Client Experiences
What clients say when the advice was genuinely independent
Honest accounts from Malaysian organisations that have used our placement advisory to make better-informed decisions about AI infrastructure.
Back to Home60+
Engagements completed
4.7
Average client satisfaction
4+
Years of practice
100%
Deliverables on schedule
Client Feedback
From organisations that have been through it
Farouk Khalid
Head of Platform Engineering, Petaling Jaya
"We came in thinking we needed a cloud-first strategy because that's what everyone around us seemed to be doing. After the Placement Options Briefing, we realised our inference latency requirements made on-prem more appropriate for at least two of our workloads. The trade-off sheet was clear enough that I could share it directly with our CTO without any additional explanation."
May 2025
Shanthini Thiagarajan
IT Director, Shah Alam
"The Hybrid Design Workshop gave our team an architecture diagram we could actually use. Previous advisory we'd received had been thorough but too abstract to implement. Lumen Array asked different questions — about data movement costs and operational burden — and the output reflected that. The follow-up note was also more useful than expected."
April 2025
Ahmad Wazir
VP Technology, Kuala Lumpur
"Good advisory. We appreciated that they didn't push us toward cloud just because it was the more obvious path. The engagement was efficient — one session, a clear summary, and we had what we needed to move forward internally. The only thing I'd suggest is a slightly longer pre-session questionnaire so they can go deeper faster."
May 2025
Lim May Yin
Chief Data Officer, Subang Jaya
"As a financial services organisation we had specific PDPA concerns about where training data could sit. Lumen Array were the first advisors we'd spoken to who brought those considerations into the placement framing from the start rather than treating them as a legal afterthought. The Strategy Advisory engagement was worth every ringgit."
April 2025
Rohani Ibrahim
Infrastructure Manager, Cyberjaya
"We used the briefing to help us align on a direction before committing to a larger infrastructure project. It was refreshing to get analysis that actually considered both sides — the trade-off sheet became a reference document we kept coming back to in our internal discussions. Would come back for the workshop when we're ready to design."
May 2025
Zulfikri Yusoff
Senior Architect, Kota Kinabalu
"Based in Sabah, I was a bit hesitant about whether remote advisory would be as thorough as in-person. It was — the virtual sessions were well-structured and the shared diagrams worked well. Their understanding of the connectivity constraints that affect on-prem viability in East Malaysia was also notably better than I expected from a KL-based firm."
April 2025
Case Studies
How the engagement played out
Anonymised accounts of real placement decisions Lumen Array supported.
Challenge
A mid-size financial institution was preparing to deploy its first production AI workloads — a fraud detection model and a customer segmentation pipeline. Leadership had agreed to "cloud-first" as a general policy but the infrastructure team was concerned about latency for the fraud model and data sovereignty for the customer data used in segmentation.
How we helped
We ran a Hybrid Design Workshop that analysed each workload separately. The fraud detection model was evaluated against latency thresholds — cloud round-trip times were insufficient for their transaction processing pipeline. The segmentation workload, by contrast, was batch-oriented and had less sensitive data; cloud was workable. We produced a hybrid blueprint showing a clear workload boundary.
Outcome
The team committed to on-prem for fraud detection and cloud for segmentation — a decision grounded in the blueprint rather than general preference. The strategy also satisfied their compliance team's data residency requirements without requiring a fully on-prem deployment. Internal alignment was reached within two weeks of receiving the deliverable.
"For the first time we had a document that explained our infrastructure decision in a way that made sense to both the technical team and the risk committee."
Challenge
A manufacturing group was evaluating whether to run AI-powered predictive maintenance on a cloud platform or on local edge servers at its Johor facility. The decision was stalling because different parts of the organisation had different assumptions about cost and reliability — there was no shared analytical framework to resolve the disagreement.
How we helped
A single Placement Options Briefing was sufficient. We structured the session around their specific constraints: internet connectivity at the facility, acceptable downtime, data volume, and the operational capacity of the local IT team. The resulting trade-off sheet made the cost and risk assumptions explicit for both options, which gave the group a basis for a decision rather than a continuation of the debate.
Outcome
The group chose on-prem edge servers — a decision that had seemed obvious to some stakeholders but had been blocked by others who assumed cloud was more reliable. The briefing made the reliability case for edge explicit and quantified. The total engagement — from first contact to written summary — was completed in under two weeks.
"We'd been going in circles for months. One focused session got us to a decision everyone could support."
Challenge
A private healthcare group was expanding its AI use across diagnostics support, patient scheduling, and resource planning. Each workload had different data sensitivity profiles and different teams sponsoring them. The organisation needed a coherent placement strategy across all three, not three separate decisions made in isolation.
How we helped
We ran a three-month Placement Strategy Advisory engagement. The first month established placement criteria and assessed each workload. The second month refined the strategy as the diagnostics team raised additional clinical data handling requirements. By month three the strategy document captured the reasoning for each workload placement along with the constraints that would trigger a review of that decision.
Outcome
The group adopted a three-way split: diagnostics support on on-prem infrastructure (due to patient data sensitivity), scheduling on a private cloud environment, and resource planning on a public cloud. The strategy document became a governance reference that each sponsoring team could use independently. It has since been used to evaluate two further AI initiatives without additional advisory.
"The document outlasted the engagement. We're still using it to frame decisions nine months later."
Contact
Speak with us directly
Phone
+60 3-2287 6041Office
Level 7, Menara UOA Bangsar
59000 Kuala Lumpur
Hours
Mon–Fri: 9:00 AM – 6:00 PM
Sat: 10:00 AM – 1:00 PM
Professional Standing
Add your own account to this list
Most clients start with the single-session briefing at RM 530. It is a low-risk way to see whether our approach fits your situation.
Book a Briefing