Aivance · Singapore
Independent AI governance reviews for regulated institutions
For each AI system, including generative AI and agents, Aivance checks what it can actually do when it runs and whether the limits your risk rating relies on are enforced. Founded and run by Arjen Hendrikse in Singapore.
Where to start
AIRG Use-Case Baseline
A fixed-price review of one AI use case against the MAS AI Risk Management Guidelines. For this use case, what would you show MAS, your board or internal audit today, and what is missing?
One use case · 10 working days · S$4,800 fixed
The call is to choose the use case and confirm it fits.
What you receive
- An inventory entry in a format you can reuse across the firm.
- A risk materiality rating with the reasoning written out, ready for your control function to challenge.
- A verdict on each limit the rating relies on: proven, assumed or unbounded.
- A third-party evidence check where a vendor is involved: what is self-attested and what is independently assessed.
- A prioritised gap list and a two-page management note.
What the MAS AIRG are
The Monetary Authority of Singapore (MAS) is Singapore's central bank and financial regulator. On 7 October 2026 it issued its Guidelines on Artificial Intelligence Risk Management, known as the AIRG. They apply to the financial institutions MAS regulates, including banks, insurers, fund managers, financial advisers and payment firms.
Under the Guidelines, each firm should identify where it uses AI, keep an inventory, assess the risk materiality of each use case, and apply controls in proportion to that assessment. They are supervisory expectations, worded as "should". AI vendors selling to these firms can expect to be asked for evidence against them.
- 7 October 2026: MAS issued the final Guidelines.
- 7 October 2027: Expectations on oversight, identification, inventory and risk materiality assessment apply.
- 7 October 2028: Life cycle expectations apply.
Limits: this is my independent reading of one use case against the Guidelines, within the agreed scope and evidence access. It carries no legal advice, certification or compliance sign-off. Risk acceptance and deployment approval remain with the institution. One invoice on delivery. The fee is credited if you order a firm-wide baseline within 60 days.
Outside Singapore? The same review can be run against ISO/IEC 42001 or your own AI policy. Email contact@aivance.com.sg.
Other work
Is this your situation?
When to bring in Aivance.
-
Unclear delegated authority
Your agents have access, but their delegated authority is unclear.
-
Untested controls
Your runtime controls exist, but nobody has tested them against real authority failures.
-
Decision traceability
Your organisation cannot reconstruct who approved an exception, under which rule, and with what evidence.
What Aivance assesses
From delegated authority to observable control
Aivance begins with the business authority that must hold, then tests whether the deployed technical controls enforce it.
- 01
Consequential action
What can the agent actually do?
- 02
Authority rule
Who or what is permitted to approve, deny, or escalate that action?
- 03
Configured control
Which runtime, identity, workflow, or application control should enforce the rule?
- 04
Adversarial test
Does the control hold when the action is attempted through normal and misuse paths?
- 05
Decision evidence
Can the organisation reconstruct the action, authority, approval, policy/configuration version, outcome, and retained evidence?
Three- to four-week assessment
Agent Authority & Control Assessment
A three- to four-week independent assessment of the authority rules, configured controls, exception paths, and decision evidence around consequential agent actions.
Explore the assessmentEngagement sequence
Assess → Prioritise → Design → Implement with internal teams or specialist partners → Retest
Aivance can define remediation requirements and validate remediation evidence. Implementation is performed by the client’s internal teams or selected specialist partners unless explicitly agreed otherwise.
Singapore and Southeast Asia
Independent assessment for teams accountable for consequential agents.
Founder-led work for CROs, CISOs, internal audit leaders, AI programme leaders, enterprise architects, compliance leaders, technology executives, SI partners, and AI vendors selling into regulated enterprises.