Search

See how SCS automated a week-long quarterly workflow with local AI. Read the case study.

SCS automated a week-long quarterly workflow with AI that never leaves the building

An independent wealth-management firm gave its operations lead back a week of work every quarter using a local AI application that keeps every record on the lead’s own Windows® PC.

Reduce costs

~200 hours/year

of specialist time saved

Compliance

100%

locally run on the employee’s hardware 

100%

of client data stays on the local machine, with no cloud or server needed

FullColor_SCS_Advisory

The challenge

SCS Advisory is an independent wealth-management and retirement-plan firm based in Bellevue, Washington. Like most firms in its category, SCS competes on advisor relationships. Motivating its team requires operational rigor, including the quarterly work of calculating advisor compensation across advisors and roles. Historically, the burden of performing these calculations has rested entirely with one person. 

Alisha Yonan, an operations lead at SCS for more than 15 years, owns many business-critical tasks, including end-to-end bonus calculations. There is no documented standard operating procedure (SOP) for this process; it lives entirely in Yonan’s head. The process has evolved over her long tenure, running across five Microsoft® Excel® spreadsheets plus a Microsoft® Power BI® dashboard with limited automation. It’s meticulous work, requiring manual calculation and constant navigation between tabs, culminating in a full week of Yonan’s time every quarter, and closer to a week and a half at year-end. That’s roughly 200 hours a year of a senior operator locked into a mechanical task. 

SCS Advisory President Brendan Sullivan joined in 2025 to accelerate the firm’s growth and move the firm forward. An open-minded leader, Sullivan engaged with Prowess Consulting to evaluate AI-based solutions. But the obvious 2026 answers were off the table: SCS’s policies had not yet been revised to allow cloud-based AI products regardless of vendors’ security postures, which meant that well-known cloud AI providers like Microsoft® Copilot®, ChatGPT®, and cloud workflow tools that permitted data egress into LLM environments simply weren’t going to work. A traditional software-as-a-service (SaaS) workflow-management tool the team had begun evaluating would have replaced one specialist burden with another, as it would still require Yonan to configure, touch, and service the workflow for every run. 

“SCS is an incredible group of talented and committed people, but they spend too much time on tedious work that technology should handle. My goal wasn't to replace our experts; it was too free them to apply their insight to grow the business and better serve clients.” 
Brendan Sullivan President, SCS

The SCS team needed to automate a process no traditional system could touch, and to do so in a way that respected data egress policy, kept the expertise with the expert, and eliminated tedium without eliminating judgement. The team needed a new path, not a trade-off; one that would unlock key people like Yonan and move the entire firm forward. 

“Every quarter I spend an entire week or more doing this. It is holding me back from taking on more strategic projects and advancing my own role and the business.” 
Alisha Yonan Operations Lead, SCS

The solution

Prowess Consulting built SCS Advisory a custom compensation-splitting application that runs entirely on a Windows® PC. It can be run on Yonan’s own workstation with no cloud connection, no server, and no dependency on any external AI services. The engagement used a frontier model exactly once, during design. All AI services thereafter run on small on-device models.

The extraction step came first. Prowess Consulting’s AI engineers worked with Yonan to capture the process using an audio-less video recording and a frontier model to derive an SOP in two paired versions:

  1. A natural-language rulebook that had never existed before, and that needed to be precise enough to explain the process to another person
  2. A matching formula set expressing each rule as an executable notebook-compatible expression

The same model produced the design specifications for the application itself.

Figure 1. The Commission Splitting solution for SCS is executed fully on-device, however frontier cloud models also play a role in development, analysis, and architecture.

The runtime is where the local models earn their keep. Two on-device language models sit alongside the app: Qwen 3.6 handles logic, reading an existing formula, understanding a requested change, and updating the formula. Qwen 3.5 NPU, a smaller model, handles the narrow linguistic task of rewriting the plain-English rule to match. Yonan types a change (such as “update Marcus Chen’s standard rate from 5% to 12%”), and the local system then updates the formula, rewrites the corresponding rule, and shows her both versions to confirm. Response time is less than a second, and Yonan never needs to open the code or file a ticket.

In fact, much of the application still operates like the trustworthy, battle-tested Excel formulas people already rely on. AI isn’t driving every part of the experience; it’s intentionally scoped to the tasks it’s uniquely good at: interpreting requirements, comparing logic, rewriting formulas, and improving written explanations. By using the smallest capable model for each of those targeted tasks, the application stays fast, accurate, and efficient enough to run on local hardware without compromising performance or reliability.

“I still understand every calculation that’s being made. The AI helps me update the rules, but I can review the logic before anything runs. That transparency was really important to me.” 
Alisha Yonan Operations Lead, SCS

The result

With this application, compensation splitting condensed from a full week of manual spreadsheet work per quarter to a one-click execution against the same source data with rules readable, updatable, and owned by the person who actually understands them. “For the first time in years, I’m not dreading compensation analysis and deployment week. Instead of spending days buried in spreadsheets, I can focus on the work that actually helps move the business forward,” explains Yonan.

Approximately 200 hours of Yonan’s time is saved per year, redirected to a broader body of operations work she is uniquely qualified to do. The sub-second response times on local hardware mean the on-device solution isn’t a slower compromise of a cloud tool; for the specific work being done here, it’s often faster.

Just as important as the time gain, nothing about the constraint that started this conversation had to change. The AI data policy is a non-issue because no data ever leaves the machine. The commission-splitting logic that relied on a single person’s expertise is now memorialized as a readable, maintainable specification, reducing key-person risk that had been an open question and scale limitation at SCS for years. And the tool runs where the work already happens: on the same Windows desktop as Microsoft® Word, Excel, and Edge®, with no VPN and no separate server environment.

What changed

  • A week of quarterly specialist time reduced to a one-click process, saving ~200 hours per year
  • Rule updates handled in plain English by the operations lead herself, with no developer in the loop
  • Institutional knowledge, previously undocumented, captured as a paired natural-language and executable ruleset
  • Every record kept on the local Windows machine, satisfying a parent-company cloud-AI prohibition that had blocked earlier options
  • A repeatable pattern established for applying local AI to other specialist-bound workflows at SCS

What’s next

With compensation splitting running smoothly, SCS Advisory is evaluating other similarly shaped workflows: quarterly reporting, advisor onboarding, and any other processes where tribal knowledge meets manual spreadsheet work. With the pattern validated, Prowess Consulting and SCS expect to move even more swiftly when implementing subsequent solutions. More broadly, the engagement gives SCS Advisory and Prowess Consulting a proven model for wealth firms, registered investment advisors (RIAs), and credit unions operating under similar cloud-AI data restrictions: a way to say yes to AI adoption on terms that a compliance-sensitive organization can actually approve.

 

About Prowess Consulting

 

Prowess Consulting helps organizations adopt agentic AI that works with existing systems, not against them. With over 15 years of experience in AI, automation, and data—and deep roots in financial services and enterprise—we reduce time-to-value from months to weeks and empower teams to lead the change.

Learn how Prowess can drive measurable impact in your content and marketing operations

Learn more about Prowess marketing services

Prowess and the Prowess logo are trademarks of Prowess Consulting, LLC.

Copyright © 2026 Prowess Consulting, LLC. All rights reserved.

Other trademarks are the property of their respective owners.

Related Posts

Are you ready to expand your reach?

Get in touch today to discuss how Prowess can make a real difference for your enterprise.