SELECTED WORK / THREE DEEP DIVES

The decisions.
The work behind them.

Enterprise complexity, product direction and the details that make a workflow work. Three projects, with the original artifacts and an honest account of the evidence.

Re-designing the product with a newly acquired startup: original interface artifact01 / CASE STUDY

ENTERPRISE SYSTEMS / PRODUCT INTEGRATION

Make complex automation feel like one product.

Integrating an acquired no-code builder into an enterprise platform, from the first template to the failed run.40% faster setup, reported in the published caseRead the decisions & evidence ↗
Deep Data based Product Designing: original interface artifact02 / CASE STUDY

PRODUCT STRATEGY / RESEARCH & EXPERIMENTATION

Turn behavioral evidence into product direction.

Reframing a multi-product website around how prospects evaluate, how customers return, and what actually leads to retained use.129% conversion increase, reported in the published caseRead the decisions & evidence ↗
Redesigning a new Ad Platform: original interface artifact03 / CASE STUDY

B2B WORKFLOWS / PRIORITIZATION & DELIVERY

Make an audience wizard easier to finish.

Using research severity, a focused scope and engineering collaboration to improve a complex advertising setup workflow.7 users informed the focused Audience Wizard redesignRead the decisions & evidence ↗

Selected from my published project archive. Outcome figures are source-reported; each case makes its measurement limits explicit.

THE AI-NATIVE DESIGN KIT · EVERYTHING THE PROCESS NEEDS

The kit behind the work.
Process, templates, components, patterns.

01 · PROCESS MANUALFrom one-line brief to shipped design

The 10-step AI-native design process: 52 tools, 19 templates, 30 failure modes - applied end to end on a live agentic product.

02 · HARNESSES & WORKFLOWSThe workflows that run the process

Repeatable AI-native design workflows, documented in the open as markdown files.

03 · TEMPLATE KIT17 fill-in artifacts

Behavior specs, grounding briefs, eval rubrics, failure taxonomies.

04 · COMPONENTS12 AI-native components

The states to design for each, plus the 6-step guide to creating your own.

05 · LIVE PATTERNS78 patterns, running in the browser

Streaming, approvals, confidence, sources, recovery - inspect the behavior live.

06 · DESIGN SYSTEMAgent-ready foundation

Tokens and rules agents can consume, with a machine-readable manifest.

07 · SKILLSInstallable design skills

SKILL.md files that make the judgment portable to your own agent.

BENCHMARK SNAPSHOT · 13 NOV 2026Kimi K3 leads open-weight text-to-HTML at 1335 Elo↗5,139 contributions · Sep 1 - Nov 13, 2026 · github.com/ayodhyamohanthy ↗
More from the project archive

Earlier studies and explicitly labeled concepts, outside the three selected deep dives.