Agentic AI in Financial Services: A Research Roundup for 2026

Agentic AI is inevitable for financial services, but most firms are structurally unprepared to deploy it safely. 

As a systems integrator and consultancy specialized in AI for financial services, we’ve gathered the research from leading sources like Deloitte, McKinsey, KPMG, Ernst & Young (EY), and other finance-specific studies, giving you a clear view of the implications, opportunities, challenges, and benefits of agentic AI. 

By the end, you’ll have the information needed to improve your decision-making about your own agentic AI strategy for 2026.

In this article:

• 6 of the Most Interesting Statistics on Agentic AI for Financial Services
• What Changes for Financial Services with Agentic AI
• Opportunities and Benefits of AI Agents for BFSIs
• The Financial Services Use Cases Where Agentic AI Is Making an Impact
• The Implementation Challenges Financial Services Firms Face with AI Agents
• How to Use these Insights for Implementing Agentic AI in Your BFSI
• FAQs about Agentic AI in Financial Services

6 of the Most Interesting Statistics on Agentic AI for Financial Services

1. Agentic AI will lead to $3 trillion in corporate productivity and a 5.4% EBITDA improvement for the average company annually based on research on more than 17 million companies worldwide.
2. Organizations can achieve an average 2.3x return on agentic AI investments within 13 months, with ROI expected to grow as adoption scales.
3. 99% of companies plan to put agents into production but only 11% have done so due to implementation challenges related to data, governance and security.
4. Data challenges are the most pressing with 48% of organizations citing governance concerns and 30% flagging privacy issues, and 20% admitting their own data isn’t ready.
5. Data implementation and governance challenges result in financial and reputational losses (77% and 55% respectively across AI incidents).
6. Spending on ready-made artificial-intelligence solutions has dropped from 38% to 32% in 2025, with 84% of organizations believing success now depends on working with specialist providers, like system integrators and consultancies, that can accelerate their AI capabilities.

What Changes for the Financial Services Industry with Agentic AI?

Until recently, artificial intelligence has relied on pre-defined rules and and machine learning models trained on datasets to execute specific tasks. An example of this is bank customer service chatbots that answer simple questions based on an internal knowledge base. But agentic AI can plan, reason, and adapt in real time to handle more complex and multi-step workflows, such as managing portfolios, detecting fraud, and automating compliance, to improve employee and customer engagement.

This potential is driving rapid enterprise implementation and significant investment. KPMG places global market spend on agentic AI at an estimated $50 billion in 2025 (1). According to Wolters Kluwer, 44% of finance teams will use agentic AI in 2026, representing an increase of over 600% (2). Deloitte predicts that 50% of companies that have already implemented generative AI (GenAI) will deploy agentic AI pilots or proof of concepts by 2027 (3).

Agentic AI is also set to fundamentally reshape workplaces. According to a global MIT Sloan study, employees believe artificial intelligence now performs 23% more of their tasks than a year ago and expect it to handle 46% of their tasks within three years (4).

95% of employees at organizations with advanced agentic AI adoption report higher job satisfaction. Additionally, among organizations already using agentic AI extensively, 66% expect to change their operating model and redefine roles, for example by flattening hierarchies and reducing middle management.

In a McKinsey report, there’s a 30% likelihood that artificial intelligence substantially reshapes the global banking sector as a whole by enabling agentic systems to take over core workflows and change how consumers manage their finances. And this could potentially put $170 billion in global profits at risk for banks that don’t adapt their business models (5).

What are the Opportunities and Benefits of AI Agents for BFSIs?

The general consensus is that agentic AI applications in financial services present massive opportunities that can bring notable return on investment in terms of cost savings and operational efficiency to financial services. 

Agentic AI returns are visible in efficiency, productivity and operational savings

KPMG reports estimates agentic AI will lead to $3 trillion in corporate productivity and a 5.4% EBITDA improvement for the average company annually based on research on more than 17 million firms. The report further states that as of June 2025, AI agents’ ability to automate tasks is doubling every three to seven months. 

On average, companies earn $3.50 for every $1 they invest in agentic AI, while the top 5% globally earn about $8 per $1. The report further states that agentic AI could drive a 30% increase in workforce efficiency and a 25% decrease in operational costs by 2027. 

In separate findings, KPMG found that companies using AI agents report 55% higher operational efficiency and an average cost reduction of 35%. 

Agentic AI returns by financial sector

For wealth management specifically, agentic AI can: 

• Cut advisor time on manual prospecting by 40 to 50%.
• Increase prospecting efficiency and increase net new AUM by 30 to 40%.
• Reduce onboarding costs by 30 to 40% while accelerating onboarding by 50% (6).

According to PwC, AI agents can lead to up to 90% time savings in key processes (7). They can also redirect 60% of finance teams’ time to insight work, while resulting in a 40% improvement in forecasting accuracy and speed.

In another 2024 report by PwC, finance teams are spending more time generating insights and less time on automatable tasks, leading to nearly 25% in cost savings because of agentic AI (8).

According to the previously mentioned McKinsey report, artificial intelligence could reduce certain cost categories by as much as 70% across the banking industry. The net effect, however, is expected to be 15%- 20% or $700 billion to $800 billion, due to rising ai technology costs.

First movers see higher returns  

IDC reports that organizations achieve an average 2.3x return on agentic AI investments within 13 months, with ROI expected to grow as adoption of AI scales (9).

McKinsey highlights that the effects won’t be felt equally. Pioneers or first movers are set to gain a 4% return on tangible equity (ROTE) advantage—a key profitability metric—while slow movers are likely to be stuck with an uncompetitive cost base.

IDC’s findings back up this divide. Frontier firms leading in AI adoption achieve returns of 2.84x on their investments, compared to just 0.84x for laggards.

Financial Services Use Cases: Where is Agentic AI Making an Impact?

Across multiple reports, key use cases in financial services lie in AI-driven compliance, client onboarding, and fraud detection where agentic AI can enhance efficiency and augment human expertise. 

Deloitte sees four key areas where artificial intelligence could unlock innovative use cases across the value chain, namely:

• Multi-agent sales acceleration and customer retention
• Automating underwriting
• Enhanced KYC, next-generation transaction monitoring and investigations
• Operational excellence and employee productivity

In a 2025 report, Citi outlines agentic AI in financial services examples, such as fraud prevention, financial forecasting and more, across wealth management, corporate banking, institutional investors and insurance (10), as shown below.

In the McKinsey report, early agentic AI use cases have shown significant potential, enabling zero-touch operations and reducing manual workloads by 30%-50% and this impact is expected to increase. 

Banking Use Cases

The McKinsey report also  highlights that in 2025 alone, 50 of the world’s largest banks announced more than 160 use cases. Of these, some have already shown significant transformative potential. 

For example, a US bank that used AI agents to change the way it creates credit risk memos, experienced a 20%-60% increase in productivity and a 30% improvement in credit turnaround.

According to PwC, agents can reduce cycle times by up to 80% in purchase order (PO) transaction processing and matching while improving audit trails, reducing compliance risk and enabling scale without added cost.

Deloitte also provides key examples of agentic AI use cases that are showing positive results:

• A large Dutch financial institution has been using a combination of AI innovations for its KYC and compliance processes, achieving a 90% reduction in onboarding time and cutting staff workload by 30%. 
• An American financial institution’s employee-facing agent reduced calls to the human-run IT desk by more than 50%.

EY found that when used for manual time intensive Anti-Money Laundering (AML) investigations, agentic AI led to a 50% time reduction per investigation or a saving of two hours of human labor per case (11).

Discover the use cases of agentic AI for banking chatbots and agents

Sardine reports that at one financial institution, Know Your Customer (KYC) workflows resolution rates exceeded 98% on average (12). For more complex tasks, such as sanctions screening or negative news reviews, resolution rates were closer to 55%. 

Wealth, Asset and Investment Management Use Cases

Agentic AI also has valuable applications in wealth and investment management. According to the Sardine report, firms using agentic AI achieved 100% precision in decisioning, compared to under 95% for humans and a four-eyes review process. Additionally, while humans frequently deviated from established policies, agents were much less likely to do so.

KPMG provides two separate examples of agentic AI in financial services.

• They used artificial intelligence to analyze thousands of customer interactions and spot skill gaps that affect customer experiences for a top five wealth manager. This cut analyst time by 66% and helped the firm refine its digital platforms and advisor workflows.
• They built an agentic AI assistant for a top 10 investment management firm that reviews advisor profiles, meeting notes, and other data to create personalized agendas. It also generates summaries and is estimated to cut meeting prep time by up to 50%, saving 20,000 hours a year and supporting added sales coverage.

What are the Implementation Challenges Financial Services Firms Face with AI Agents?

Financial services firms wanting to explore the new wave of agentic AI face implementation, data, security, and risk management challenges.

According to KPMG, 99% of companies plan to put autonomous agents into production but only 11% have done so already. Meanwhile, in another EY survey, 34% of leaders have started using AI agents, but only 14% have fully implemented them (13). These findings highlight the current gap between intent and execution. 

So what’s holding companies back?

In a Forrester AWS paper, 57% of respondents believe that their organizations lack the internal capabilities necessary to take advantage of agentic AI (14). The challenges they cite include:

• Security and risk (63%)
• Lack of interoperability across the technology ecosystem (55%)
• Tech debt (55%)
• Poor data governance (48%)

Similarly, 87% of senior leaders in the previously mentioned EY survey say their main challenges are cybersecurity (35%) and data privacy (30%), as shown in the image below.

In the same study, 64% also say employees worry about AI taking their jobs, which slows down adoption. This resistance is echoed in KPMG’s AI survey, where 45% of respondents stated that employees were resistant to change.

Data Readiness 

Data emerges as a recurring concern across multiple dimensions, from governance (48% in Forrester AWS) to privacy (30% in EY) to overall readiness. In fact, 70% of senior leaders in the EY study say organizations don’t understand how important data readiness is, and 20% admit their own organization’s data isn’t ready, both of which are blocking adoption. 

That’s because agentic AI raises the need for strong data foundations, including well-governed datasets. Poor-quality or fragmented data, common in financial services, increases the risks of hallucination and inaccuracy across AI models, especially as agents scale and interact. 

Build vs Buy Dilemma 

Beyond these challenges, organizations must also navigate the build vs. buy dilemma. 

In the KPMG survey, 67% of respondents stated buying existing agentic AI capabilities was the quickest, most direct option and most preferred method for acquiring AI agents. 

However, while buying may seem like the faster route, most internal builds in financial services fail not just due to skills gaps, but because generic AI teams lack the experience to navigate regulatory requirements, integration complexity, and model risk governance. Off-the-shelf agents often break down in financial audits, fail to integrate with legacy core systems, or don’t meet internal data and security standards. 

This tension is driving a notable shift. According to EY, spending on ready-made AI solutions dropped from 38% to 32% in just six months. More organizations are now building custom AI in-house, reflecting a desire to develop their own AI capabilities internally while maintaining greater control over customization and security. 

Co-creation with experts who understand both the technical and regulatory landscape is emerging as the most sustainable approach. In fact, 84% of respondents in the Forrester AWS paper believe their success now depends on working with specialist providers, like system integrators and consultancies, that can accelerate their AI capabilities.

These partnerships help financial services firms launch safely and compliantly while also enabling the transfer of knowledge to prevent vendor lock-in and enable long-term ownership. 

At Neurons Lab, we emphasize co-development to accelerate time-to-value and ensure long-term ownership, compliance, and adaptability. Reach out to learn more. 

Security and Risk Issues

As AI agents can independently access and act on sensitive customer data and company systems, this presents major security and risk issues, summarized in the table below.

*Summary of Agentic AI security and risk challenges outlined in IBM’s study

Agentic AI significantly expands the attack surface by integrating with multiple systems across an organization, all of which share data (15). This interconnectedness, combined with agents’ autonomy and broad system access, creates new cybersecurity risks and increases the potential for breaches or misuse.

The previously mentioned Citi report states that 50% of all fraud today involves some form of AI and this figure is set to rise. Deepfake scams, for example, have increased more than 2000% over the last three years, with financial institutions among the most targeted victims. 

Agentic AI threatens to accelerate this trend by enabling the mass production and distribution of deepfakes. In finance, this could be used to manipulate transactions, create synthetic identity fraud, and automate large-scale scamming operations. AI agents must access and act on large amounts of sensitive consumer data which increases privacy risk. 

A separate 2025 EY report reveals that while over 75% of financial services firms disclose their use of artificial intelligence to customers, controls in other critical areas are still lacking (16). For instance, 30% of the firms studied had limited or no controls to ensure AI is free from bias.

In a broader industry study by Infosys, only 2% of companies had adequate AI guardrails in place in 2025 (17). As a result, 95% of respondents had experienced at least one AI incident. As shown in the image below, this includes privacy violations (33%), systemic failures (33%) and inaccurate or harmful predictions (32%). 

Of these incidents, 77% resulted in financial losses, while 55% resulted in reputational harm.

The same report states that 86% of executives are aware agentic AI will pose additional risks and compliance challenges. And while a further 83% believe future AI regulations will support rather than slow adoption, they are underinvesting in responsible AI by about 30%.

How to Use These Insights For Implementing Agentic AI in Your Financial Services Firm

Realizing the potential of agentic AI while navigating its challenges and risks requires a strategic, structured approach. According to the research and our own experience, we outline what that entails below:

1. Start with a Clear AI Strategy and Plan

Before deploying agents, establish the strategic foundation that will guide implementation and maximize return on investment:

• Align your leadership and teams around change to reduce resistance. 
• Understand your business priorities and identify the agentic AI use cases and initiatives that will deliver the highest ROI. 
• Assess your current infrastructure, data readiness, governance frameworks, and overall risk assessment processes to ensure they can support AI agents at scale. 
• Build internal capabilities through AI literacy programs, training, or by hiring skilled development teams to implement and maintain autonomous systems.

2. Prioritize Governance to Ensure Safety and Trust

Agentic AI requires the highest level of governance. According to Deloitte, this includes safeguards like agent control rooms, real-time auditing, action logging, human oversight, kill switches, and human override. 

KPMG echoes this by recommending firms run regular stress tests, check for bias, and put clear fail-safe mechanisms in place to prevent unintended outcomes. This means you need to identify where your agents could fail, assess the harm, set clear controls and monitor them continuously so issues are caught early.

Data governance is equally important. IBM argues that a reactive, traditional approach to data management is no longer enough. You’ll need to:

• Proactively identify the data that matters for AI agents, ensure it comes from reliable sources and check its quality. 
• Limit what data your agents can access or share with strict permissions to protect sensitive information.
• Develop a responsible AI framework, strengthen security at every integration point, and ensure you can detect and respond to emerging threats.

3. Ensure compliance and scalability from the start

Agentic AI solutions must be designed to meet regulatory compliance standards, such as GDPR, OCC, MAS, and ISO/IEC 27001, from the start. This is important because agentic AI increases the need for strong data foundations. Poor-quality or fragmented data can lead to hallucinations and unexpected outcomes, especially as agents scale and interact. 

By building with compliance and data quality in mind you’ll ensure your systems can scale into new use cases without costly rebuilds down the line.

Partner with proven Agentic AI and Financial Services experts

Developing the strategy, governance, security, and compliance capabilities for agentic AI requires expertise that many financial institutions don’t have in-house. According to EY, this is where strategic partnerships can be a key differentiator. 

Working with partners that have proven expertise in both agentic AI and financial services means you can set a feasible AI strategy and roadmap. You’ll also have the help you need to prepare your data, integrate AI into your existing processes and optimize implementation while accelerating adoption and reducing risk.

Neurons Lab is a global AI consultancy with AWS advanced competencies in AI and Financial Services. We help organizations become AI-native through our comprehensive approach that combines leadership alignment with technology integration to achieve measurable outcomes.

If you’d like to learn more about how we can help you develop, implement, and scale agentic AI systems tailored to your organization’s needs, get in touch with us today

FAQs about Agentic AI in Financial Services

Will agentic AI replace human finance professionals?

No, agentic AI will not replace human finance professionals. Instead, it will augment their capabilities, free them from routine, automate tasks and leave more time for strategic work.

How can financial services firms prepare employees for agentic AI?

Financial services firms can prepare their employees for agentic AI by establishing clear governance and risk frameworks, investing in AI training and upskilling on the responsible use of AI as well as prioritising change management to foster a culture of human-AI collaboration.

How much does it cost to develop an agentic AI system as a bank?

The costs to develop AI as a bank vary based on the complexity of your implementation, whether you choose to build inhouse or work with an external partner and other factors. Online estimates vary widely placing the cost at anywhere from as low as $15,000 to well over $500,000.

How do I ensure a high ROI with agentic AI in financial services?

To ensure a high ROI with agentic AI in finance, align your AI projects to a clear strategic goal, target high-volume use cases that create scalable value, define measurable success metrics, and track progress with reliable baseline data.

When is it better to build agentic AI capabilities in house for wealth management firms?

It may be better to build If you already have established AI teams with a proven track record of deploying and maintaining complex systems at scale. If you don’t have that expertise yet,  partnering with an AI consultancy like Neurons Lab can help you deploy agentic AI solutions while developing your own team’s skills and capabilities to manage and expand the systems independently over time.

Sources:

1. https://kpmg.com/kpmg-us/content/dam/kpmg/pdf/2025/kpmg-agentic-ai-advantage.pdf
2. https://www.wolterskluwer.com/en/news/pr-2025-wolters-kluwer-survey-increasing-adoption-agentic-ai
3. https://www.deloitte.com/content/dam/assets-zone1/in/en/docs/services/engineering-ai-data/2025/in-eaid-agentic-ai-in-financial-services.pdf
4. https://sloanreview.mit.edu/projects/the-emerging-agentic-enterprise-how-leaders-must-navigate-a-new-age-of-ai/
5. https://www.mckinsey.com/de/~/media/mckinsey/industries/financial%20services/our%20insights/global%20banking%20annual%20review/why-precision-not-heft-defines-the-future-of-banking.pdf
6. https://kpmg.com/kpmg-us/content/dam/kpmg/pdf/2025/agentic-ai-changing-wealth-mgmt.pdf
7. https://www.pwc.com/us/en/tech-effect/ai-analytics/ai-agents-for-finance.html
8. https://explore.pwc.com/febs2024/finance_effectiveness_study_2024
9. https://marketingassets.microsoft.com/gdc/gdcflXNT6/original
10. https://www.citiwarrants.com/home/upload/citi_research/rsch_pdf_30305836.pdf
11. https://www.ey.com/content/dam/ey-unified-site/ey-com/en-ca/services/ai/documents/ey-how-agentic-ai-can-improve-anti-money-laundering-investigations.pdf
12. https://go.sardine.ai/hubfs/Whitepapers/The%20Agentic%20Oversight%20Framework%20-%20Procedures%2C%20Accountability%2C%20and%20Best%20Practices%20for%20Agentic%20AI%20Use%20in%20Regulated%20Financial%20Services.pdf
13. https://www.ey.com/content/dam/ey-unified-site/ey-com/nl-nl/services/ai/documents/ey-ai-pulse-survey-report-v05.pdf
14. https://pages.awscloud.com/rs/112-TZM-766/images/Forrester_AWS%20Marketplace_How%20Financial%20Services%20Leaders%20Are%20Approaching%20Security%20And%20Innovation.pdf?version=1
15. https://www.ibm.com/downloads/documents/gb-en/12f5a71117cdc329
16. https://www.ey.com/content/dam/ey-unified-site/ey-com/en-gl/industries/financial-servicesemeia/documents/ey-gl-european-ai-financial-services-pulse-survey-09-2025.pdf
17. https://www.infosys.com/iki/research/responsible-enterprise-ai-agentic.html