His big idea is to create "frequent commuter programs" in which people who travel on public transit would be rewarded for patronizing the system varying amounts depending on when and how far they travel. Prabhakar thinks the system could help create greater public transit usage and simultaneously decrease congestion. And he's deploying behavioral economics to transform the small monetary rewards a city could offer into something more. They tried a pilot program with Infosys in Bangalore and are rolling out a larger program with Singapore soon.
I think this is a promising idea and I know of a few other studies underway along similar lines. For crowded systems this can benefit all riders by spreading ridership over a longer period of the day, which should improve service on the edges of rush hour and off-peak through increased demand. This also reward many low-income riders who do not work 9-5 and would disproportionately benefit without changing travel behavior or taking service away from other riders.