ScienceSoft Global
ScienceSoft Finance
Q3 2026 Investment AI Trends: AI Rollouts Accelerate, Scaling Lags. AI Autonomy Remains Limited
Related topics:
5 min read
Last updated:
An investment AI transformation and software development company with 19 years in the field, ScienceSoft creates custom AI solutions that improve investment decision-making, operational efficiency, and client experience. Drawing on ScienceSoft’s Q3 2026 Investment AI Market Watch and recent project work, we sum up where AI is delivering value now and what investment firms and wealthtech vendors should prepare for over the next 12 months.
At a glance:
- Investment AI adoption accelerates. Firms are raising AI budgets despite modest financial returns, prioritizing immediate efficiency gains and aiming to expand business capacity.
- Agentic AI moves into production, but autonomy remains limited. Enterprises and vendors focus on human-supervised agents for financial research and administrative workflows.
- Investment firms prefer buying commercial AI tools to building proprietary AI solutions. Business process outsourcing to AI-first firms is emerging as an alternative to in-house AI adoption.
- Slow organizational transformation remains a major barrier to AI scaling. Data readiness, governance, workforce preparation, and workflow redesign are lagging behind AI adoption.
AI Technologies Covered in This Report
|
Assistive AI |
|
|---|---|
|
AI that responds to requests from humans but doesn’t act autonomously Generative AI (GenAI) technology powering copilots, chatbots, and voice assistants that help investment professionals and clients with everyday tasks. GenAI assistants can analyze investment documents, retrieve and summarize financial data, draft reports and communications, and support advisors. |
AI that can act autonomously or semi-autonomously Technology that combines GenAI with workflow automation to carry out multi-step operations. AI agents can use large language models (LLMs), traditional machine learning (ML) models, and rule-based engines to analyze data, perform tasks, and trigger approved actions across systems, typically under selective human oversight. |
Investment AI Adoption Grows, With Efficiency the Primary Near-Term Goal
AI adoption continued to accelerate across investment and wealth management in Q3 2026, with organizations increasingly moving from AI exploration and experiments to practical use. The Sixth Annual Defined Contribution Consultant Study by T. Rowe Price found that the share of financial advisory firms still evaluating AI fell from 44% in 2025 to 14% in 2026. At the same time, the study showed a substantial increase in the share of firms reporting AI use (roughly 2–3.5x uplift across several key advisor tasks).
Financial advisors' interest in AI remained strong. The 2026 AssetMark Advisor Insights: Artificial Intelligence report found that 85% of US financial advisors have adopted AI within their practices, and 80% expect to increase AI use over the following 12 months. 41% of surveyed advisors reported widespread AI adoption across multiple processes, and 14% indicated AI use across most of their practices.
The asset management domain showed similarly strong interest in AI adoption but more modest deployment rates. Acuity Analytics’ 2026 annual survey of asset managers found that 81% of asset managers across the US, Europe, and Asia-Pacific have AI initiatives underway. However, only 37% have achieved moderate adoption, and 5% have fully integrated AI into their operations.
Assistive AI continued to dominate live deployments. AssetMark identified meeting summary preparation, investment performance report drafting, and research support among advisors' leading AI use cases. Similarly, a 2026 report by London Stock Exchange Group (LSEG), The blueprint for growth: AI’s role in Wealth Management, found that wealth management firms primarily use generative AI to assist with client data analysis, communications, and summarizing complex documents.
In Q3 2026, operational efficiency remained the strongest immediate target for AI adoption. T. Rowe Price found that the largest share of investment advisors (78%) use AI most frequently for efficiency-related activities. Acuity Analytics reported a similar emphasis on efficiency in its survey of asset managers: operational efficiency was the most frequently cited primary driver of AI use (24%), and incremental efficiency improvements topped the list of AI’s expected transformative industry impacts over the next 1–3 years (21%).
Input from ScienceSoft’s investment clients suggests that longer-term expectations for AI center on increasing business capacity through AI-supported automation. Broader investment industry sentiment points in the same direction. AJ Emanuele, Senior Vice President of Product Management at Dynasty Financial Partners, noted in a discussion with Trevor Hicks, Chief Technology Officer at Advisor360°, that AI's strategic economic opportunity lies in enabling investment firms to “add clients and AUM without needing to add people at anywhere close to the same rate, and without the client experience getting watered down.”
AI outcomes reported through Q3 aligned with the industry's priorities. The 2026 State of Wealth Management AI Adoption study, conducted by Cerulli Associates in partnership with Vista Equity Partners, found that the three most frequently reported measurable AI outcomes among US registered investment advisers (RIAs) and wealth management firms are operational efficiency gains: reduced manual and administrative work (64%), improved client communication (46%), and streamlined investment research (33%).
AssetMark also highlighted AI's role in enhancing work quality and business capacity. In its 2026 survey, advisors cited improved work quality most often (50%), followed by business growth (43%) and time savings and greater efficiency (42%). 54% of advisors who had adopted AI reported saving at least 4 hours per week, equivalent to more than 200 hours, or 26 eight-hour working days, annually.
Investment Firms Increase AI Budgets Despite Moderate Financial Returns
In Q3 2026, investment firms continued to raise their AI budgets steadily. The aforementioned Cerulli Associates study found that 70% of US RIAs and wealth management firms expected to increase AI spending in 2026, with a median firm planning to raise its AI investments by roughly two-thirds. The same study projected average AI expenditure to rise from 8% of firms' technology budgets in 2025 to 15% by the end of 2026, and total technology budgets to grow by an average of 13%.
In Q3, AI increasingly became a dedicated technology budget line item. A 2026 survey of RIAs, wealth managers, and broker-dealers by F2 Strategy found that the share of firms with a dedicated AI line in their technology budgets rose from 14% in 2025 to 67% in 2026.
Large investment firms are directing AI budgets toward advisor productivity, business growth, and client experience. Ameriprise Financial disclosed an annual technology and AI budget of $1 billion (5.5% of its 2025 net revenue), including AI investments. The firm emphasized three major areas of AI spending: helping advisors identify business growth opportunities, automating administrative workflows, and enabling more personalized client interactions.
At the same time, financial returns have yet to match the pace of investment. A survey by LSEG found that 69% of wealth management firms see only small (0.1–2.9%) or moderate (3–6.9%) positive returns on AI investments, and 12% see no or negative returns.
ScienceSoft consultants note that an average AI investment payback period of around two years, also reported by LSEG, may partly explain the modest ROI, as AI solutions introduced in 2025–2026 may not have had enough time to deliver their full financial value. The consultants also highlight that while targeted AI use cases can deliver quick returns, early investments in AI-enabling infrastructure and data foundations may generate little direct financial value on their own, weighing on overall AI ROI figures until more applications are deployed on top of them.
Measuring AI's financial impact presented a challenge in itself. F2 Strategy found that 65% of investment firms have no formal AI success metrics. In LSEG's survey, 30% of wealth managers identified unclear ROI as a major barrier to scaling AI.
There's an interesting contradiction here. Investment firms spend their days valuing opportunities and calculating returns. Yet many still struggle to put a reliable number on their own AI investments. When I ask clients and industry contacts what makes AI ROI so hard to measure, one issue comes up repeatedly: there's no direct link between the operational capacity AI frees up and the actual financial returns.
Say an AI assistant saves an advisor four hours a week. That's a measurable efficiency gain, but what happens to those four hours? Does the advisor spend more time with existing clients, bring in new assets, or just end up with a lighter workload?
When working with investment firms, we usually recommend they track several things separately: efficiency gains from AI, freed-up capacity, how that capacity is used, and the resulting financial impact. This helps firms see not just whether AI works, but whether they're making good use of it. If AI frees up staff but that doesn't translate into financial gains — higher revenue, avoided hiring, lower overtime expenses, or similar — it usually means the firm isn't directing the extra capacity toward value-generating work. In that case, investing in more AI would not improve returns, but adjusting workflows or responsibilities might.
Executive Director of Solution Design & Competencies, Financial IT & AI Principal Consultant, ScienceSoft
Generative AI Expands in Advisor Support. Investment Firms Pursue AI-Assisted Portfolio Workflows
In Q3 2026, investment firms continued deploying generative AI primarily as an assistive technology for advisor operations, investment research, and client communications. At the same time, enterprise initiatives signaled growing interest in applying GenAI to more complex portfolio management activities.
Advisor workflow assistance remained a major implementation area. Enterprise deployments focused on GenAI assistants that help investment and finance professionals retrieve and summarize information, analyze documents, navigate internal processes, and prepare recommendations and next steps. Prominent examples include Wells Fargo's launch of AI Teammate within its Advisor Gateway desktop platform, Insigneo's implementation of the Claude-powered Insigneo Executive AI Assistant, and LPL Financial Holdings' rollout of GenAI-powered advisor workflows within its Latitude platform.
Investment research remained another priority for enterprise GenAI adoption. T. Rowe Price announced plans to deploy Anthropic's Claude Cowork to help investment professionals consolidate financial data and support multi-step research workflows. The initiative aims to give analysts and portfolio managers more time for proprietary research, investment analysis, and decision-making.
ScienceSoft's recent project experience also indicates sustained interest in GenAI-assisted investment research. One recent engagement involved developing GenAI capabilities to support traders in investment research and analysis.
Portfolio management emerged as the next major AI application area. In Acuity Analytics' survey, 64% of asset managers identified portfolio management as the area where AI could have a significant impact. ScienceSoft expects investment firms to expand GenAI use in portfolio analysis, risk assessment, and investment decision support through 2027, as they strengthen the operational foundations and governance needed for more complex use cases.
Some early enterprise initiatives supported that direction. Millennium Management announced a collaboration with Anthropic to develop an AI-powered digital risk analyst. The GenAI tool will help risk managers analyze portfolio risk positions and interpret daily changes in investment risk. It will also recommend responses to risk exposures across asset classes.
In Q3 2026, technology vendors expanded GenAI capabilities to support advisors and more complex portfolio workflows in response to enterprise demand. SS&C Technologies introduced a pack of commercial GenAI tools for wealth advisor assistance, personalized retirement outreach, fund valuation and allocation reviews, and asset management prospecting. The vendor also released the Genesis AI portfolio management assistant, enabling investment teams to consolidate portfolio insights, discover opportunities, and initiate workflows through natural-language requests.
Anthropic launched Claude for Financial Advisors, a specialized GenAI assistant for investment and wealth management that supports client meeting preparation, investment research, portfolio rebalance reviews, compliance documentation, and client follow-ups. Prebuilt integrations with major custodian, asset management, and wealthtech platforms simplify assistant access to firms’ existing financial data sources and operational systems.
Agentic AI Moves Into Production. Investment Firms Favor Controlled Autonomy
Agentic AI adoption in investment gradually progressed through Q3 2026. Acuity Analytics found that 64% of asset management firms across the US, Europe, and Asia-Pacific have agentic AI initiatives underway; 16% reported moderate adoption and 15% reported significant usage.
However, AI agent autonomy was limited, as investment professionals remained cautious about delegating responsibility for investment, regulatory, or client outcomes to AI. AssetMark found that 50% of financial advisors would not trust AI to handle client-facing work, 45% would not let it make portfolio decisions, and 43% identified compliance as an area that should remain with humans. ScienceSoft expects controlled, semi-autonomous agents to dominate adoption through 2027.
In the enterprise segment, investment research emerged as a primary use case for agentic AI. Balyasny Asset Management (BAM) launched BAMAgent, a proprietary platform supporting thousands of AI agents. Early agents focus on coordinating and executing multi-step research tasks: preparing research packages, analyzing earnings and macroeconomic events, and translating new data into financial modeling scenarios. BAM’s research agents operate within approved systems and permissions and with human review gateways.
Middle-office reconciliation was another area for agentic AI deployment. Pacific Investment Management Company (PIMCO) implemented AI agents to investigate discrepancies between financial records. The agents identify likely causes, distinguish genuine exceptions from false alarms, and flag discrepancies requiring specialist attention, reducing the time employees spend investigating reconciliation issues.
Large-scale agentic automation also began attracting investment. Mariner announced a five-year partnership with Humanity Labs to deploy AI agents representing the equivalent work capacity of more than 700 full-time employees. Planned applications include client onboarding, account opening, billing, reporting, cross-sell opportunity capturing, and compliance reviews. The initiative aims to increase Mariner’s operational capacity without a comparable increase in headcount.
The market for commercial investment AI agents remained relatively narrow in Q3 2026. Everest Group's Innovation Watch: Agentic AI in Wealth Management Technology describes agentic AI products for wealth management as an emerging market still dominated by specialized agents for isolated advisor tasks. However, the quarter saw technology vendors begin to introduce agents that coordinate broader administrative tasks within workflows and interact with multiple enterprise systems.
One major release came from Advisor360°. The vendor introduced Meeting Prep, an AI agent that continuously analyzes CRM records, portfolio performance, financial plans, market data, and advisors’ open tasks and proactively prepares client meeting agendas and talking points. The agent can also identify potential operational issues and opportunities, including missing beneficiaries and tax-loss harvesting options. In live tests, Meeting Prep helped advisors reduce meeting preparation time by 60%.
Investment Firms Mostly Access AI Through Commercial Tools. BPO to AI-First Firms Emerges as an Alternative to AI Adoption
In Q3 2026, most investment firms adopted off-the-shelf AI tools rather than building proprietary AI solutions. The Cerulli Associates survey found that 84% of US RIAs and wealth management firms use general-purpose GenAI assistants like ChatGPT or Microsoft Copilot; 59% rely on AI capabilities built into their existing technology platforms; and 54% use specialized commercial AI tools. Only 24% have custom AI solutions developed internally or specifically for their organization.
Looking ahead, hybrid sourcing will likely become the preferred approach. The same survey revealed that 66% of firms plan to combine packaged AI capabilities from existing technology partners with specialized commercial tools for selected workflows. Only 10% intend to build proprietary AI solutions.
Commercial AI has become so capable that investment firms rarely need to build the core technology themselves. But there are still areas where custom engineering is hard to avoid. In ScienceSoft's projects, it's often the data foundation that requires the most custom work. An AI tool may be ready to use, but if a firm's data is scattered across systems, stored in inconsistent formats, or lacks proper quality and access controls, the tool won't be able to use it reliably. And there's no off-the-shelf product that can fix every firm's data problems.
Data preparation needs to happen well before AI deployment, and it's essential for scaling. You can get one AI use case working with a few dedicated data integrations. But when you're rolling out dozens of use cases across functions, you need a reliable, reusable way to give AI access to the right data. Otherwise, every rollout turns into a major integration project.
Security is another area where custom work often makes sense. From our security work with some of the world's largest asset managers, we know that cloud platforms and AI vendors rarely provide every protection a firm needs to address its specific risks and requirements. You may need additional safeguards to control what AI can access, what data it can share, and which actions it can take.
Head of AI, Principal Architect, ScienceSoft
Notably, Q3 2026 also saw investment firms turning to AI-enabled business process outsourcing (BPO) as an alternative to in-house AI adoption. Under this model, investment firms entrust the execution of their business processes to third-party service providers that use AI alongside their own specialists, while retaining oversight of the outsourced activities. Cabot Wealth Management became the first production customer of an AI-powered managed reconciliation service by Ridgeline. The company now has Ridgeline's reconciliation specialists and vendor-operated AI agents handling daily reconciliation of its investment positions, cash, and transactions.
ScienceSoft expects AI-enabled BPO to attract growing interest among smaller investment firms that lack the resources to build, integrate, and supervise agentic AI independently.
Continuing a trend ScienceSoft tracked in Q2 2026, large investment firms expanded their role in raising and providing capital for AI infrastructure. In Q3, NVIDIA partnered with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to facilitate financing for global AI computing infrastructure, with an ambition to mobilize more than $500 billion in third-party capital.
Investment Firms Pursue Structured AI Transformation to Address Persisting Scaling Barriers
Industry consultants and regulators increasingly emphasize the need for structured organizational transformation to scale AI enterprise-wide and maximize returns on AI investments. The AI Playbook for Financial Services, developed in 2026 by the World Economic Forum (WEF) in collaboration with Accenture, outlines four priorities for successful AI transformation:
- Establishing an AI vision and strategy with board-level accountability.
- Building the technology, data, and governance foundations for responsible AI scaling.
- Balancing immediate efficiency gains with longer-term transformation.
- Preparing employees, processes, and organizational culture for AI-enabled operations.
Q3 2026 studies point to a close link between AI governance and maturity. Cerulli Associates found that wealth management firms with well-defined AI governance models score more than 2.5x higher on AI maturity than those without established governance approaches.
T. Rowe Price identified a similar relationship between formal AI usage policies and adoption. Its study found that firms with established AI policies report using AI 50% more frequently across business functions, suggesting that clear guidelines can foster adoption.
AI capabilities are also becoming increasingly important for talent retention. The survey by AssetMark found that 69% of financial advisors would consider changing their current employer if it fell behind competitors in AI capabilities. The figure increased from 55% among advisors managing less than $250 million to 78% among those managing $500+ million.
Higher AI maturity was also associated with stronger financial returns. LSEG found that firms with more mature AI capabilities achieve average returns on AI investments of 4.7%, compared to 3.1% among less mature firms.
Yet, ScienceSoft's Q3 2026 findings suggest that organizational change at investment firms is lagging behind AI adoption, and only a small fraction of firms have scaled AI enterprise-wide. Based on how wealth management executives described their firms’ progress in Cerulli Associates' study, 97% of firms have AI initiatives underway, but only 26% have AI in production and are actively scaling it, and just 1% have successfully scaled AI across functions.
The study's separate AI maturity assessment provided a broader view of firms' progress: among AI adopters, 50% remain in early stages, with limited formal governance, training, and measurable outcomes; 38% are gradually scaling AI, with some policies and controls established or underway but inconsistent outcome measurement; and 12% have deployed AI at scale, supported by established governance, data readiness, and employee training.
ScienceSoft's analysts suggest that the gap may stem from differing definitions of successful AI scaling: wealth managers may measure success by the breadth of deployment, while the maturity scoring model also factors in progress in governance and operational readiness.
Data readiness remained a persistent constraint. LSEG found that 50% of wealth management firms lack established processes to clean, normalize, and tag data or obtain trusted external data. As noted by ScienceSoft experts, these limitations complicate AI work and make it hard to deliver reliable results across connected investment operations.
At the same time, Q3 2026 industry studies indicated progress in preparing for enterprise AI transformation. In the LSEG survey, 77% of wealth management firms reported developing AI strategies and roadmaps, 69% were training or acquiring AI talent, and 67% were working on responsible AI governance policies.
Large investment firms began announcing structured AI transformation programs. State Street Corporation outlined plans to transition to a technology- and AI-enabled product platform operating model by 2029. Its program involves reorganizing employee work, modernizing legacy tools, expanding agentic AI use, and strengthening the enterprise data foundation. The firm expects around $1 billion in annual run-rate transformation benefits by 2029, with 75% coming from cost efficiency and 25% from revenue uplift.
State Street is a good example of how much more operational transformation can cost than AI technology. In its Q2 2026 earnings call, the firm estimated around $500 million in one-time transformation costs (roughly 4.5% of its annual fee revenue), primarily for severance and contract terminations. The absolute figure would be lower for midsize firms, but the revenue share gives a sense of the potential scale of investment.
A fair question I hear from clients is: how do we plan something this large without taking on too much risk? In ScienceSoft's engagements, we typically set a multi-year transformation direction while breaking technology implementation and business changes into smaller initiatives, each with its own budget, ROI forecast, dependencies, and risk assessment. This staging makes it easier to evaluate expected returns and approve investments.
A critical move is separating investments that enable AI from those expected to generate direct returns. For example, data modernization, governance, and workforce adaptation may not alone deliver attractive ROI, but they support profitable AI use cases. If you judge every foundational investment by direct impact, you risk cutting the very work needed to scale AI.
And don't lock in every transformation decision on day one. Over a three-year program, AI capabilities, vendor options, and regulatory rules will change. You need a clear business destination but enough flexibility to adjust the plan as new risks and opportunities emerge.
Executive Director of Solution Design & Competencies, Financial IT & AI Principal Consultant, ScienceSoft
References
- Sixth Annual Defined Contribution Consultant Study (T. Rowe Price, September 2026).
- More Than Half of Advisors Using AI Save 4+ Hours a Week, AssetMark Research Finds (AssetMark, September 15, 2026).
- Annual survey of asset managers 2026 (Acuity Analytics, 2026).
- The blueprint for growth: AI’s role in Wealth Management (LSEG, 2026).
- What Do Advisors Want From AI? (Advisor360°, August 28, 2026).
- State of Wealth Management AI Adoption (Cerulli Associates, September 9, 2026).
- Q2 2026 Trend Report: AI Use, Spend, and ROI in Wealth Management (F2 Strategy, 2026).
- Ameriprise Advances AI Innovation Across the Advisor and Client Experience with $1B Annual Spend on Technology and AI Capabilities (Ameriprise Financial, July 27, 2026).
- Wells Fargo introduces AI Teammate, bringing AI-powered support to the advisor experience (Wells Fargo, July 15, 2026).
- Insigneo Executive AI Assistant: Compliant Generative AI for a Regulated Wealth Management Firm (Wildnet Edge, September 28, 2026).
- LPL Financial’s Latitude Unifies Technology Built for the Future of Advice (LPL Financial, July 28, 2026).
- T. Rowe Price Works With Anthropic To Bring Claude To More Of Its Investment Process (T. Rowe Price, September 10, 2026).
- Millennium and Anthropic are building a digital risk analyst with Claude (Claude, August 6, 2026).
- SS&C Technologies Unveils Portfolio of AI-Enabled Enhancements; Debuts Six-Part Agentic Blueprint Docuseries at SS&C Deliver (SS&C Technologies, September 23, 2026).
- Anthropic Launches Claude for Financial Advisors (Yahoo Finance, September 14, 2026).
- Working at the frontier: How Balyasny Asset Management evaluates and governs Claude Fable 5 (Claude, September 17, 2026).
- AI Agents Streamline PIMCO Middle-Office Reconciliations (Oncorps, July 24, 2026).
- Mariner and Humanity Labs Announce Five-Year Partnership to Leverage an AI Workforce of 700 FTEs (Humanity Labs, July 14, 2026).
- Everest Group launches Innovation Watch on Agentic AI in Wealth Management Technology (Everest Group, July 21, 2026).
- Advisor360° Launches Meeting Prep, an AI Agent for Financial Advisors (Advisor360°, July 7, 2026).
- Cabot Wealth Management Becomes First Customer to Leverage Ridgeline Intelligent Outcomes (Yahoo Finance, September 1, 2026).
- NVIDIA Partners with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to Establish AI Compute Infrastructure Financing Platforms to Mobilize Over $500 Billion of Third-Party Capital (Blackstone, August 10, 2026).
- Financial Institutions Race to Scale AI as Trust and Governance Become Critical Tests (World Economic Forum, June 24, 2026).
- State Street (STT) Q2 2026 Earnings Call Transcript (The Motley Fool, July 23, 2026).
Version History
Browse earlier quarterly editions of our investment AI trends report:
