27 topics. Three decisive. Now visible.
Your client brings 27 topics in 60 minutes. You listen, you structure in your head — and at the same time you know that something important is getting lost between the lines.
DIN 77230 creates structure — but captures no half-sentences, no nuances, no ambivalence. The emotional subtext of a conversation doesn't fit in a checkbox.
AI-assisted analysis is a powerful tool. But when the result carries your name, you need to be able to trace it fully — step by step.
Three years later your client asks: "Why did we decide that back then?" A good answer needs an artifact — not just a memory.
You don't want to decide for your client. You want to show them what their priorities and tensions really look like — and empower them to judge for themselves.
From the half-sentence to the externalized mental model.
From the idea to the documented decision.
Want to experience it yourself?
[ TRY STEERABLE → ]The CFP conducts a normal first session — no change in behavior, no predefined structure.
Audio or free text. The client notices nothing. Steerable extracts in the background.
Priorities, tensions, and constraints are identified from the narrative.
No algorithm decides — the system structures. Pillar nodes and edges with confidence scores.
The client calibrates their own model. Every confidence adjustable, every assumption visible.
"From 0.6 to 0.9." → Human Override, documented. The system makes structure visible. No preset decides for your client.
Zero Baseline + assumptions + confidence + evaluated alternatives. Append-only.
In 3 years: "Why did we decide that back then?" The answer is in the Packet.
Conversation running —
structure not yet visible
27 topics → P1 confirmed · P2 P3 emerging · P4 P5 still open
Client: "Weight equities higher."
Human Override, documented.
"Why did we decide that back then?"
The answer is in the Packet.
Steerable runs in the background. Audio or free text. No change in behavior needed — the conversation stays natural.
The client notices nothing. The advisor works as usual. Steerable listens.
Priorities, tensions, and constraints are identified from the narrative. No algorithm decides — the system structures.
27 topics in 60 minutes → 5–7 Pillars with weighted connections.
Every weight is contestable. Every confidence adjustable. The system doesn't suggest what to change — it makes the structure visible.
"The confidence for equities needs to be higher — from 0.6 to 0.9." → Human Override, documented.
Zero Baseline + assumptions + confidence + evaluated alternatives. Append-only. Forensically usable.
In 3 years the client asks: "Why did we decide that back then?" The answer is in the Packet.
No report. No summary. A verifiable decision artifact.
Your client speaks. The system extracts priorities, tensions, constraints — and makes them visible as a network.
Every weight is contestable. The client remains the author of their decision — the system opens the space, doesn't close it.
A verifiable artifact. Who decided what, when, with what confidence, based on what assumptions. The missing record.
Five dimensions of a Decision Packet — visualized live.
Three real CFP scenarios — editable directly in the browser.
Independent financial planner in the DACH region. Conducts deep first sessions with high-net-worth private clients. Needs structure — not another tool that decides for them.
“My client decided. I showed the structure. The system documented.”
Decision Governance is not a marketing category. It is a documented response to MiFID II Article 25, the EU AI Act (Regulation EU 2024/1689), and the Inverse Governance Paradox in AI-assisted financial advice. Steerable is grounded in structural patterns — not normative recommendations.
“Investment firms providing investment advice shall, before recommending a suitable financial service to clients, provide information demonstrating why the recommendation given corresponds to the needs, characteristics and objectives of the client.”
“High-risk AI systems in financial services must be designed so that decision processes are recorded, human oversight is ensured, and outcomes remain traceable.”
“Better AI does not lower governance risk — it raises it. The less the advisor overrides, the less there is to document. Steerable addresses the shrinking documentation surface, not AI quality.”
· Steerable structures decisions — never recommends.