Undated
Ajay Swamy says he has seen banks process millions of transactions every hour without error.
“I've seen banks process millions of transactions every single hour flawlessly.”
Topic
18 facts from 1 episode, in date order. These are what speakers said on the shows, each linked to the moment it was said. They aren't independently verified.
Undated
Ajay Swamy says he has seen banks process millions of transactions every hour without error.
“I've seen banks process millions of transactions every single hour flawlessly.”
Undated
Swamy says bank workflows touch separate core, trading, document, risk and SharePoint systems, producing data fragmentation.
“There's one workflow that touches a core system or a trading system.”
Undated
Swamy says document-processing AI needs a golden corpus of evaluation documents to measure accuracy confidence.
“you also need some sort of a golden corpus of eval documents where you can evaluate and say, hey, this is actually giving me to a state of confidence, accuracy, confidence that I'm comfortable with.”
Undated
Swamy recommends a human-in-the-loop check, following the four eyes principle, for AI document processing.
“you also need to plug in some level of a human in the loop so they can double check, sort of like the four eyes principle.”
Undated
Swamy says a bank's loan rejection decision must be explainable as a regulatory principle.
“a bank rejects somebody for a loan. Now that decision has to be explainable, and that's that's a regulatory matter of fact principle.”
Undated
Swamy says frontier AI models are becoming commoditized month by month.
“the models that they're using underneath, the frontier models, et cetera, they're all getting commoditized on a month by month basis.”
Undated
Swamy says any model in a business-critical process must be auditable end to end, whether machine learning or generative AI.
“any model, whether it's machine learning or GEN EI, that sits in a business critical process, it has to be auditable end to end.”
Undated
Swamy says he rejects software that cannot show how its decisions were reached.
“So if your software can't show that, then it's a no.”
Undated
Swamy's startup FundLens.ai learns how a fund operates based on its philosophy, strategy and risk posture.
“For my startup FundLens.ai, what we do is that we learn how a fund operates based on its philosophy and its strategy and its risk posture.”
Undated
Swamy says a trading model carries much higher risk than a model that extracts invoice information.
“a model that's being used in trading is going to be highly, highly more riskier than a model that's being used to like. Maybe extract invoice information”
Undated
Swamy says that in his career deployed systems have missed amounts and triggered rules that should not have fired.
“There have been a couple of examples across my career where things have gone south, where amounts have been missed.”
Undated
Swamy predicts 'tokenomics', meaning the return on investment from AI token usage, will become important.
“There's going to be this notion of tokenomics, where the amount of tokens you're using and how valuable it is, right?”
Undated
Swamy predicts that institutions which rebuild workflows and governance controls will win over the next couple of years.
“the institutions that will win over the next couple of years are the ones that have rebuilt their workflows and their governance controls”
Undated
Swamy says product managers on his team no longer write requirements documents from scratch and use AI instead.
“There's product managers on my team who don't write from scratch anymore.”
Undated
Swamy says AI does roughly 70 to 80 percent of the heavy lifting in drafting requirements documents.
“Because that'll do seventy to 80% of the heavy work.”
Undated
Swamy says coding assistants help teams deliver faster with higher quality when grounded with manifest files and skills.
“you will be able to deliver faster with more quality, as long as you sort of ground the A I with the specifics manifest files and the skills”
Undated
Swamy argues financial operations should be made defensible at speed, saying the bottleneck was trust rather than speed.
“I think you need to make them defensible at speed. So the bottleneck was never really the speed.”
Undated
Swamy advises engineering trust deliberately into processes rather than trying to automate it.
“Don't try and automate trust, engineer it deliberately into your processes, make sure that's grounded”
Episodes