Engineering delivery / Human–AI work design
A decision is only the beginning.
I lead engineering delivery in fintech, and I am building a specialisation in deciding what should stay with people when AI takes part.
When AI began drafting quality scores for customer calls, no score could count until a reviewer had checked it. Internal reporting showed evaluation time still fell by 75%.
Selected work
Three decisions, and what they changed.
Human–AI workflow
Making AI-assisted quality review faster without removing human judgement
The decision: AI drafts every score. Nothing counts until a reviewer has checked it.
Financial services · Delivery coordination
Reported · 75% less evaluation time · ≈R200k a year savedWorkflow automation
When professional verification becomes an automation problem
The decision: Any check that failed or was uncertain went to a person; the automation never approved or rejected one.
AEMI · Independent engagement
2024–2025Fintech delivery
Coordinating integrations across a multi-partner payments platform
The decision: No partner was committed to a release until its readiness and dependencies were explicit.
Yoyo · Delivery Manager
2022–2023
Reported · 15+ partners · 8 releases
One decision, start to finish
From a request to a result.
From case 01 · Reported figures
How a request for faster reviews became a design where AI does the preparation and people keep the judgement.
The request
Faster quality reviews.
Reviewers were spending most of each evaluation transcribing and structuring a call before they could judge it.
The diagnosis
The slow part was preparation.
And the scores fed into how individual staff were assessed, so speed could not cost explainability or a person’s right to challenge a score.
The decision
AI drafts. A reviewer decides what counts.
AI produces the transcript and a draft of every score. A reviewer can change any score, and nothing counts until a reviewer has checked it.
The result
75% less time per evaluation
and about R200k a year in cost saving.
Figures from internal reporting, confirmed by me for publication. The baseline and measurement period are not published.
Reported · internal reporting
Now · Since March 2026
Engineering Delivery Lead at Lula.
I coordinate delivery across engineering and technology support at Lula, a fintech in Cape Town: planning and prioritisation, dependencies and delivery risk, and keeping engineering and business priorities aligned.
Before that: business operations at Lula, payments and commerce delivery at Teya and Yoyo, and earlier, business analysis and Scrum roles. I write Monorati, an independent publication on the decisions behind modern work.
About meLinkedIn (opens in a new tab)Monorati (external site)
Let’s talk about the work.
Open to senior roles and collaboration in delivery, especially where AI is part of the work.
Get in touch