State of AI Trust in 2026 and Beyond- McKinsey, Bain, Deloitte, BCG, EY

Kakunin is a specialized governance and trust platform for deploying autonomous AI agents in highly regulated industries (primarily finance, healthcare, and public sector). It provides cryptographic identity, real-time behavioral monitoring, automated revocation, immutable auditing, and regulator-ready compliance reporting—positioning itself as the “missing primitive” for governing live AI agents.

### Core Offering and Positioning

Kakunin issues **X.509 certificates** (familiar in finance, backed by AWS KMS) that cryptographically bind an agent’s identity, operator, model hash, and permitted actions/scope (e.g., transaction caps). It then monitors both *actions* (API calls, transactions) and *outputs* (content) in real time, assigns risk scores, and can automatically revoke certificates if risk exceeds thresholds. It delivers immutable audit logs, compliance reports mapped to standards (EU AI Act, NIST AI RMF, MiCA, GDPR, ISO 27001), and a public verification endpoint so third parties or regulators can instantly check an agent’s status without accounts.

Additional features include AgentMail (verifiable email inboxes for agents) and SDKs (TypeScript, Python, REST) for quick integration. It emphasizes “from zero to auditable in one afternoon,” content-aware risk scoring (blocks prohibited outputs per regulations), and high-performance scaling (1,000 events/sec, sub-second revocation/verification).

**Target users**: Compliance officers, CTOs/security engineers, operations/finops teams, API platform leads, and AI startups selling into enterprises. It is explicitly *complementary* to human KYC tools and model governance platforms (e.g., Credo AI), not a replacement. It focuses on runtime governance of *deployed agents* in production workflows (trading, claims processing, payments, customs, reconciliation).

It has early traction with Tier-1 EU banks, fintechs, and platforms, with claims of 12k+ certificates issued and real use cases showing 3x faster execution, zero violations, or dramatically reduced fraud/loss.

### How It Fits the 2026 Research Landscape

The reports (Deloitte *State of AI in the Enterprise*, BCG AI Radar 2026, McKinsey State of AI Trust, EY Global AI Sentiment, and Bain insights) paint a clear picture: **agentic AI is scaling explosively while governance, trust, and responsible oversight lag dangerously behind**. Kakunin directly targets this gap.

**Agentic AI surge vs. governance lag**:

- Deloitte: Only 21% of companies have mature governance models for autonomous agents, despite 23% already using them moderately and 74% expecting to within two years.

- McKinsey: Added “agentic AI governance and controls” as a new maturity dimension—it lags other areas; security/risk is the top barrier to scaling.

- BCG: Nearly all CEOs expect measurable returns from agents in 2026; Trailblazers invest heavily in them.

- Bain: Governance is the central challenge for scaling autonomous agents; very few companies run fully autonomous agents in production today.

- EY: AI is shifting from assistive advice to autonomous authority/actions (16% of people have already used autonomous AI for real transactions); trust and defining limits are critical.

**Kakunin’s fit**: It supplies the exact runtime governance layer these reports say is missing—cryptographic identity + behavioral monitoring + automatic revocation + auditability. It turns the “proof-of-concept trap” (Deloitte) into production-ready, regulator-auditable agents by addressing integration, security, compliance, and monitoring needs that cause pilots to stall.

**Responsible AI / Trust as a business enabler**:

- McKinsey: Organizations with explicit RAI ownership and higher maturity achieve better outcomes; treat trust/governance as a core capability, not just compliance.

- Deloitte: Governance must be built *before* scaling and made everyone’s role; it should enable innovation rather than constrain it.

- EY: Human oversight is essential (70%+); organizations must define boundaries, ensure accountability, and build trust through design and transparency as autonomy increases.

- BCG: Concerns around data privacy, cybersecurity, and lack of control/understanding of AI decisions are widespread.

**Kakunin’s fit**: It maps directly to major frameworks (NIST AI RMF, EU AI Act, etc.), provides human-in-the-loop alerts and revocation, enables transparent public verification, and delivers regulator-ready proof. This helps organizations move from “we need to review your AI governance” (a deal-killer per their testimonials) to defensible, auditable deployment. It supports the human-AI collaboration emphasized across reports—AI handles execution while humans/compliance teams retain oversight and control.

**Scaling, activation, and operational readiness**:

- Deloitte: The gap between access and meaningful activation/usage is the primary barrier; most companies are not yet redesigning work or infrastructure for AI at scale.

- Bain: Budgets are rising and tech works, but ROI lags due to data, process, and change-management issues; 2026 is the year boards demand bottom-line results.

- BCG: Trailblazers succeed by focusing on end-to-end workflows and broad upskilling; many others remain in pilots.

**Kakunin’s fit**: It accelerates the leap from pilots to governed production by making compliance and trust built-in and fast. Features like high-throughput event streaming, cost-overrun prevention (via monitoring), and quick certification directly tackle operational scaling hurdles in regulated environments.

**Sovereign / Regulatory and regional considerations**:

- Deloitte: 77% of companies now factor country of origin into vendor choices; sovereign AI (data residency, localization, compliance) is a boardroom issue.

- McKinsey: Asia-Pacific leads in RAI maturity; regional differences matter.

- EY: “Pioneer” markets adopt autonomous AI faster; varying regulatory readiness creates deployment challenges.

**Kakunin’s fit**: EU data residency (eu-west-1), explicit mapping to EU regulations (EU AI Act, MiCA), and use of globally trusted X.509 standards help organizations navigate sovereign and cross-border requirements. It is particularly well-suited for heavily regulated Western/European contexts highlighted in the reports.

**Physical AI**: Indirect fit at best. Kakunin is a software/runtime governance layer for digital agents. It could extend to agents controlling physical systems (e.g., in logistics or customs use cases mentioned), but the reports’ physical AI discussion (Deloitte’s emphasis on controlled environments, cost of ownership, safety) is not its primary focus.

### Overall Positioning

Kakunin positions itself as **specialized infrastructure for trustworthy agentic AI in regulated industries**—the governance, identity, and compliance “operating system” layer that the 2026 reports identify as the critical missing piece preventing safe, scalable adoption. 

It aligns closely with the “Trailblazer” or high-maturity organizations in BCG/McKinsey that invest early in governance to unlock value, and directly solves Deloitte’s call to “build governance before you scale.” It addresses EY’s question of “who defines its limits” through technical enforcement + cryptographic proof rather than policy alone. It is not a general-purpose AI platform, model builder, or broad MLOps tool—it is narrowly and powerfully focused on making autonomous agents *safe, verifiable, and regulator-acceptable* at production scale.

In the context of the reports, Kakunin sits at the intersection of the biggest 2026 pain points: exploding agentic adoption, lagging governance/trust, regulatory pressure, and the need to move beyond pilots to defensible value in high-stakes environments. It is purpose-built for the regulated-enterprise segment where the reports show both the highest risk *and* the highest potential reward for getting governance right.