Following his recent panel discussion at TSAM New York, MAIA’s CEO, Mark Veevers, has been reflecting on what scalability really means for asset managers. He believes that whilst growth creates opportunity, it can also expose weaknesses in the operating model. New strategies, clients, markets and investor requirements bring more processes, integrations and exceptions; without the right foundations, complexity can grow just as quickly as the business.
“Scale is not about processing more. It is about adding complexity without adding equivalent cost, people or risk.”
Standardise what does not need to be bespoke
Time-to-market is often lost in data readiness, integration dependencies, testing and decision-making rather than the core technology. Treating every new strategy, client or investor type as completely unique makes it difficult to scale efficiently. The aim should be to industrialise what is repeatable and reserve bespoke effort for what genuinely differentiates the investment proposition.
Put trusted data at the centre
Open architecture matters because it determines how easily firms can access, combine and act on their data. An API-first platform such as MAIA allows clients to get deep inside the system, using both their own data and the data created through the investment process without introducing another silo.
“AI becomes far more powerful when it is working across a contiguous dataset that is created, controlled and audited at the heart of the investment process.”
Make the platform AI-ready
Access alone is not enough. AI needs context, lineage, permissions and reliable data if it is to reason across the investment lifecycle rather than respond to disconnected prompts. That is why strong governance and a coherent data model are becoming prerequisites for meaningful AI adoption.
Embed AI into the workflow
Next-generation platforms should not treat AI as a bolt-on assistant. AI needs to sit inside the core workflow, able to inspect data, understand context, identify patterns and interact with the platform itself. The opportunity is to remove repetitive tasks that still require logical thinking and evaluation, not just simple mechanical activity.
Move from alerts to recommendations
The next step is for the platform to do more than surface an exception. It should help evaluate what has happened, why it happened, recommend a course of action and, where appropriate, prepare a draft outcome for the user to review and approve. This keeps the human in the loop while materially reducing the effort needed to reach a decision.
“We should automate tasks, not accountability: let AI do the repetitive analysis and preparation, while the user retains judgement, approval and control.”
Use AI to create capacity
Repetitive checking, chasing, reconciling and investigating consume experienced people who could be applying judgement elsewhere. The goal is not simply fewer people; it is an operating model that can support more strategies, clients and complexity without requiring headcount and operational burden to rise at the same rate.
Two routes to AI-enabled scale
For MAIA, the opportunity has two dimensions: giving clients open, governed access so they can apply their own AI capabilities to the platform and embedding AI directly into MAIA so the system itself can reason, recommend and prepare actions within controlled workflows. Both depend on the same foundation: trusted data, open architecture and human accountability.
