Make: From Concept to Code

"Vision without execution is just hallucination." That line, attributed to Edison, captures the gap that defines most failed AI initiatives. The Think phase is where firms explore ideas, define problems, and identify opportunities for AI to drive transformation. But thinking alone doesn't create impact. The Make phase is the critical bridge between strategy and execution.

While AI has been romanticized as a near-magical force, the reality is that AI solutions are built, not conjured. The process of making AI work in real-world financial settings requires not just cutting-edge models, but deep expertise in software development, data integration, and user experience.

So how do we turn ideas into functional AI-driven solutions?

The blueprint: understand the problem before writing the first line of code

Every great AI solution starts with a deep understanding of the problem. In financial services, this often means navigating:

• Complex regulations and compliance constraints
• High-stakes decision-making processes
• The need for accuracy, transparency, and auditability
• The coexistence of AI with human expertise

A successful Make phase does not begin with coding. It begins with defining a clear problem statement, data strategy, and success metrics. The goal is to refine use cases before building anything, ensuring that AI aligns with real-world needs rather than being a solution in search of a problem.

From idea to intelligent software

1. Architect the AI solution

Making AI work is not just about selecting the most powerful model. It is about engineering a solution that integrates seamlessly into existing workflows. Before building, the right questions are:

• What data sources will power the AI? Structured financial reports? Analyst transcripts? Operational data?
• How will AI output be used? Will it assist due diligence analysts, automate compliance checks, or drive investment insights?
• What level of explainability is required? In regulated environments, black-box AI is unacceptable.

Answering these questions upfront produces AI architectures that are not only intelligent but also practical and compliant.

2. Engineer the data foundation

AI is only as good as the data it learns from. A critical part of the Make process involves:

• Extracting and cleaning massive amounts of structured and unstructured data
• Training models on industry-specific datasets to improve accuracy
• Ensuring continuous learning so AI evolves with new financial trends

Strong data pipelines allow AI models to ingest, interpret, and synthesize financial data at scale. This gives analysts a clearer, faster path to insights.

3. Develop and train the model

Once the foundation is set, models can be trained and fine-tuned to perform tasks such as:

• Analyzing investment documents for risk factors
• Identifying anomalies in private placement documents
• Extracting key insights from earnings calls and regulatory filings

AI development is not one-size-fits-all. Models should be customized to each firm's unique needs, whether that means increasing accuracy in alternative investment analysis or automating due diligence workflows. The investment in time on the front end ensures that when the tools are deployed, they integrate immediately and deliver meaningful ROI.

4. Build the interface

Even the most advanced AI is useless if end users cannot or will not use it. A critical step in the Make phase is designing an interface that integrates AI into everyday workflows:

• Embedding AI into existing software platforms to avoid workflow disruption
• Providing intuitive dashboards and natural language search tools
• Building audit trails and compliance reporting into the system

AI in wealth and asset management must enhance decision-making, not replace human expertise. Prioritizing usability ensures that AI is an asset rather than an obstacle.

5. Test, iterate, deploy

AI solutions are not static. After development, the work is to deploy in real-world environments and gather user feedback to refine accuracy, efficiency, and experience.

• Does the AI produce relevant, reliable insights?
• How does it perform under different financial market conditions?
• Is it aligning with compliance and risk management expectations?

Through constant iteration, solutions evolve to stay ahead of industry challenges.

Why Make matters

Without a structured, disciplined approach to the Make phase, AI remains just an exciting idea. The challenge is moving from ideation to impact. Firms that skip the rigor of this phase end up with prototypes that never become products, or products that never become operational.

Strategy without engineering is theater. Engineering without strategy is wasted code. The Make phase is where the two have to meet.

This is the second piece in the Think. Make. Do. series. Next: where AI becomes operational in the Do phase.