Forward-deployed AI consulting in Singapore
Singapore is where many banks and enterprises have pilots that work in a demonstration and stall before production. The stall is usually about ownership and evidence, not about the model: who runs the agent after launch, what the risk team can inspect, and how the system is tested when it changes.
I am Mayur Patil, based in Nashik, India. I work with Singapore teams on forward-deployed AI consulting, and I am in Singapore often for engagements.
Why AI pilots stall in Singapore
Three reasons come up most often. The first is that the risk team was not part of the design, so approval arrives after the build. The second is that the evidence is a demonstration rather than a test set that the business owner has agreed. The third is that the agent can act, but no one has defined when a person must approve or stop an action.
The IMDA framework addresses all three. It is organised around four dimensions: assess and bound the risks upfront, make humans meaningfully accountable, implement technical controls and processes, and enable end-user responsibility. It is voluntary, and it is a living document, so the version and date should be quoted in any submission. The current version is 1.5, published on 20 May 2026.
What the rules mean for the build
Singapore's rules set what an agent may do with personal data and what a risk team will inspect. The Personal Data Protection Act applies to personal data in an agent's inputs, outputs and logs. The PDPC's advisory guidelines on AI recommendation and decision systems say that a service provider who builds a bespoke system can be a data intermediary, which changes who holds the protection and retention duties. Each of these is settled in the design, and the Singapore pillar summarises the framework and its sources.
Industries I work with
I work in banking, fintech, consulting and IT services, manufacturing and capability centres. The Singapore engagements I can describe are in banking. Crédit Agricole CIB, Singapore, is listed among the training engagements on the AIXL Academy showcase, and the work there covered web solutions with AI built in, prompt-based working methods, and upskilling for the teams that use them.
How an engagement runs in Singapore
An engagement starts with a two-week readiness review on one workflow. It ends with a written recommendation to proceed, reframe or pause, and the recommendation may be to stop. If the work proceeds, an architecture sprint or a forward-deployed delivery engagement follows, with the evaluation set agreed by the business owner and a handover to a named owner.
Singapore is 2 hours and 30 minutes ahead of India time. Decisions can be scheduled for the Singapore afternoon, and the build can run from India in the Indian morning. I run back-to-back full-day corporate trainings, mostly in IST, and some programmes in US Pacific time for teams in the Bay Area.
Three ways to start
AI Delivery Readiness Review. Two weeks on one workflow, with a written readout and a recommendation to proceed, reframe or pause.
Architecture Sprint. Two to four weeks on one agreed use case. It produces the architecture, the recorded decisions, the evaluation plan and a backlog.
Forward-deployed AI delivery. One bounded use case, built inside your systems, with a runbook and a named owner at handover.
Each is quoted on its scope. The engagements page describes what each includes. I do not publish prices.
Relevant work
The organisations I have worked with and can name are listed on the about page. Among them are Tredence, Persistent and Walmart, and I have worked with Wipro and Infosys on capability programmes.
Where to go next
The FDE consulting hub covers the three ways to engage in more detail. For teams that want to learn the method first, the FDE training for Singapore page describes the options. For the rules, read the Singapore pillar. For the method, read the four gates.
Common questions
Is the IMDA agentic AI framework mandatory for our agent?
No. The IMDA Model AI Governance Framework for Agentic AI is voluntary. It sets four dimensions that a risk team in Singapore will expect to see addressed, and the design work follows them.
Are you a data intermediary under the PDPA when you build a system for us?
The PDPC advisory guidelines on AI recommendation and decision systems say that service providers who build bespoke AI systems can be data intermediaries, with protection and retention obligations. The engagement records which party holds each obligation.
We are a bank. Do the MAS AI risk management guidelines cover agents?
The MAS guidelines are final and apply to financial institutions with phased dates from October 2027. Whether a particular agent falls under a particular section is answered in the readiness review, against the guidelines as issued.
Can the work run from India with on-site discovery in Singapore?
Yes. I am in Singapore often for engagements. Discovery and the handover can be on site, and the build can run from India with a fixed overlap for decisions.
Sources
- IMDA, Model AI Governance Framework for Agentic AI (launch, 22 January 2026)
- IMDA, Model AI Governance Framework for Agentic AI, version 1.5 (20 May 2026)
- PDPC, Advisory Guidelines on the Use of Personal Data in AI Recommendation and Decision Systems (1 March 2024)
- MAS, Guidelines on Artificial Intelligence Risk Management (7 October 2026)
- Crédit Agricole CIB, Singapore: AIXL Academy showcase