Kaagam

We run a bio-engineering company that operates at the boundary between medicine and personal care.

Most of our customers are also someone's patient. Their doctor handles what is going on inside the body. We handle what reaches the skin. The space between the two is where we work. Reaching it takes awareness-driven research no other company has been willing to do.

A patient with trimethylaminuria has a defective FMO3 enzyme. They cannot break down trimethylamine, and it leaves the body through their skin, breath, and urine as a fishy smell no soap can wash off.

The apocrine glands produce a different kind of sweat when hormones shift. The smell that follows cannot be stopped by any deodorant on the shelf.

As the skin ages, it produces 2-nonenal, a waxy compound that lodges in the lipid layer and brings the smell back within hours of every shower.

Bacteria living inside the hair follicle have learned to hide from the antimicrobial chemistry the deodorant aisle was built on. Every product on that aisle slowly loses against them.

We mapped more than a dozen of these pathways and built a sequence of formulations to address each one. For some customers we recommend one product. For others a combination. For a small number we tell them which of our own products would make them worse, and exactly why.

Every formulation we ship traces back to peer-reviewed research. Every claim has a source. We read at the depth their specialists read, often deeper than the consultation has time for, because anything shallower puts a guess on their body.

In our line of work, we read research papers cover to cover. Each biochemical pathway has its own body of literature, often thousands of papers deep. We turned to AI to get through them faster.

Every tool we tried fabricated answers. One pointed us to papers that did not exist. Another gave us DOIs that resolved to nothing. A third produced a reference list that looked authoritative, until the third citation in it failed to open. Verifying every answer by hand took longer than reading the literature ourselves. A company built on what is verifiable cannot run on what is invented.

Neither can a doctor.

That was the second realisation. Our customers' clinicians were drowning in the same literature, with the same speed pressure, with the same broken AI tools, with the same impossible verification burden. The personal care answer for a customer with TMAU exists in dozens of papers. The clinical answer for a doctor managing the same patient exists in thousands. The same broken tools were failing both of us. The cost of getting it wrong was much higher on the clinical side.

So we built Kaa.