what why

The gap between these two words is a regulatory liability.

If you are a Head of Compliance, Risk, CCO or CRO at a PRA-regulated firm with AI agents in production, you already carry personal accountability for their decisions under SM&CR. Right now, you cannot independently evidence why those decisions were made. whattowhy is the only platform that produces that evidence continuously, traced to source data, structured for regulators, from a party with no stake in the outcome.

Reasoning trace Live capture
Credit application · Agent: CreditUnderwriter-v2
Application declined
Ref #CU-2024-88421 · 14:32:07 GMT
Data Policy Context Decision
Why, captured at decision moment
"Debt-to-income ratio of 4.8× exceeded SS1/23 §3.4 policy ceiling of 4.2×. Challenger model concurs. Escalation not required. Human review threshold not met."
Tamper-evident FCA-structured Independent SS1/23 aligned
54%

of the world's largest banks are already piloting AI agents in live production workflows.

IIF-EY, 2025
1 in 5

enterprises has a mature governance model for the autonomous AI agents they are deploying.

Deloitte, 2025
0

vendors today offer truly independent evidence of why an AI agent made a consequential decision.

Market analysis, 2026

Watch a decision move through whattowhy.

A credit agent declines an application. Without whattowhy, a log entry. With whattowhy, a complete evidential record — in under a second.

🤖
AI Agent
w→w
whattowhy
🔍
Trace
📊
Score
📑
Evidence
1 of 5

Not governance. Not security. Continuous Assurance.

Tools give you telemetry. Advisors give you frameworks. Neither produces the independent, regulator-legible evidence that proves the right decision was made, under the right policy, at the right moment.

Capability
Market today
whattowhy
Produces evidence, not just telemetry
 Telemetry only
 Regulator-legible
Traces decisions to source data
 Output logs
 Full chain
Structurally independent of agent vendor
 Conflict of interest
 Permanently
Continuous in-life monitoring
 Point-in-time
 Real time
Risk scoring by materiality tier
 Binary alerts
 SS1/23 aligned
Independent of the firm advising you
 Built by advisors
 No advisory stake

An independent manager agent for your AI agents.

Five steps. Each valuable alone. Each making the previous step more powerful over time.

01

Sits outside your agent stack

Integrated at the layer between your agents and the systems they access, not inside the agent itself. No vendor partnership. No internal access. Structurally independent from the system being assessed. The same vendor cannot provide the agent and the assurance of it, ever.

Independent
02

Capture the why at the moment of decision

What information the agent had. What policy it applied. How it interpreted that policy in context. Why it acted as it did. Reasoning traced all the way back to source data. Captured live, not reconstructed after the fact.

Capture
03

Score and produce evidence that holds up

Each decision is scored by risk materiality, exactly as SS1/23 requires. High-risk decisions get full trace and human escalation. The output is tamper-evident, queryable, and structured for FCA, PRA and Consumer Duty — legible to a compliance officer, not just an engineer. Every decision added to the corpus makes the next phase more powerful.

Prove
04

Surface risk before it materialises

As the corpus grows across your agent fleet, whattowhy identifies which agents are drifting toward bad outcomes before they appear in a log. Which decision patterns are accumulating tail risk across hundreds of low-materiality calls that each look fine individually. Which use cases pose a greater risk or have a lower level of control. Not reactive. Predictive.

Predict
05

Turn evidence into better policy

The corpus identifies where policy is ambiguous, where agents consistently misinterpret intent, and where outcomes are diverging from what the policy was designed to produce. These surface as structured policy recommendations — turning operational evidence into continuous governance improvement, not a static rulebook reviewed once a year.

Improve
Firms must assess, test, understand and evidence the outcomes their AI systems deliver to customers."
FCA Consumer Duty, active enforcement, 2024-2026

The only platform that moves from evidence to prediction.

Every phase is valuable on its own. Together they compound into something no point-in-time audit or telemetry dashboard can replicate.

02

We surface risk before it materialises

As the decision corpus grows, whattowhy identifies which agents are drifting toward bad outcomes before they appear in a log. Which decision patterns are accumulating tail risk across hundreds of calls that each look fine individually. Which use cases have a lower level of control than your risk appetite allows. Not reactive monitoring. Predictive risk intelligence, answering the question regulators are starting to ask: do you know where your risk is concentrating before something goes wrong?

Predictive risk
03

We are structurally independent

whattowhy sits outside the agent stack, with no vendor partnerships and no internal access. The same entity cannot produce the AI agent and the independent assurance of it. Architecturally identical to external audit. Any firm that also provides the agents, the infrastructure, or the advisory wrapper has a conflict of interest that no policy can resolve.

Structural independence
04

The corpus turns evidence into better policy

Over time the decision corpus identifies where policy is ambiguous, where agents consistently misinterpret intent, and where outcomes are diverging from what the policy was designed to produce. These surface as structured recommendations, turning operational evidence into continuous governance improvement rather than a static rulebook reviewed once a year.

Policy intelligence
Design partner conversations

We are having 10 conversations
with Heads of Compliance,
Risk and CROs.

If you lead compliance, risk or AI governance at a PRA-regulated firm and have agents in production, we want to speak with you. No pitch. We will show you a single decision, fully traced, and ask whether you can get that from anything you have today.