Person
1 episodes · 8 quoted claims
Factories have used digital twins, virtual representations of production lines connected to live sensors, for around twenty years.
“Factories have been doing this for twenty or so years, uh, in a factory setting, a digital twin is is a virtual representation of the production line”
In factories, digital twins are used for predictive maintenance and bottleneck identification.
“In factories, what you can do is predictive maintenance. You can spot where a piece might be wearing out, you can identify bottlenecks”
In Webanck's loan example, a digital twin agent would detect a failed income validation control and block disbursement for affected applications.
“In the twin, an agent would detect the failed control and block disbursement of the loan for all the affected applications.”
A risk and control self assessment is an exercise, usually annual, in which a business area identifies what could go wrong in its processes and checks its controls.
“A risk and control self assessment is, uh, an exercise. It's usually done once a year, but ideally more often”
In one client pilot, McKinsey identified almost 200 concentration risks, including one provider supporting more than thirty processes.
“in in one pilot with a client, we identified almost 200 concentration risks, and we identified one provider that was supporting more than thirty processes”
Today a bank may run only about eight tests or tabletop exercises a year to assess performance under stress, per Webanck.
“maybe you do eight tests or tabletop exercises a year to know how you would perform in a stress scenario.”
Webanck said early agents can be Python code on top of existing systems, without a data lake or new system.
“It's some Python code right now that sits on top of your other systems. You don't necessarily need a data lake, you don't need to build a new system.”
A challenger RCSA has AI agents build a risk and control self assessment in parallel to the existing one and compare the two to find discrepancies.
“Gen AI can pressure test with a what we call a challenger RCSA, where you provide some of the raw information and agents, then build an RCSA sort of in parallel”