The archive · AI & Models · Strategic decision · 2025–2026
kAIgentic bets governed operations, not models, win enterprise AI; SMBC invests $10M
Ex-Microsoft APAC president Ahmed Mazhari launched Singapore agentic-AI startup kAIgentic in Aug 2025; SMBC Group put in $10M and took a majority stake.
kAIgentic
What the business is
Agentic AI deployment infrastructure that captures an enterprise's unwritten operational logic and turns it into supervised, governed AI agents for banking and other regulated industries.
Starting capital：US$10M strategic investment from SMBC Group; SMFG agreed to acquire a majority stake (Nov 2025)
How it started
Ahmed Mazhari, former president of Microsoft Asia Pacific and already SMBC Group's AI Transformation Advisor, incorporated kAIgentic in Singapore in August 2025 to attack the gap between AI pilots and governed production deployment.
What happened
SMBC Group committed US$10M and agreed to acquire a majority stake, announced in November 2025. The company runs a 'dual-engine' structure: one engine embeds kAIgentic software inside SMBC's banking operations as customer zero, while an independent platform company with its own board builds generalisable products for other regulated enterprises, starting with healthcare, retail, and telecoms.
How it ended up
Still live and scaling; proving the platform inside SMBC banking operations while hiring engineering teams in India for broader enterprise rollout.
Background
kAIgentic is a Singapore-based agentic AI startup founded in August 2025 by Ahmed Mazhari, former president of Microsoft Asia Pacific. Its bet is that enterprise AI is won at the application layer — capturing how work actually happens inside a company and converting that unwritten operational logic into domain-specific AI agents that run under continuous human supervision and governance.
The launch came with an unusual structure: SMBC Group, one of Japan's largest financial groups, invested US$10 million and agreed in November 2025 to acquire a majority stake. kAIgentic's first deployment is inside SMBC's own banking workflows as 'customer zero', giving the platform a live, heavily regulated environment to prove itself in before it sells to other enterprises.
Mazhari's thesis is that model access is no longer the bottleneck in enterprise AI; the bottleneck is whether large organisations can absorb intelligence into critical operations at speed while maintaining their standards. The company runs a dual-engine structure — an internal engine embedded in SMBC's operations and an independent product platform with its own board — and plans to expand into healthcare, retail, consumer goods, and telecommunications after banking.
What has to be true
- Most enterprise AI vendors sell models or point tools; kAIgentic sells governed deployment, which is what regulated buyers actually stall on.
- Deploying inside SMBC first gives the platform a real stress test with audit-grade requirements from day one.
- A majority owner as anchor customer removes the classic startup cold-start problem of no reference customers.
- The risk is structural: being majority-owned by its first customer can limit how independent the platform looks to other banks.
What can be applied
Anchor a regulated institution as first customer to de-risk AI adoption — but selling that trust to other banks depends on proving governance, not model quality.
Aftermath
As of June 2026, kAIgentic is operating with SMBC Group as its first customer, embedding its agents in live banking workflows under human supervision. It is headquartered in Singapore with engineering operations in India, hiring for production-grade AI development, and positioning itself to expand beyond finance into other regulated sectors. SMFG's disclosed agreement to acquire a majority stake dates to November 2025.
Sources
- Ex Microsoft APAC President Launches kAIgentic With $10 Mn Funding To Build AI For Global Enterprises
- Former Microsoft APAC President Launches AI Startup kAIgentic With $10 million From SMBC Group
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card