An AI co-pilot that took a retail trading platform upmarket to fund managers.
StratX is Motilal Oswal’s in-house algorithmic F&O trading platform, built for retail traders: a marketplace of exchange-approved back-tested strategies, plus no-code and Python strategy builders.
Fund managers running HNI and UHNI mandates do a different job. They are not looking for a strategy to enter. They are running many strategies across many mandates and need to know how the book is performing, which positions have drifted, and what changed since they last looked. StratX did not answer those questions, so the managers paid for a separate third-party tool that did.
A colleague started Night Owl to close that gap. I joined the initiative and led it as techno-functional lead, from architecture through to deployment on AWS.
The two audiences generate different workloads.
A retail trader asks one question at a time and waits for the answer. The interaction is short, unpredictable, and worth optimising for latency and cost.
A fund manager works to a calendar. The review that matters is the one already waiting when they sit down at the end of the day, the week, the month or the quarter. Nobody watches a spinner while a quarterly review is written.
Night Owl runs on two tiers because of that split. Gemini 2.5 Flash answers on demand, where response time is what the user notices. Gemini 2.5 Pro runs the scheduled end-of-day, end-of-week, end-of-month and end-of-quarter reviews, covering both strategy performance and platform health. Those jobs run unattended, so they can use the slower and more capable model without anyone waiting on it.
Splitting the work this way meant the expensive model was only ever spent where its output was read carefully.
Night Owl displaced the third-party performance-tracking tool the managers had been using, and moved that reporting inside the platform where the positions already live.