SEI is working with IBM to modernise financial operations with agentic AI, but the partnership begins somewhere less glamorous than an autonomous agent: a detailed review of existing work.
The financial infrastructure provider and IBM Consulting are examining SEI’s systems, data architecture and daily processes before deciding where intelligent agents belong. That sequence reflects a practical truth. Automation cannot create reliable value when it is placed on top of a broken pipeline.
Process discovery comes before deployment
SEI subject-matter experts are working with IBM to map current operations and identify where employees spend time on repetitive administration. The goal is to redesign processes and make targeted system changes rather than apply AI indiscriminately.
This discovery stage supports governance as well as efficiency. Teams can define which actions an agent may take, locate the data it needs and decide where human approval remains necessary. IBM Enterprise Advantage will provide the technical foundation for deployment, but the business map determines what should be built.
The source notes that automating standard queries and basic data entry can reduce processing times by as much as 40%. That potential is attractive, yet a percentage alone does not identify the right workflow. A fast process with poor data or unclear ownership may simply produce mistakes sooner.
SEI Chief Financial and Chief Operating Officer Sean Denham describes investment in how the company operates as being as important as investment in what it delivers. The stated ambition is to deploy AI with discipline, improve decisions and scale with confidence.
Clean data gives agents useful boundaries
Agents need more than a capable foundation model. They require accurate information, explicit permissions and a dependable connection to operational systems. In finance, every automated action must also fit regulatory and risk controls.
That makes data hygiene a financial issue. Inconsistent records and fragmented definitions can cause an agent to act on the wrong context. Governance establishes which source is authoritative, how an action is logged and what happens when information conflicts.
IBM financial-services leader Glenn Finch says the combination of SEI’s business knowledge and IBM’s process intelligence can create efficiency while embedding data-centric insight into daily work. The partnership model joins regulatory and operational knowledge with engineering capacity.
The intended role of agents is targeted. They can handle routine entries, standard questions and repeatable administrative steps, while people retain complex problem-solving and client responsibility. This is augmentation at the workflow level rather than a general promise to automate an occupation.
Productivity should improve the client relationship
Denham frames the workforce outcome around time. If teams spend less of it on repetitive manual work, they can spend more on relationship-driven service, professional development and decisions that require judgment.
That is a stronger test of value than counting agent interactions. An automated process can look productive while shifting hidden correction work to another team. SEI and IBM will need to measure consistency, exception rates and client outcomes alongside raw speed.
Operational resilience must remain the constraint. Financial systems cannot sacrifice traceability for autonomy, particularly when an agent updates records or influences a client-facing process. Monitoring and escalation paths should be designed before the first production action.
SEI’s method is deliberately unhurried: inspect, redesign, establish a data foundation and then embed agents where their boundaries are clear. It may lack the spectacle of a company-wide AI launch, but it addresses why so many enterprise programmes remain stuck in pilots.
Agentic automation creates value when it disappears into a better process. By starting with the process rather than the model, SEI and IBM are treating AI as an operating change—one that must improve both the economics of work and the experience delivered to clients.
