Our Company
Advisory rooted in how telcos actually work.
We are a small practice that combines telecommunications operations experience with a measured understanding of what AI can, and cannot, do for customer-facing teams.
← Back to HomeOur Story
A practice built from the inside
velvettwwilight came out of a frustration familiar to anyone who has sat in a Malaysian telco operations centre: well-meaning AI tools arriving through procurement, landing in front of care teams, and creating more confusion than clarity — because nobody had thought carefully about where the AI sat relative to the network.
Our founders spent years working within telco customer operations and OSS/BSS environments before stepping back to ask a simpler question: what would genuinely useful AI assistance look like for a care team leader in Petaling Jaya, Ipoh, or Johor Bahru? Not a product pitch. An honest answer.
The result is a small advisory practice with three specific engagements, each designed to stay entirely outside the operational network. We read correspondence, draft replies, summarise ticket threads — and we do it with full attention to MCMC guidelines and PDPA obligations, because those frameworks shape how our clients operate every day.
We work with a small number of clients at a time. That suits our approach: deliberate, documented, and based on a careful reading of what your team actually handles.
Our Mission
To help Malaysian telecommunications teams introduce AI assistance into customer operations with care — keeping decision-making with engineers and care leads, not delegating it to algorithms.
Our Approach
Every engagement starts with reading. We read your correspondence, your ticket threads, your outage notices. We write a brief. We do not arrive with a pre-formed solution looking for a fit.
Our Boundary
AI read-only, always outside OSS/BSS. Your network engineers make every operational call. We support the correspondence layer, not the infrastructure layer.
The Team
The people behind the briefs
Razif Hasan
Principal Adviser
Former senior operations manager at a Klang Valley telco, with twelve years reading customer-facing processes and OSS/BSS boundary conditions before moving into advisory practice.
Nadia Ismail
AI Integration Lead
Specialises in natural language processing for bilingual (EN/BM) customer correspondence. Worked previously on PDPA compliance frameworks for managed service providers.
Khairul Latif
Regulatory Affairs
Tracks MCMC regulatory developments and advises on how new guidance affects AI deployment boundaries in telco customer operations. Based in Petaling Jaya.
Standards & Protocols
How we keep the work clean
PDPA Compliance
Each engagement includes a written data handling policy covering what customer data is read, where it is processed, and how long it is retained — reviewed with your data compliance officer before deployment.
MCMC Awareness
Our work accounts for current MCMC consumer and licensing frameworks. Advisory briefs reference applicable guidelines so your operations team has context alongside findings.
OSS/BSS Separation
AI tools we recommend or deploy have no integration pathway to your operational systems. Architecture separation is documented and confirmed before any pilot begins.
Human Approval at Every Step
AI outputs are drafts, not decisions. No reply is sent and no ticket is closed without a human care team member reviewing and approving it. This is a structural requirement, not a preference.
Written Deliverables Only
Every engagement ends with a structured written brief. Verbal-only findings are not accepted. This gives your leadership team a document they can circulate, challenge, and archive.
Quarterly Output Review
For ongoing engagements, we sample AI outputs each quarter, check boundary adherence, and document findings. This gives your COO a clear picture of where AI assist is working and where adjustments are needed.
Our Expertise
AI advisory built for Malaysian telecommunications
Malaysian telecommunications operations sit within a specific regulatory environment shaped by MCMC licensing conditions, PDPA data obligations, and the practical realities of bilingual customer service across a diverse subscriber base. AI assistance that works in one jurisdiction does not automatically work here.
velvettwwilight focuses specifically on the correspondence and ticket-handling layer — the part of telco operations where customer sentiment is read, complaints are categorised, and outage communications are drafted. This is where AI summarisation and drafting assistance has the most direct value, and also where data handling discipline matters most.
Our advisory work draws on direct familiarity with Malaysian telco customer operations: the volume of repetitive tier-one enquiries, the complexity of network-related complaint triage, the bilingual nature of subscriber communications, and the documentation expectations from MCMC when customer service processes are reviewed.
We do not offer a one-size approach. The Customer Operations Review suits teams that are exploring whether AI assistance is appropriate. The Drafting & Triage Pilot suits teams ready to test a structured deployment. The Quarterly Stewardship suits operations leaders who need ongoing governance of an AI assist already in use. Each is scoped to a specific stage in that journey.
Start Here
A brief conversation costs nothing
If you are thinking about how AI might fit into your customer operations — or wondering whether it should — we are glad to discuss it without any prior commitment. Reach out and let us know a little about your setup.
Arrange a Conversation