AI Automation & Operations Leader · Telehealth · Systems & Process Design

I know the operational problems you're hiring someone to automate — because I've spent the last 2.5 years inside them.

Medicinal-cannabis telehealth operator. Systems thinker. Independent AI builder. I identify messy manual work, determine where AI adds value, build the workflow, test it, measure it and iterate.

Explore AI builds Why me

30,000+

monthly patient consults

~200

clinicians: ~120 doctors and 80 nurses

24h → <3h

patient rebooking turnaround

independent AI builds

Why me

Operations first. AI where it earns its place.

My career has been built around finding operational friction: unnecessary hand-offs, capacity constraints, fragmented information, slow processes and work being done manually because nobody has redesigned the system.

In medicinal-cannabis telehealth, I've worked across patient support, doctors, nurses, scheduling, capacity and clinical operations. Alongside that work, I've independently built AI products that turn messy human inputs into structured, useful outcomes.

That intersection — operational discovery + AI-enabled delivery — is where I do my best work.

AI builds

Don't just read about them. Try them.

Primary build

Bespokeyo

AI Process Discovery & Automated Quoting

Independently designed and built an AI-powered application that converts an unstructured description of a user's operational requirements into a structured process.

The application captures requirements, uses AI to interpret and organise the information, develops a process map and applies predetermined business criteria to determine the next action. Depending on the outcome, the system can provide process documentation or generate a quote for implementation.

I designed the product, workflow architecture, AI behaviour, business rules, prompting, back-office administration, automated email sequence, testing and iterative refinement using Lovable.

At code level, I personally modified the TypeScript schema validation, connected the project to GitHub and wrote four automated Bun tests with seven passing assertions.

Try the live build

How it works

Capture the real operational requirement in the user's own words.

Bespokeyo case study

Behind the build

From an operational problem to a working AI product.

“Bespokeyo didn't start with ‘what can I build with AI?’ It started with an operational problem. The technology came second.”

Stage 01

Find the real problem

Start with the work, not the technology.

Businesses often know something in their operation is inefficient, repetitive or frustrating, but struggle to articulate the underlying workflow clearly enough to determine what should actually change.

Bespokeyo therefore begins by allowing the user to describe what they need in their own words rather than requiring them to understand process mapping, automation architecture or technical terminology.

Messy operational problem
Understand the work first

1 / 7

What this build demonstrates

Operational discovery
Finding the actual problem before choosing the technology.
AI workflow design
Giving AI a defined role inside a broader operational process.
Business logic
Connecting AI outputs to rules and downstream actions.
Prompt development
Shaping AI behaviour through iterative prompting and scenario testing.
End-to-end thinking
Designing the customer experience, AI behaviour, back-office process and communications together.
Iteration
Testing actual behaviour and refining the product when it doesn't perform as intended.
Deployment
Taking the idea through to a functioning live application.

Claireity Health

Women's Health Advocacy Platform

Designed to help women capture health symptoms and relevant information longitudinally rather than trying to reconstruct everything during a short medical consultation.

The application organises this information and allows the user to generate a structured, clinician-friendly advocacy report to support a better-prepared medical conversation.

The product is designed to support the human clinical interaction rather than replace clinical judgement.

Independently developed through AI-assisted product design, workflow development, prompting, testing and iterative refinement.

Explore Claireity Health

Original AI & workforce design framework

Don't automate the role. Deconstruct the work.

Workforce Intelligence is an AI workforce methodology I developed independently, in my own time, informed by firsthand operational experience in medicinal-cannabis telehealth.

“The question isn't simply ‘Can AI do this job?’

Most roles contain a mixture of repetitive processing, human judgement and emotional interaction. Workforce Intelligence breaks work down to task level so those components can be treated differently.”

The objective: find the work AI should take on — and protect the work humans should own.

The ACE model

AArtificial

Work technology can perform, accelerate or augment.

auto-triaging tickets
predictive response suggestions
sentiment detection
knowledge surfacing
voice/chat transcription
pattern recognition

CCognitive

Work requiring interpretation, judgement and contextual decision-making.

interpreting the patient's actual need
managing multi-step or unusual processes
adapting to compliance requirements
balancing competing priorities
handling exceptions

EEmotional

Work where human connection materially affects the outcome.

empathetic reassurance
tone calibration
de-escalation
explaining sensitive information
providing a personal human response

These aren't mutually exclusive — a single workflow can contain elements of all three. The aim is the right combination.

The people principle

AI transformation should happen with people, not to them.

“One of the reasons I designed ACE was psychological safety.

When AI arrives as a top-down initiative, people can understandably fill the information gap themselves: What does this mean for my job? Is the objective to replace me? What happens to the work I do?

ACE creates a different conversation.

Employees can be involved in breaking down the work they actually perform and identifying which parts are repetitive or technology-suitable, which require human judgement, and which depend on empathy, trust and connection.

That makes the workforce part of the redesign rather than the subject of it.”

AI done TO the workforce

Uncertainty
Assumptions
Anxiety
Resistance

AI designed WITH the workforce

Visibility
Participation
Understanding
Better change adoption

The design intent and change-management philosophy behind ACE — not a guaranteed outcome.

“The people doing the work often know better than anyone where the friction is. They shouldn't just be told what AI will change — they should help identify what should change.”

A Artificial

What are we doing manually that technology could reasonably take on?

C Cognitive

Where does our knowledge, judgement and contextual understanding create value?

E Emotional

Where does being human materially affect the outcome?

ACE makes AI opportunity discovery a workforce conversation, not simply a technology assessment.

The method

From a role to an AI opportunity

Stage 01 of 08

Define the outcome

What is this role or workflow actually trying to achieve?

Interactive example

See ACE in action

A simplified medicinal-cannabis patient-support example

Illustrative scenario — not a description of Dispensed's workflow

A patient contacts support because their prescription or medication process has been delayed. They are frustrated and want to know what is happening.

Task 1 of 5

Detect the issue and urgency

A Artificial

AI opportunity

Analyse the incoming message, identify the issue category, detect sentiment/urgency and route appropriately.

Human role

Review exceptions or high-risk situations.

The opportunity isn't “replace patient support with AI.”

It's redesigning the workflow so AI handles more of the searching, sorting, detecting and drafting — while people spend more time on judgement, exceptions and human interaction.

Automation discovery

From task analysis to an AI agent

“Rather than starting with a predetermined AI solution, Workforce Intelligence identifies clusters of work where automation could create measurable operational value.”

Individual tasks

sentiment detectionissue classificationinformation retrievalresponse draftingpattern detection

Potential AI capability

Operational workflow / AI agent opportunity

Workforce design

ACE taxonomy
Task database
AI substitution / augmentation assessment
Workload composition
Workforce mix forecast
Strategic workforce planning

“Automation changes more than a process. It changes what work remains, what skills become more valuable and how workforce requirements evolve.”

For the role

Why this matters for AI Automation

“The hardest part of operational AI isn't finding something that AI can technically do. It's finding the right work to change.

Workforce Intelligence gives me a repeatable way to enter an unfamiliar workflow, understand what people are actually doing, involve the people doing it, establish where time is being consumed, identify suitable AI opportunities and redesign the surrounding operation.”

Understand the work
Involve the people doing it
Establish the baseline
Identify AI opportunities
Build and evaluate
Redesign the work

“For me, successful AI transformation isn't just an automation that works technically. It needs to fit the operation, create measurable value and be understood by the people whose work it changes.”

AI isn't the starting point. The work is.

First 90 Days

I wouldn't arrive with the answer. I'd arrive with a method for finding it.

“The fastest way to build the wrong automation is to decide what people need before understanding how the work actually happens.

My first 90 days would start close to the operation: observing the work, talking to the people doing it, establishing the baseline and finding the friction worth solving.”

The goal for Month 1: identify one high-value problem and get something useful into the hands of the team quickly.

Patient Support

Find the first problem worth solving.

Patient support is the first operational area identified in the role, so that's where I'd begin.

Step 1 of 8

Listen

Spend time with the people doing the work.

Understand

what patients contact the team about
what staff repeatedly search for
where information lives
what requires copying or re-keying
where hand-offs occur
what causes queues or rework
what employees find frustrating
which exceptions require judgement

The People Principle in practice: the people doing the work help identify the friction, rather than having automation imposed on them.

What I'd want by the end of Month 1

Intended objectives, subject to operational access, governance and technical constraints.

A mapped operation

Clear understanding of the workflow, friction and major manual activities.

An opportunity backlog

Prioritised AI/automation opportunities based on operational value, feasibility and risk.

A baseline

Evidence of current performance against which change can be measured.

One focused build underway or in users' hands

A deliberately narrow first use case being tested with the operation.

Operator feedback

Direct feedback from the people whose workflow is changing.

Build for six months from now

The finish line isn't deployment.

“An automation that works in a demo but becomes unusable, untrusted or dependent on its creator isn't operational transformation.

I would design for the point where the automation becomes part of normal work.”

Fit

Does it fit naturally into how the team works?

Measure

Is it still creating the value it was built to create?

Document

Can someone other than the builder understand how it works and what to do when it doesn't?

Hand over

Can the operation own the stable workflow without needing the Head of AI involved every day?

Discover→Measure→Build→Evaluate→Embed→Hand over→Find the next problem

That's the operating model I'd bring to AI Automation.

Experience

I've been redesigning operations long before I started building with AI.

AI is a new tool in my toolkit. Operational transformation isn't.

“My route into AI automation isn't a traditional engineering path. It comes from more than 10 years working inside complex operations — understanding demand, finding friction, redesigning workflows, building systems and measuring whether the change actually worked.

Over the last approximately 2.5 years, I've done that inside medicinal-cannabis telehealth. More recently, I've begun combining that operational experience with hands-on AI application building.”

01 · Problem
Patient rebooking involved a fragmented six-step process and turnaround could take approximately 24 hours.
02 · Intervention
Bibi mapped and redesigned the workflow, reducing it from six steps to three and creating clearer ownership and faster movement through the process.
03 · Result
Rebooking turnaround reduced from approximately 24 hours to under three hours within the first three months.

What it demonstrates

Process discovery
Workflow redesign
Removing unnecessary work
Measurable operational improvement
Patient experience

Before automating a process, make sure the process deserves to be automated.

A pattern, not a one-off

The tools change. The way I approach problems doesn't.

01
CIBC
Workforce planning in a large contact-centre environment handling more than one million calls per month.
02
Sitel
Workforce management experience in an operation of approximately 500 agents.
03
Baby Bunting
Process improvement and operational response work, including Bibi's contribution to the khapra beetle biosecurity response, which received external recognition from Australian biosecurity authorities.
04
Sportsbet
Senior workforce planning across forecasting, rostering, adherence and operational visibility.
05
Medicinal-cannabis telehealth
Workforce strategy, clinical capacity, patient operations, workflow redesign and operational systems.
06
AI application building
Independently designing, testing and deploying AI-assisted products including Bespokeyo and Claireity Health.

What changed was the technology.

01
Forecasting
Understand demand.
02
Process improvement
Understand how work moves.
03
Workforce design
Understand who should do the work.
04
Workforce Intelligence / ACE
Understand which parts technology and humans should own.
Now
AI Automation
Build technology into the work.

“AI didn't teach me how to find operational problems. It gave me a new way to solve them.”

Interactive application

You've seen the evidence. Now interrogate it.

Ask Bibi's AI

“Ask about my experience, the applications I've built, Workforce Intelligence / ACE, measurable operational outcomes or how I'd approach AI automation at Dispensed.”

This assistant is designed to answer from the evidence in this portfolio and its approved knowledge base — not invent an answer when the evidence isn't there.

Grounded in Bibi's approved portfolio evidence