In December, I shared an example of a structured private banking prompt.
The idea was simple:
Don’t treat advisory as free-form interaction, but as structured logic.
Define the parameters – base currency, risk appetite, investment horizon, allocation ranges – and embed constraints and guardrails.
I started with a structured prompt for Private Banking workflows.
A few hours later, it was a working application.No classical development cycle.
No backlog. No sprint.
From concept to application
What started as a small experiment quickly turned into something more structured.
Using recent experience with vibe coding, the initial prompt was iteratively translated into a working tool – not through traditional development steps, but through continuous refinement of logic, structure, and output.
Instead of writing specifications, you iterate with the system:
define intent, adjust structure, validate output.
Testing was fully automated – with test cases that run after each change and dynamically extend across both logic and UI.
The complete logic of all checks and rules is automatically and cleanly documented in a .md file.
What began as a simple idea became a functioning application within hours.
No classical development cycle – just a direct translation of domain logic into a working solution.
What the Application Does
The application translates structured investor inputs into a fully defined prompt for portfolio construction.
It combines domain logic, validation, and implementation rules into a reusable advisory workflow.
This is not a static template.
It is a rule-based orchestration of inputs, validation, and output.
Functionality
- Select base currency, risk appetite, investment horizon, equity allocation, ETF count, and preferred exchange Choose between 4 standard strategies per currency between:
- Basic mode – fast prompt for daily use
- Pro mode – advanced configuration for investment professionals
- Plausibility checks for:
- Risk appetite
- Investment horizon
- Required minimum selections
- Automatically adjusts the equity allocation range to the selected risk appetite
- Applies dynamic home-bias logic based on the selected base currency
- Optional requirements for:
- Currency hedging
- ETF look-through
- Synthetic ETF assessment
- Selectable output sections:
- Target stratetigc asset allocation (SAA)
- ETF implementation
- Currency overview
- Rebalancing
- TER estimate
- Choose between English and German for the generated prompt Ready-to-copy prompt for structured portfolio analysis
The result: a consistent, reproducible prompt for strategic asset allocation and ETF-based portfolio construction.
Or simply:
Structured inputs → reusable advisory process.
Open source – but not neutral
The code is fully transparent and open to use:
https://github.com/volkmarritter/allocation-prompt-engine
But it is not neutral.
The allocation logic, rules, and defaults already reflect specific investment decisions – such as risk levels, asset allocation ranges, and implementation choices.
They can be changed, extended, and adapted via a configuration file.
But they are never neutral.
- The code is open.
- The outcome depends on how the logic is defined and configured.
This is essentially a way to structure and operationalize investment logic.
What is changing
What used to take weeks of specification, alignment, and implementation…
now starts with a prompt.
The constraint is no longer implementation, but the quality of the underlying domain logic.
This is where the shift becomes visible.
The constraint is no longer coding.
It is how well the underlying logic is structured.
Tools like OpenAI Codex do not just generate code.
They act as development partners – translating structured intent into working systems.
What matters is no longer how fast code is written,
but how clearly logic, rules, and interactions are defined.
Coding is losing its scarcity
Coding remains relevant, but it is no longer the limiting factor.
As standardizable elements become automated, value creation shifts – from implementation to structuring, from execution to domain clarity, and from isolated features to integrated system logic.
The new bottleneck: structure
Outcome quality increasingly depends on how precisely the underlying logic is defined.
Unclear inputs lead to unclear outputs – regardless of technological capability.
The prompt becomes the specification
More speed, more responsibility
As speed increases, so does the need for control.
Validation, governance, and traceability become critical, as more output without control does not create value.
Strategic perspective
This is less a technological shift and more a change in control logic.
Organizations that are able to structure and operationalize domain logic can unlock both speed and scale – but only if governance is embedded from the outset.
Conclusion
Digital capability is no longer defined by systems.
It is defined by how well logic is structured, governed, and scaled.
Disclaimer: This does not constitute investment advice, an investment recommendation, or a solicitation to buy or sell financial instruments.