pl8ypus

Greg Staunton / Lead AI Builder

Building governed AI systems for marketing operations.

pl8ypus is my AI builder portfolio - focused on translation AI, marketing intelligence, campaign operations, and practical AI workflows that can be reviewed, trusted, and scaled.

View active builds

02

Active systems

LIVE

Build mode

OPEN

Speaking track

Build Status

Fast prototypes. Enterprise delivery discipline.

Every build is framed around workflow value, governance, review states, failure paths, and evidence that senior stakeholders can trust.

Mode

BUILDING

Frontend

HTML / Tailwind / JS

Source

GitHub

Output

DEMOS

Enterprise Controls

The model is only one part of the system.

The durable value sits around the model: source control, protected terminology, review gates, data quality checks, audit records, fallback paths, and clear ownership.

Review states

Human approval

Human approval remains visible before outputs are treated as final.

Audit trail

Traceable decisions

Inputs, outputs, prompts, rules, and decisions can be traced.

Rollback path

Recoverable versions

Working versions and approved versions are separated so recovery is possible.

Data quality

Validation first

Bad or incomplete inputs are surfaced before AI summaries or outputs are trusted.

Current Build Evidence

Proof that the systems are moving.

The site now tracks active build state, architecture direction, and next evidence steps. The goal is not to claim finished products early - it is to show visible, controlled progress as the systems move from architecture to working demos.

Click to enlarge

Translation AI

Architecture live
  • Governance model documented
  • Translation workflow page published
  • Controls defined for QA, review, rollback, and protected terms

Next evidence: upload-to-translation demo workflow

Click to enlarge

Marketing Intelligence AI

Option A selected
  • Four-part intelligence model published
  • Three architecture options documented
  • Data quality, ingestion, and review gates defined

Next evidence: dashboard screenshots and ingestion demo

Click to enlarge

Builder Log

Weekly record started
  • Build decisions captured
  • Delivery notes organised by progress, controls, and next steps
  • Public proof-of-work structure in place

Next evidence: first proper build note

AI Builder Stack

Lean build stack. Serious workflow architecture.

The stack is intentionally lightweight where speed matters, but the architecture is not casual: version control, server-side orchestration, review gates, source validation, and deployment discipline are part of the build.

AI Models

OpenAI Anthropic DeepSeek

Model choice depends on the task: orchestration, translation quality, reasoning, structured output, review support, and cost control.

Build Layer

HTML Tailwind CSS Vanilla JS Cloudflare GitHub

Fast, inspectable, portable builds. No unnecessary framework weight where a clean static or worker-based architecture can prove the system faster.

Marketing Systems

Eloqua Salesforce Campaign Ops Reporting

The niche is not generic AI. The focus is AI for marketing automation, campaign operations, asset workflows, reporting, stakeholder visibility, and operational scale.

Active AI Systems

Two active systems. Built around real operational problems.

Build 01

Active Build

Translation AI

AI orchestration for structured marketing asset translation, GitHub versioning, review workflows, and output governance.

Operational value

Reduces manual copy and paste translation work, creates governed review states, protects approved source content, and gives teams a repeatable workflow instead of isolated AI experiments.

Open Translation AI Builds overview

Build 02

Active Build

Marketing Intelligence AI

AI-assisted dashboard and ingestion layer for campaign intelligence, competitor signals, channel analysis, and executive readouts.

Decision value

Turns scattered campaign, competitor, channel, and performance inputs into a stable intelligence layer for decisions, reporting, prioritisation, and strategic visibility.

Open Intelligence AI Builds overview
Open full builds page

Builder Operating Model

Prototype. Test. Govern. Ship.

01

Frame

Define the workflow, user, input, output, risk, and decision point.

02

Build

Create the smallest useful system that proves the workflow.

03

Govern

Add review states, audit trails, rollback, QA, and failure paths.

04

Ship

Publish the demo, document the trade-offs, and capture what changed.

Speaking Track Open

Live demos with the architecture underneath.

pl8ypus is shaped for practical AI talks, build-led demos, and senior conversations about what it takes to move AI from experiments into governed operating workflows.

Open speaking page

Talk themes

> From prompt experiments to governed workflows

> Building AI systems for marketing automation and campaign operations

> Why AI builders need demos, not slideware

> Enterprise AI reality: workflows, trust, review, reporting, and rollback

Builder Log

Weekly notes from the build.

The log will track what shipped, what broke, what changed, and what the next version needs. This is where the build journey becomes visible for both technical readers and senior stakeholders.

Open builder log

Contact

Discuss a build, demo, or speaking session.

For demos, AI builder events, speaking opportunities, or collaboration around practical AI systems for marketing automation and intelligence workflows.