CONSULTING · TRACK B · AI ENGINEERING & DELIVERY
The system a readiness diagnostic called for, built and delivered
Bounded-scope delivery of the system a workshop's roadmap or a governance review identified, built by the same engineering team behind the case studies below: machine vision quality control, robotics, IoT integration and applied AI, delivered end to end rather than recommended in a report.
Engagement structure
Scoped delivery, quote-based
Delivered by
Orion Five Engineering
Delivery
Singapore-based
Funding
EDG-eligible scopes assessed at intake, for qualifying Singapore SMEs
Led by
Veronica Loh, Managing Director & Chief Sustainability Officer
Track record
ISO 14064 Lead Auditor, twenty-six years in engineering delivery
Overview
Most AI projects fail before they reach production, and leadership is the usual reason
More than eighty percent of AI projects fail to deliver their intended business value, roughly twice the failure rate of non-AI IT projects, according to a RAND Corporation study built on interviews with sixty-five experienced data scientists and engineers. Eighty-four percent of the industry interviewees who had lived through a failure named leadership-driven decisions, not a technical limitation, as the primary cause: a problem scoped by someone who never had to build it, a pilot approved before anyone checked whether the process behind it could actually be measured.
A separate 2025 MIT study reviewing three hundred public AI deployments found ninety-five percent of generative AI pilots delivered no measurable return on the balance sheet, even where the pilot itself was judged a technical success. Neither figure is an argument against building. It is an argument for the feasibility discipline this track applies before a system is commissioned, the same checklist Veronica Loh built from systems she personally delivered rather than a generic automation framework, so a client finds out whether a process is actually ready before spending the budget to discover it the hard way.
A diagnostic or a governance framework describes what a system should do. This track builds it: the same engineering discipline behind Orion Five Engineering's delivered case studies, applied to the problem a client has already scoped, whether that scoping happened in one of the workshops, in a Governance & Assurance engagement, or on its own.
This track is bounded-scope systems delivery: quality control systems, robotics, IoT integration, and applied AI built end to end. It is not seed-stage software or MVP development for a startup building its first product. A funded AI startup whose need is governance credibility for an enterprise deal or a funding round is better served by the Governance & Assurance track.
Why this track
Every case study on this page ships with my name attached because I was on site for the commissioning, not reading a status report from an account manager afterward. Twenty-six years of engineering delivery taught me that the gap between a system that works in a vendor demo and a system that survives its first month of real production use is usually where a project's actual risk sits, and it is not a gap a report can close from a distance.
This track exists because a readiness diagnostic or a governance framework still leaves an organisation holding a plan, not a running system. I do not publish a rate card for the same reason none of the delivery work behind these case studies was ever a fixed product: a robotics line for an egg distributor and a vision system for a pharmaceutical packaging line do not share a price, only a discipline. What I ask at scoping is the same question every time, whether the client is a government agency or a single SME owner claiming an Enterprise Development Grant: what has to be true about the process before it is worth building a system for it.
Capabilities
The disciplines behind the case studies below
Drawn on individually or together, depending on what the scoped system actually needs.
-
Machine vision & quality control
Inspection systems trained on a client's own production data, with full traceability of every pass and every rejected unit.
-
Robotics integration
Six-axis and mobile robotics scoped against a real production or logistics line, not demonstrated in a lab.
-
IoT & sensor systems
Environment, asset and condition monitoring wired into a dashboard someone actually checks, not a data feed nobody reads.
-
Applied AI & ML pipelines
Models trained and deployed against one bounded operational problem, measured against a baseline set before deployment.
-
EDG & grant scoping
Singapore's Enterprise Development Grant and comparable funding instruments assessed against the delivery scope from the first conversation, not bolted on afterward.
-
Commissioning & handover
Systems handed over running and documented, with the client's own team able to operate it, not a pilot nobody owns after go-live.
Track record
Systems delivered, not proposed
Four of the case studies already published on Veronica Loh's speaker profile, chosen for their AI and applied-engineering relevance.
-
Machine-learning quality control for pill packs
A machine-learning vision system trained on roughly 6,000 samples to inspect pill packs for foreign objects and double or empty pouches, with automatic barcode-based routing of failures and a full traceability log of every pack and rejected pouch.
-
RFID asset and inventory management, with self-service kiosk
An RFID-based asset and inventory system replacing manual paperwork and physical searches, with a self-service kiosk that maps a user's body or item dimensions to the correct rack or shelving location and a real-time dashboard of stock on hold.
“A clear understanding of warehouse operations complexity, translated into a technically coherent, user-centric design. Completed to specification, on schedule, and met every requirement in our scope.”
Programme Lead, Logistics and Public-Sector Client, 2025 -
AI diagnostic and e-commerce platform for scalp longevity
A full diagnose-treat-maintain platform for a scalp longevity brand: an AI-customised diagnostic engine decodes a customer's scalp biological signature, an adaptive treatment system tailors care, and an integrated e-commerce shop with a persistent cart and admin dashboard keeps the relationship going.
-
Robotic carton removal line for egg distribution
An Enterprise Development Grant-funded robotics line that replaced manual carton unboxing for imported egg shipments end to end, lifting operational efficiency by over 1,000 percent and cutting the process's dependence on manual labour.
“Engaged as consultant and solution provider for our egg unpacking automation project, covering gap assessment, mechanisation design, and Enterprise Development Grant proposal support, through to project management of design, development, and commissioning. The objectives of the project were completed and met.”
Director, Poultry Processing Client, 2022
A clear understanding of warehouse operations complexity, translated into a technically coherent, user-centric design.
Programme Lead, Logistics and Public-Sector Client, 2025Who this is for
Bounded scope, on a defined problem
- SME owners with a specific, bounded operational problem, a QC step, an inventory process, a repetitive manual task, where a delivered system pays for itself against a clear before-and-after baseline.
- Enterprises extending a readiness diagnostic or governance framework into a working system: the same engineering team credited in the case studies below, not a subcontracted build.
- Organisations already exploring Singapore's Enterprise Development Grant or comparable funding instruments, where the delivery scope needs to be assessed against eligibility from the outset.
Led by
Veronica Loh
Co-Founder & Head, Operations · Managing Director & Chief Sustainability Officer
Co-founder and Head of Operations at Praxora Lab, and Managing Director and Chief Sustainability Officer of Orion Five Engineering, with twenty-six years managing business operations across logistics, food technology R&D, manufacturing mechanisation and digitalisation in Singapore and the region.
Read the full profile ›From the practice
Recent writing on operations & systems
Plant-floor and engineering detail written from the site walk, not the vendor slide deck this track's own delivery work is measured against.
-
Why the AI layer has to sit outside the safety-rated control loop
A machine's safety function runs on a certified, bounded response time. A model's inference time is neither certified nor bounded. Most of what gets called an AI integration challenge on the plant floor is really that mismatch, misdiagnosed as a data or model problem.
-
Robotics ROI on the SME floor: where it's real, and where it's oversold
A cobot handles the four-thousandth repetition exactly like the first. That is the entire case for buying one, and it stops being the case the moment the task changes more often than the cycle repeats.
-
The Scope 3 number your customer needs is sitting in your OT data
From the 2026 financial year, Singapore's largest listed companies must disclose Scope 3 emissions, and a supplier with no activity-based data to offer becomes the industry-average estimate their customer is stuck reporting.
Common questions
Before you enquire
-
What exactly gets delivered?
A working system, not a proof of concept or a report recommending one. The case studies below describe what that has meant in practice: a machine-learning vision system inspecting pill packs on a live packaging line, a robotic carton removal line, an RFID asset system with a self-service kiosk.
-
Does this include AI strategy or governance work?
Not on its own. This track builds the system a strategy or governance process has already scoped. Where that scoping has not happened yet, it usually starts with the AI Governance & ROI Executive Programme or the Governance & Assurance consulting track, though it can also start directly from a defined operational problem.
-
What does this cost?
Quote-based, scoped to the system once the problem, the data involved, and the constraints are understood. We do not publish a rate card for delivery work, and neither does any comparable firm at this end of the market; the fee is set at scoping, not before it.
-
Are we too small for this?
The case studies below include work delivered for individual SME clients under Singapore's Enterprise Development Grant, not only large enterprise or government engagements. Scale is assessed case by case at scoping.
-
Can you build an MVP for our startup?
No, and we would rather say that plainly than let you find out partway through an engagement. This track's evidence is bounded-scope industrial and enterprise systems delivery, not early-stage software product development. If your actual need is governance credibility for an enterprise buyer's due diligence, the Governance & Assurance track is built for exactly that instead.
-
Who actually does the work?
Veronica Loh and the Orion Five Engineering team credited in the case studies below, not a subcontracted delivery partner.
Investment
Quote-based, scoped to the system
No two engagements on this track look alike, so we do not publish a rate card. The fee is set once the problem, the data involved and the constraints are understood at scoping, and Singapore SME clients should raise Enterprise Development Grant eligibility at that same conversation.
