Agentic AI needs trusted financial data. Here's why.
Picture a family with pensions scattered across former employers, investments held with different providers and a home insurance policy coming up for renewal. They may be paying more in fees than they realise, missing opportunities to grow their savings or simply lacking a clear picture of their finances.
The first wave of consumer AI tools could explain what pension consolidation means or how investment fees affect long-term returns. The next wave is likely to go further.
Agentic AI could help this family identify pensions worth consolidating, compare investment costs, review their insurance coverage and take the next steps when they're ready to act.
The potential is significant. But there's a fundamental challenge: even the most capable AI agent cannot make informed financial decisions without access to accurate information about the person it's helping.
And today, that information is scattered across multiple providers, systems and markets.
An AI agent is only as good as its data
To identify opportunities to reduce pension fees, an AI agent needs to know where someone's pensions are held, what they're worth and what they cost.
To evaluate an investment portfolio, it needs to understand the customer's actual holdings, allocation and financial circumstances.
And to assess whether an insurance policy provides suitable coverage, it needs access to the policy itself, not just general information about insurance products.
Without this context, even the most sophisticated AI model is working with an incomplete picture.
This is especially important in financial services, where decisions can have long-term consequences. A recommendation based on missing or outdated information could lead to unnecessary costs, unsuitable products or missed opportunities.
The intelligence of the model matters. But so does the quality of the information it can access.
Consumers are already finding their own ways to share financial data
Demand for AI-powered financial services is growing.
A recent EY survey across 23 countries found that almost half of respondents had used AI to help guide savings or investment decisions within the previous six months.
Yet accessing the financial information needed to make these tools genuinely useful remains complicated.
In the absence of seamless connections, consumers are finding workarounds. They upload PDFs and screenshots, copy information from account statements or, in some cases, share login credentials with services that access their accounts on their behalf.
These approaches may provide a temporary solution, but they also introduce limitations and risks.
Documents can quickly become outdated. Manually shared information may be incomplete. Sharing login credentials can create security concerns, while customers may have limited visibility into how their information is stored, used or accessed.
For financial institutions, these improvised connections also make it harder to understand which services are accessing their systems and under what conditions.
The demand for connected financial experiences already exists. The challenge is making that connectivity secure, reliable and scalable.
As AI agents become more capable, the need for better ways to access financial information will only increase.
What trusted financial data access looks like
For AI agents to become genuinely useful in financial services, they need more than access to information. They need access they can rely on.
That starts with customer consent.
People should understand which services are accessing their financial information, for what purpose and under what conditions. They should also be able to control and withdraw that access.
The information itself needs to be accurate, consistent and usable across different financial providers.
A pension statement, an investment portfolio and an insurance policy may come from entirely different systems, each with its own terminology, formats and data structures. Bringing this information together requires more than simply establishing a connection.
It requires an infrastructure layer that can collect, standardise and structure financial information so that applications and AI agents can work with it.
Security and accountability are equally important. Financial institutions and technology providers need to understand how information is accessed, how permissions are managed and which responsibilities apply when data is used.
These principles become particularly important with agentic AI.
There is a meaningful difference between giving an application permission to retrieve information and authorising an AI agent to take action.
A customer might want an agent to analyse their pensions and recommend opportunities for consolidation. That does not necessarily mean they want the agent to initiate a transfer.
As AI moves from answering questions to carrying out tasks, distinguishing between accessing information, making recommendations and executing actions becomes essential.
The underlying requirement is clear: AI agents need financial data that is consented, reliable and structured, delivered through infrastructure designed for secure access and responsible use.
From individual products to a complete financial picture
For decades, financial services have largely been organised around individual products and providers.
Customers log into one application to see their investments, another to review their pension and a third to manage their insurance.
Each provider understands the products it holds, but rarely has a complete view of the customer’s financial situation.
This creates limitations for both consumers and financial institutions.
A wealth platform may be able to recommend investments based on the assets held within its own system, but without visibility into external pensions and savings, it is working with only part of the customer’s financial picture.
An insurer may understand a customer’s existing policy but have limited insight into their broader financial needs.
And an AI assistant may be capable of sophisticated analysis while lacking the information required to make that analysis personally relevant.
Open finance changes what is possible by enabling customers to share financial information across providers.
When that information is brought together and made usable, financial services can move beyond isolated products towards a more complete understanding of each customer’s circumstances.
For AI, this financial context is particularly valuable.
Instead of providing generic answers, an assistant could help someone understand how their pension savings relate to their investments, identify opportunities to reduce unnecessary costs or assess whether their insurance coverage still meets their needs.
The opportunity is not simply to make existing financial services faster. It is to build services that understand more of the customer’s financial reality.
Financial data infrastructure becomes essential for AI
Much of the attention around agentic AI focuses on what models can do: their reasoning capabilities, ability to use tools and potential to complete increasingly complex tasks.
But in financial services, the infrastructure connecting those models to real financial information is just as important.
Without dependable access to external financial data, developers face a difficult choice.
They can build and maintain individual integrations with financial institutions, manage different authentication processes and reconcile inconsistent data formats themselves.
Or they can rely on incomplete information supplied manually by users.
Neither approach provides an ideal foundation for scalable financial AI.
A financial data infrastructure layer addresses this challenge by handling the complexity of connecting to providers, collecting customer-authorised information and transforming it into consistent data models.
For developers, this means they can focus on the experiences and intelligence they want to build rather than the underlying connectivity.
For financial institutions, it creates opportunities to develop more personalised services, improve advisory experiences and reduce friction in processes such as investment transfers and pension consolidation.
And for customers, it means AI-powered financial services can become more relevant to their actual circumstances.
At Anessa, this is the infrastructure we are building: connecting financial services and AI applications to the customer’s broader financial context across investments, pensions, savings, insurance, loans and bank accounts.
Through our financial data infrastructure, APIs and MCP connector, developers can build AI-powered experiences using customer-authorised financial information without having to create and maintain the underlying integration layer themselves.
The next challenge for financial services
Agentic AI has the potential to change how people interact with financial services.
Instead of navigating multiple applications, comparing products manually and trying to piece together their financial situation, customers could increasingly rely on intelligent assistants to help them understand their options and take the next steps.
For banks, insurers, wealth platforms and fintechs, this creates an opportunity to build services that are more useful, more personalised and better connected to customers’ financial lives.
But realising that opportunity requires looking beyond the AI model itself.
An agent cannot identify unnecessary pension fees if it cannot see what the customer is paying.
It cannot help someone understand their investments if it only has part of the picture.
And it cannot responsibly act on someone’s behalf without clear permissions and appropriate safeguards.
The next generation of financial services will not be defined by AI capabilities alone. It will depend on how effectively those capabilities connect to the financial lives of the people they are designed to serve.
Because even the most capable AI is only as good as the data it can access.
