Debanjan Saha says some people call AI agents silicon agents and human employees carbon agents.
“some people call it the silicon agent versus the carbon agent, which is the human employees, right?”
Person
Debanjan Saha
1 episodes · 21 quoted claims
Debanjan Saha describes personal productivity agents, or copilots, as helping with tasks such as email and running scripts.
“personal productivity agents. I mean, these are people who are your co-pilots and they help you with your personal productivity, whether it is email or whether it is running some script”
Debanjan Saha says line-of-business agents inside applications such as Salesforce, ServiceNow and Workday make those applications more user-friendly through chat interfaces.
“These are agents which are in Salesforce, in ServiceNow, in Workday, a and in many ways, they are essentially making those application a lot more user-friendly”
Debanjan Saha uses the term coordination tax for the cost of coordinating work across siloed enterprise applications and departments.
“one of the problems that we see in enterprises, what I call the coordination tax.”
Debanjan Saha gives a consumer packaged goods port-closure scenario to illustrate cross-department coordination that agents could automate.
“let's say you are in in a CPG industry and there is a port closure somewhere, and you have to figure out that what are the customers who are going to be affected?”
DataRobot built an agent for its oil and gas customer Chevron, according to Debanjan Saha.
“we recently did, uh, you know, one of those agents with one of our, uh, oil and gas customer, Chevron.”
Debanjan Saha says Chevron uses drones and robotic dogs to detect leaks in pressurized pipelines instead of sending humans.
“they have a set of drones and a and digital dogs that they send in order to sniff out where there may be a leak.”
Debanjan Saha says drones that approach a leak can explode because of their lithium-ion batteries, so digital simulation is used to identify no-fly zones.
“if there is a leak and they go too close to it, they have lithium ion batteries and it can blow up.”
Debanjan Saha says DataRobot treats agents like human employees, with a lifecycle of interviewing, evaluating, hiring, onboarding, performance management and decommissioning.
“you interview them, you evaluate them, you hire them, you then onboard them, and then you performance manage them, and then, you know, sometimes they leave and you decommission them, right?”
DataRobot is working with a customer to integrate agents with Okta so that each agent has its own identity.
“we are working with a customer where we are integrating these agents with Octa, so that every agent has an identity”
Debanjan Saha says every agent action is traced and monitored so mistakes can be corrected for future runs.
“every action that they take, we trace them and monitor them. So that if they make a mistake, you can actually go and correct that, and the next time they do the job, it's actually corrected.”
Debanjan Saha says that when an agent is decommissioned, someone must take over its work, context and memory, as when a human employee leaves.
“when you decommission an agent, like the way when a human employee leaves, somebody has to pick up that work, that context, that memory, and you have to do that, right?”
Debanjan Saha describes insurance claim processing where customers upload video of car damage and agents analyze it, match it to policies and schedule repairs.
“Today, you have customers who will take their camera and then, uh, go and take a video of their car if they're in an accident and upload that, and lot of that is going to be automated.”
Debanjan Saha says enterprise workflows with human and digital employees start with multimodal perception, such as scanning forms, reviewing videos or analyzing images.
“there is a multimodal perception that comes first, right? Whatever you are trying to do, there is a perceptive step”
Debanjan Saha gives a hurricane-driven shipping reroute as an example of feeding signals into a digital simulator to evaluate alternative paths.
“You have picked up some signals that there is a hurricane coming, and the shipping line is probably have to reroute or a port is going to close, right?”
Debanjan Saha argues that medicine is crude because compounds are tested on about a thousand people and then applied to roughly 8 billion people.
“we kind of come up with some chemical compounds, try it on thousand people, and then apply it to 8 billion people on on earth, right?”
Debanjan Saha envisions Netflix generating personalized content on request based on a viewer's mood.
“You can tell Netflix, this is what I am in the mood for, and they will cook something up for you and show it to you, right?”
Debanjan Saha says accountability and attribution are foundational to governing a hybrid workforce of humans and agents.
“foundational to this, by the way, is accountability and attribution.”
Debanjan Saha says actions in a hybrid workforce may be taken by humans, by agents on behalf of humans, or by agents acting autonomously, and each must be attributable.
“sometime people are going to take action, sometime agents on behalf of the people are going to take action, and sometime agents autonomously are going to take actions, right?”
Debanjan Saha says agents must follow domain-specific and jurisdiction-specific policies and produce artifacts that maintain an audit trail.
“there are domain-specific and jurisdiction-specific policies that the agents need to know, and sometimes they have to produce artifacts which can maintain an audit trail”
Debanjan Saha says DataRobot's customer uses Okta as an agent identity management system to control agents' access privileges and shut them down.
“using Okta as an agent identity management system. That is one place from where you can control the access privileges an agent can have to various different internal information”