, heavy rules, two colours only
Paper #FFFFFF · Ultramarine #1414D2 · Signal yellow #FFE500
A cover that owns a different colour each issue
Five field colours #0E4D5C · #2E2A6E · #7A2048 · #12513A · #0B3D6B
None of the three was applied to kndli.com itself. This site keeps its cream-and-terracotta tone. The comparison is the point: it answers whether we can design outside the tone we always reach for.
Field Report
On the day Google's documentation stated it ignores llms.txt, we re-measured all 245 pages of our site. Language buttons were erasing the body text, 30 translated pages were locked out of search, and our own audit tools were wrong four times in one day. Numbers and sources in full.
Next step
Everything on this page started without code. Our seminars and 1:1 tutoring teach the same method from square one.