S1
HR & operations
Top 3
What changed in HR systems, compliance and workforce data for a team under 500 people, and what it costs or saves.
Operations
The EEOC recovered more than $97,000 in two pregnancy discrimination conciliations announced September 17. Vital Systems paid $57,000 after firing a worker who brought a note barring radiation exposure, and Condor Club paid $40,000 for cutting shifts.
$97,000
recovered in two EEOC pregnancy conciliations
Why this mattersThe settlement amount is the small part. The real cost was a manager deciding what a pregnant employee could handle. Route every medical note to one trained person, record the restriction and the accommodation offered, and keep managers out of the medical judgment.
Operations
LRN surveyed about 2,000 full-time employees and found fewer than 1 in 10 codes of conduct address AI or technology ethics. Separately, HR Acuity found 55% of employees experienced or witnessed misconduct in 2025, up from 41% a year earlier.
1 in 10
codes of conduct that mention AI at all
Why this mattersStaff are already pasting company data into chat tools, and nothing on file tells them what is allowed. Add three lines to the code of conduct: which tools are approved, what data may never go into them, and who to ask. One page closes most of the gap.
Operations
The Labor Department removed the 7% disability representation goal for federal contractors and dropped the requirement to collect disability status from applicants and employees, effective September 21. Nondiscrimination and accommodation duties stay.
Sept 21
date the 7% contractor disability goal ends
Why this mattersIf you hold a federal contract or subcontract, one reporting burden goes away while the accommodation duty stays exactly where it was. Check which forms your ATS still collects, turn off the ones no longer required, and leave the accommodation process untouched.
S2
Science & technology
Top 3
Research and technology that is making money, saving money, or about to change how work gets done.
Science
MIT CSAIL researchers Vivek Gopalakrishnan and Polina Golland, with clinicians at Harvard Medical School and Boston Children's Hospital, built xvr, which aligns live surgical X-rays to a patient's prior CT scan to sub-millimeter accuracy in seconds. It trains on scans from over 2,000 patients, then adapts to one patient in about five minutes, published in Nature.
5 min
to tune the model to one specific patient
Why this mattersThe team stopped chasing one model that works for everyone and tuned a copy per case. HR tooling usually fails the same way, with one onboarding flow for sales, warehouse and engineering. Take your slowest process and build a second version for the group it fits worst.
Science
Spencer Axani, now at the University of Delaware, built CosmicWatch as an MIT graduate student: a muon detector roughly the size of a cracker box that costs about $100 to assemble. Thousands have been built in the eight years since, and the third version was documented in the Journal of Instrumentation.
$100
to build a working desktop muon detector
Why this mattersCapability that used to require a budget and a specialist keeps collapsing into something anyone can buy. The same curve is running through payroll, benchmarking and policy drafting. Before you approve the next specialist invoice, check what the same job costs as software today.
Science
Chunlei Guo's group at the University of Rochester's Institute of Optics used laser-treated black metal panels to evaporate seawater with sunlight, pulling dissolved salts out as solids instead of toxic brine. The team recovered about 50% of the lithium from those salt deposits, published in Light: Science & Applications.
50%
of lithium recovered from the leftover salt
Why this mattersThe waste was the product once someone looked at it closely. Every company has a version of this sitting in the HRIS: exit reasons, time to fill, absence patterns nobody reads. Pull one of those reports this month and see what it tells you about next quarter.
S4
Three things to do this week
3 actions
Three moves with the best payoff for the least effort, each with what it is roughly worth. Even one of them saves hours of work, cash, or a risk you would rather not carry.
🛡️1
Risk
Send every medical note to one trained reviewer instead of the manager
Name one person who receives medical notes and accommodation requests, tell managers to forward instead of deciding, and log the restriction, what was offered and the date. About 90 minutes.
PayoffThe EEOC collected more than $97,000 from two small employers where a manager overrode a doctor's note. Defense fees on one charge commonly reach tens of thousands before any payout, and a clean log is the first document your counsel asks for.
$50k+
defense and settlement cost on one charge
Effort 90 minutes
📊2
Data
Add three lines on AI to your code of conduct
List the approved AI tools, name the data that may never be pasted into them, such as pay, health and candidate records, and say who to ask. Send it to everyone and add it to onboarding. About one hour.
PayoffFewer than one in ten codes of conduct mention AI, so staff are guessing with customer and employee data. One page of rules is the cheapest way to avoid a data incident, where legal review, notices and monitoring run into thousands for a small employer.
$10k+
data incident response cost avoided
Effort 1 hour
⚡3
Speed
Require whoever sends AI-drafted work to explain every number
Add one line to your AI guidance saying the person who sends the work owns it and must be able to defend each figure. Ask for that in the next three reviews so it is real. About 45 minutes.
PayoffClose to three in ten workers admit shipping work they could not explain. Rework, a wrong figure in a client deliverable and the manager time spent checking it all land on the same payroll. One line makes review cheaper than cleanup.
40+ hrs
of rework and review time saved in a year
Effort 45 minutes
Want this handled, but no time or capacity?Olga takes on HR ops projects like these on a fractional or interim basis. A 30-minute call is enough to scope one.
Reach out to Olga ↗Payoff figures are order-of-magnitude estimates for a company under 500 people, based on typical costs, not quotes for your business.