A founder answers

How do you build trust when your AI touches sensitive systems?

Keeyu touches sensitive APIs — Shopify, ERPs, warehouse data — so Jevon's team rolls out each workflow carefully: do internal dev and testing in their own environments, then run "5 5 workflows" with the customer watching, and only "increase and expand" once the customer is comfortable.

The full answer

JR
Jevon Le Roux · Keeyu
EP 21 · Co-founder, Keeyu
Show notes ↗

Keeyu touches sensitive APIs — Shopify, ERPs, warehouse data — so Jevon's team rolls out each workflow carefully: do internal dev and testing in their own environments, then run "5 5 workflows" with the customer watching, and only "increase and expand" once the customer is comfortable.

More from this episode

Jevon describes the product as semi-autonomous in a narrow sense: they've given the agent a clear start line, finish line, and parameters to work within, but "we haven't fully automated it" because the tech is "not 100% there." Right now it's "very user in the loop" — the user participates, can see the workflows, and can intervene wherever necessary. They'll "know instantly from our retailers if there isn't accuracy because it's gonna go wrong," so they're deliberately in a trust-building phase.