Across industries, the pace of AI adoption is accelerating — but so is the complexity. Recent reports highlight that leading AI firms like OpenAI, Anthropic, and Cohere are now hiring forward-deployed engineers (FDEs): specialists who sit within client organizations to co-create, customize, and implement AI models.
This “embedded engineering” approach represents a significant shift. Instead of providing one-size-fits-all tools, these AI leaders are embedding technical talent directly into their clients’ operations — helping bridge the critical gap between innovation and real-world execution.
For asset managers, who deal with complex data, sensitive IP, and strict governance requirements, this shift carries powerful implications.
Context — The AI Opportunity in Asset Management
In the financial sector, the barriers to adopting AI are often not technological, but organizational. Firms operate within tight regulatory frameworks and handle proprietary strategies they cannot easily disclose to external vendors.
AI leaders have recognized this challenge. As the Financial Times reports, companies like OpenAI and Anthropic are embedding engineers inside their clients’ organizations to accelerate adoption while safeguarding sensitive information. This model allows businesses to tailor AI systems to their unique workflows — from data ingestion to model training and compliance.
At its core, forward-deployed engineering represents a new phase in enterprise AI: one where technology doesn’t just serve the business — it grows from within it.
Our View — AI Must Be Developed Within the Business, Not Just For It
We believe asset managers have much to learn from this embedded model.
In our view, the next phase of AI adoption in financial services will be won not by firms that outsource innovation, but by those that co-develop it from the inside — aligning data science, investment expertise, and client needs from day one.
Too often, we see firms delegating AI projects to IT or external consultants, creating a gap between strategic ambition and practical delivery. The result is “shadow AI” — fragmented, ungoverned initiatives that never scale.
Instead, we believe the future lies in collaborative product discovery: cross-functional teams of engineers, data scientists, and business leaders embedded within investment functions. This model ensures that AI solutions are not only technically sound but also contextually relevant — built around the firm’s investment philosophy, risk culture, and operational nuances.
As we see it, this approach mirrors the success of forward-deployed engineers in leading AI companies: those who “learn what customers need, experiment and innovate together,” and turn insights into tangible outcomes.
Implications for Asset Managers
1. Integrate engineering and investment expertise
AI can’t thrive in isolation. The most effective models are those developed in tight collaboration between quants, portfolio managers, and technologists. Asset managers should embed data engineers within investment teams, enabling faster iteration and deeper understanding of alpha drivers.
2. Protect IP while accelerating innovation
By developing AI capabilities internally — or in hybrid embedded partnerships — firms can retain control over proprietary strategies while benefiting from external expertise. This reduces the risk of overexposure to generic vendor tools.
3. Build AI literacy at every level
Embedding engineers isn’t only about coding; it’s about translation. Forward-deployed teams serve as interpreters between data science and decision-making. For asset managers, that means ensuring portfolio managers and analysts understand enough about AI to challenge, guide, and trust the outputs.
4. Make governance and ethics a design principle
Embedding teams internally allows firms to weave explainability, auditability, and compliance directly into model design — not as afterthoughts, but as foundational features.
Conclusion — Winning with AI from the Inside Out
We believe the asset managers who will lead in the AI era are those who bring the business, IT, and investment leadership to the same table.
Forward-deployed engineering is not just a staffing model; it’s a philosophy — one that turns AI from an external tool into an internal capability.
