Posting Assistant
AI-assisted insurance payment posting workflow that kept specialists in control.
Wisdom · 2025 · Lead Product Designer
- AI Workflow
- Research
- Healthcare SaaS
- 40% — Manual time reduced
- 40+ — User interviews, this product
- 3 — Workflow stacks shipped
Challenge
Insurance payment posting was chaotic — documents came in every format, specialists jumped between screens comparing EOBs, and every posting required manual judgment calls. Slow, error-prone, and impossible to scale.
Approach
Instead of jumping to full automation, started with an assistive workflow: OCR + LLMs extracted key data from scanned insurance documents and surfaced it in a structured review interface. Preserved specialist control while reducing cognitive load. Used OBS Studio to run observational research across multiple monitors. Once trust was established, integrated with Open Dental for deeper automation.
Outcome
40% reduction in manual posting time. Transformed a messy multi-screen workflow into a structured AI-assisted tool. Created a path to scale without adding headcount, and a strong foundation for full automation in the next phase.
Posting Assistant was built to help Wisdom’s revenue specialists process insurance payments faster and with less manual work. We used OCR and LLMs to parse scanned EOBs, payment docs, and denial details, then surfaced that information in a structured workflow that made posting easier and faster. Later, we integrated with Open Dental, which helped reduce manual posting time by 40%.
Key impact
- Reduced manual posting time by 40%
- Conducted 40+ user interviews for Posting Assistant
- Turned a messy multi-screen workflow into a structured AI-assisted tool
- Helped create a path from manual assistive tooling to deeper automation

The challenge
Payment posting was one of those workflows that looked simple from the outside, but once you really got into it, it was messy and time-consuming.
Insurance documents came in all kinds of formats. Some were clean, some were hard to read, some had missing context, and a lot of them required interpretation. Revenue specialists had to look through the document, figure out what happened, understand why something was or wasn’t paid, and then manually enter that information into another system.
It was a very manual process, and it also required a lot of attention to detail. People were constantly switching between screens, comparing information, and making judgment calls. That made the work slower, harder to scale, and more mentally draining than it needed to be.
We saw an opportunity to make that whole process better by extracting the important information up front and giving specialists a workflow that supported the way they already worked.
My role
I led the design of Posting Assistant end to end, from discovery and workflow definition through prototyping, interface design, and iteration. I worked closely with product and engineering, conducted user research, and helped define how we would measure the workflow before and after launch.
I also set up observational research using OBS Studio so we could watch revenue specialists work across multiple monitors in their real environment and understand where time and effort were being lost.
Approach
Instead of jumping straight to full automation, we started with an assistive workflow. OCR and LLMs extracted key information from scanned insurance documents, and the product surfaced that information in a way revenue specialists could quickly review and act on.
This approach helped us:
- reduce cognitive load
- support trust in AI-generated output
- preserve specialist control in a complex workflow
- create a foundation for deeper automation later
Once the workflow proved value, we extended it through Open Dental integration, which removed more manual work and drove a meaningful time savings.


Impact
Reduced manual posting time by around 40%
The biggest outcome was the reduction in manual posting time. Once we introduced the product and later integrated with Open Dental, we were able to cut manual posting time by about 40%.
Made the workflow easier on the team
The product reduced how much specialists had to manually read, interpret, and re-enter on every task. Instead of starting from a blank slate every time, they were working from structured information that was already surfaced for them.
Created a path to scale without adding headcount
By making the workflow more efficient, Posting Assistant helped the team handle more volume without relying on more manual effort to get there.
Built a stronger foundation for future automation
The first version helped prove the value and establish trust. That made it possible to keep improving the experience and automate more over time.