For a lot of beef producers, in vitro fertilization (IVF) and embryo transfer (ET) fall into the “we tried that once” category — big promise, big bill, disappointing conception. Stories of 15% to 20% pregnancy rates still echo at sale barns and producer meetings.
For more than 50 years, embryo quality scoring has relied on a simple recipe — a microscope, an experienced embryologist and a visual grading system built largely on gut feel and tradition. If the embryo “looks good,” it gets loaded into a straw and placed into an expensive recipient. If it looks rough, it gets discarded — or maybe frozen and sold cheap.
During episode 23 of the Future of Beef Show, reproductive specialist and EmGenisys founder Cara Wells argues this approach is leaving real money – and pregnancies – on the table.
“We know that embryologists are unintentionally selecting to transfer dead and dying embryos into recipient animals each and every day,” Wells says. “About 15% to 20% of embryos that are typically transferred today have a very low, close to 0% chance of ever making a pregnancy.”
Her team’s answer: use short videos, motion analysis and machine learning to quantify embryo viability. By combining video, AI and a massive dataset of more than 60,000 embryos, EmGenisys is giving labs and veterinarians a number — not just a grade — to decide which embryos are worth a $400 transfer and an increasingly expensive recipient cow.
Why Embryo “Looks” Aren’t Enough
Conventional ET and IVF both aim at the same goal — multiplying elite genetics from both the bull and the cow.
- Conventional ET (in vivo): Donor cows are superovulated, bred via AI and embryos are flushed from the uterus and graded before transfer or freezing.
- IVF (in vitro): Eggs are aspirated from the ovaries via ovum pickup (OPU), fertilized in a lab, grown for seven days, then transferred or frozen.
Wells reports across the U.S., IVF yields about a 42% live calf rate. Conventional ET typically runs closer to 50%, in part because those embryos have undergone less manipulation.
From a producer’s perspective, that means 40% to 60% of transfers fail — yet each embryo represents real cost.
On top of that, the International Embryo Technology Society (IETS) grading system is inherently subjective. Embryos are scored on:
- Stage (1–9): how far along development is
- Quality (1–4): based on fragmentation, color, and cell symmetry
Wells explains “beautiful” Grade 1s are assumed to be superior. Grade 2s are often second‑class citizens. But on the podcast, both the hosts and Wells point out reality in the field: “I haven’t seen that much difference in pregnancy rate between the two groups,” Future of Beef Show Host Jim Johnson notes.
Wells adds, “We’ve found beautiful Grade 1 embryos that scored very low in our system and did not make pregnancies. And we’ve seen lower‑quality embryos with high scores that did make pregnancies.”
In other words, some embryos that look perfect are functionally dead. Some that look ugly are quietly capable of producing live, profitable calves.
Turning Invisible Biology into a Number
Wells’ journey to tackle this problem started in an undergraduate research lab at Texas Tech, deepened while she worked on her doctorate in reproductive physiology and crystallized after a TED Talk. She watched a presentation on Video Motion Magnification, an MIT-developed tool that can amplify tiny movements hidden in a normal video. In one demo, a man’s face looked normal — until the filter was applied and you could see his skin pulsing red and blue as blood flowed and his heart beat.
“Everything goes back to embryos for me,” Wells laughs. “I thought, an embryo is a rapidly growing organism. Cells are dividing and differentiating, but even with the best microscope I’ve ever used, I can’t see that activity in real time. What if we used video analysis to study those cellular dynamics and mine out signatures of life?”
Working with Texas Panhandle ET veterinarian Dr. Russell Killingsworth, Wells began capturing short videos of cattle embryos, running them through the motion filter, and then tracking which ones made pregnancies at 35 to 60 days.
Early on, she literally sat at her kitchen table during the COVID-19 shutdown, measuring embryo diameters every five seconds and treating each embryo “like a clock face” to quantify how its shape changed over time.
What emerged was a clear pattern: Extremely “fast” and extremely “slow” embryos rarely made pregnancies. Embryos with moderate, “bell‑curve” activity were far more likely to stick.
Over the last six years, EmGenisys has scaled that kitchen‑table science into a machine‑learning platform trained on more than 60,000 embryos with known pregnancy outcomes. Today, when a new embryo video is uploaded, the system:
- Analyzes micro‑movements and development patterns frame by frame.
- Compares them to known pregnancies and failures.
- Produces a pregnancy‑probability score from 0 to 100, called the Envision score.
- Provides an embryo sex prediction (currently about 80% to 85% accurate).
Wells says crucially, this is independent of traditional 1 to 4 quality grades. The software isn’t judging appearance; it’s quantifying function.
How It Fits Real-World Workflows
Wells says this isn’t a “buy a $50,000 machine” pitch. It’s designed around what veterinarians and embryologists already use. EmGenisys is a software‑only platform — no custom microscopes or hardware. The embryologist or veterinarian mounts a phone to their existing microscope. They record a 30‑second video of one or more embryos. They upload through a web app — emgenisys.app — in a browser. Within about two minutes, the report returns Envision (viability) and Engender (sex) scores.
According to Wells, the immediate value is simple but powerful: “About 15 to 20% of embryos that are typically transferred today have a very low, close to 0% chance of ever making a pregnancy,” Wells says. “Just the simple elimination of the transfer of those embryos can improve pregnancy rates up to about 15% or 20%.”
For a producer, that can mean:
- Fewer wasted recipients on “dead on arrival” embryos
- Better use of high‑genomic or high‑value embryos, by prioritizing those with the highest scores
- The ability to set thresholds based on goals, for example:
- Only put embroys with more than 60‑score into your most expensive recips
- Transfer everything above “dead” for ultra‑elite genetics where every calf matters
Wells reports some IVF labs have implemented the process and some breeders are already selling embryos with Envision scores attached, using third‑party validation to justify dropping warranties. If the buyer doesn’t get a pregnancy, they can no longer blame an inherently non‑viable embryo — the risk shifts back to recipient management and transfer technique.
The Bigger Picture: From Experiment to Routine Tool
Ultimately, Wells believes this kind of data‑driven selection can help move IVF and ET from a “risky experiment” to a routine breeding tool for both seedstock and commercial herds.
“Today producers are very hesitant to adopt IVF and embryo transfer because of our historically low pregnancy rates,” she says. “When we improve outcomes, it grows the entire industry. It makes embryos more of the norm as a breeding strategy because it’s more practical, more affordable, and higher ROI than it is today.”
If improved selection can improve national pregnancy rates even 5%, Wells estimates that represents around $148 million in added value to the U.S. cattle industry — before accounting for any increase in the number of embryos used as confidence grows.
If a producer is considering IVF or ET — especially as part of herd rebuilding or accelerating elite cow families — the question isn’t just, “Can I afford this?” With better embryo selection tools on the table, the more important question may be: Can you afford to keep putting dead embryos into live cows?
