Here are the ten key dynamics defining this moment — and what they mean for the firms trying to get ahead.

1. Investment teams are stuck in pilot mode

Teams across the industry are struggling to scale beyond initial AI experimentation. Progress is commonly choked by governance hurdles, restrictive IT environments, poor input data quality, output variability, and escalating token costs. The result is a growing collection of promising proofs-of-concept that never graduate to production. The gap between “interesting demo” and “reliable business tool” remains stubbornly wide.

2. AI ‘slop’ reflects organizational weakness, not technology failure

Poor AI outputs are frequently symptoms of longstanding operational and organizational issues rather than flaws in the technology itself. Sporadic data ownership, undocumented processes, and closed-architecture systems – these were operational, management, and governance shortcomings that existed well before AI arrived. The technology simply makes them visible. Firms that blame the model are often looking in the wrong direction.

3. Fully-agentic workflows are problematic in investment management

Highly regulated asset managers require deterministic, repeatable, and auditable outcomes. The pragmatic model is a hybrid one. AI is best suited to pattern recognition, knowledge retrieval, workflow acceleration, and interpretation, while deterministic systems continue to handle calculations, analytics, and portfolio construction. Humans remain responsible for judgement and decision-making. Fully autonomous AI workflows may work elsewhere, but in investment management the regulatory and fiduciary context demands a more measured architecture.

4. AI can accelerate the building of deterministic tasks and tools

One of AI’s most immediate benefits is its ability to speed up the creation of deterministic workflows and systems. Firms are using AI to build data-processing pipelines, analytics capabilities, and workflow management tools more quickly than traditional development approaches allow. These components then serve as reliable building blocks within broader AI-enabled operating models. The irony is instructive: AI’s clearest near-term value may lie in building the non-AI infrastructure that firms have needed for years.

5. The speed of prototyping has accelerated – creating new risks

“Vibe coding” is hitting investment teams, significantly lowering the barriers to creating new tools, models, and workflows. Investment professionals can move from concept to working prototype in a matter of hours. But many of these solutions are being built in isolation, often without appropriate IT oversight, governance, or long-term support models. The industry’s historic problem of “too many spreadsheets” is quickly becoming a “too many apps” problem.

6. Moving from prototype to production is the critical bottleneck

Building a prototype is increasingly easy; deploying a scalable, auditable, and explainable production solution is far more difficult. Success requires strong foundations – governed technology infrastructure, robust data models and ontologies, and highly modular analytics and application architectures. This is rare in many investment teams that often rely on a single monolithic system and initially focused AI on narrow test cases without addressing the broader enterprise requirements needed for production deployment.

7. Investment managers need a “catalogue of everything”

True enterprise AI adoption cannot happen in the absence of a comprehensive representation of a firm’s entities, data relationships, calculations, and workflows. What the industry needs is an investment-management ontology or semantic model that allows AI to understand and connect information across teams, systems, and data domains. Without this foundation, AI tools operate in fragments – powerful within narrow silos, but unable to reason across the full investment process.

8. AI can help standardize and enforce human-driven processes

Many of the most valuable near-term AI use cases focus on improving consistency, documentation, and workflow discipline rather than generating original investment content. AI can help ensure that processes are followed, standards are maintained, and institutional knowledge is captured consistently. In doing so, it often exposes weaknesses in existing human-led workflows and creates the foundations necessary for more advanced AI applications down the line.

9. Future technology architectures will be AI-native and interoperable

Rather than relying on a single monolithic platform, firms will adopt governed infrastructure layers and “AI control planes” operating through environments such as Microsoft Copilot or Claude. AI will continue to be embedded within specialist systems such as Bloomberg, Aladdin, or Slack, but the greater strategic challenge will be ensuring these tools function seamlessly within a broader enterprise AI ecosystem. The architecture of the future is federated, not centralized.

10. Technology vendors will be judged on interoperability, not just functionality

Poor APIs, closed architectures, and restrictive commercial models around data access are likely to become significant disadvantages. Investment teams increasingly need platforms to expose data, analytics, and workflows headlessly, enabling them to surface capabilities through enterprise AI environments and control planes such as Copilot, Claude, or other orchestration layers. In the years ahead, interoperability may become as important as functionality itself.

How Jacobi Solves the Prototype-to-Production Bottleneck

Moving from a clever prototype to an enterprise-grade, auditable application is the defining hurdle for investment teams today. That is exactly why we built Jacobi: The Platform for Investment Workflow Industrialization.

We help institutional investment managers around the globe convert proprietary investment thinking into scalable, production-grade technology – faster, securely, and with drastically less IT overhead. Jacobi’s Investment Anchor includes:

• The Jacobi AI Workbench — Our workbench provides a secure framework to build proprietary analytics. It combines Jacobi’s SDK, API, and component libraries with advanced AI-assisted coding tools, including Jacobi Skills, Rules, and Model Context Protocol (MCP) Servers.
• Jacobi Production (IaaS) — The robust enterprise engine to securely host, run, and scale your proprietary tools. We handle the infrastructure complexity — including compute scaling, containerization, information security, shared repositories, and external client access management – so your tools actually make it to production.
• Jacobi SaaS – Leverage our standard, out-of-the-box library of advanced analytical tools, workflow applications, and AI agents explicitly designed for the investment industry.
• Forward Deployed Engineering – Our expert engineers embed directly into your team to help design solutions, translate your unique investment logic into production-grade code, and manage it on an ongoing basis.