AI in Breast Health.

When the person who maintained Iowa Radiology's pick-list code left, a critical workflow was at risk. See how the team rebuilt it with AI, returned four hours a week to staff, and kept patient letters moving.

This episode is available now. Running time 20 minutes.

Premiered Wednesday, October 14 Available on demand now · 20 min
Featured Speakers
Sarah Agan
Sarah Agan
Radiology Clinic SupervisorIowa Radiology
Nick Avossa
Nick Avossa
VP of AI OperationsAbbaDox
The super user problem
What failed when one person left.
From rigid to resilient
How AI supports natural dictation.
Four hours returned each week
How the team removed repeat work.
Clearer patient communication
How better letters reduced confusion.

Free · On demand · 20 min

Watch now

One sign-up. The full conversation with Iowa Radiology.

Available on demand. Register and watch right away.
The Numbers

Half a day back, every week.

One eight-hour workday Each block = 1 hr
Before 4 hrs of rework
After Handed back

A dozen or more patient letters came back wrong every week. Twenty minutes to fix each one, an addendum filed every time, and every letter proofed by hand before it went out.

20 min to correct one mis-dictated report
12+ patient letters fixed each week
1 hr mammogram to results in the portal, now
The result

Half a day

of tech and radiologist time handed back, every single week.

4 hrs weekly  ·  200+ hrs a year

What You'll Learn

One word should not break a patient letter.

Laterality, recommendation, and timeframe had to be dictated in the right order. If one word was wrong, staff fixed the letter by hand.

Illustration: a simulated report read-out. Four mammography reports are dictated in different words and different orders. Each time, the same four data points, laterality, recommendation, follow-up timeframe, and BI-RADS category, are extracted correctly, with no addendum required.

01

When the super user leaves

How Iowa Radiology rebuilt a workflow that depended on one person.

02

Pick lists versus natural dictation

How AI captures key details without rigid phrasing.

03

The cost of small fixes

How corrections, addenda, and IT tickets added up.

04

The BI-RADS 0 follow-up problem

How automated outreach and portal access replaced manual mail runs.

Previous Episodes

More TechTalks

Ep. 11

Episode 11 · On Demand

The real ROI of AI: throughput that doesn't wait on staffing

How Capitol Imaging Services uses AI for order intake, scheduling, and follow-up across more than 60 locations.

Watch now
Ep. 10

Episode 10 · On Demand

Conversational AI: confirmations without the phone tag

How Voice AI confirms appointments, handles reschedules, and refills same-day cancellations.

Watch now
Available on demand

Watch the episode.

See how Iowa Radiology rebuilt a workflow that depended on one person.

Watch now
Free · 20 minutes · On demand

See it on your own workflows