Sabastian Niles says Salesforce has focused for about 26 years on delivering five core values.
“we have for about 26 years now have really focused on how do we deliver on five core values”
Person
Sabastian Niles is a Salesforce executive who appeared on the Analyse Asia podcast with Bernard Leong to discuss AI trust, guardrails and agentic AI. He said he oversees Salesforce's Office of Ethical and Humane Use of technology.
Niles defined agentic AI as agents, supervised by humans and organizations, that can act, reason, plan and execute tasks. He said guardrails for AI agents come in preventative, runtime and self-improving kinds. He also said enterprise-grade agentic AI needs transparency, explainability and auditability. He listed Salesforce's five core values as trust, customer success, innovation, equality and sustainability.
1 episodes · 24 quoted claims
Sabastian Niles says Salesforce has focused for about 26 years on delivering five core values.
“we have for about 26 years now have really focused on how do we deliver on five core values”
Niles lists Salesforce's five core values as trust, customer success, innovation, equality and sustainability.
“number one value is trust. Then we have customer success or customer stakeholder success. Then innovation and then equality and sustainability.”
Sabastian Niles says his career before joining Salesforce lasted almost 20 years.
“I had one career for, I don't know, it was almost 20 years or so”
Niles says a high school article about his plans stated that he wanted to be a high-tech corporate lawyer.
“it says that I want to be one day a high-tech corporate lawyer”
Niles defines agentic AI as agents, supervised by humans and organizations, that can act, reason, plan and execute tasks.
“AI agents under supervision, control, direction of the organization, of the teams, of the humans, you know, able to act, right? To reason, to plan, to execute tasks”
Niles says Singapore Airlines is partnering with Salesforce and others on thought leadership about AI trust, responsibility and impact.
“Singapore Airlines is partnering with us and others actually around what's the thought leadership around how can we think of AI trust, AI responsibility, AI impact”
Niles says OpenTable uses Agentforce to help people around the world find restaurants and make reservations.
“Open table is using agent force. Can you imagine around the world to help people who want to have kind of find connection, community, good food, reservations and whatnot.”
Niles says Salesforce deploys Agentforce agents on its external help site and in many internal uses as 'customer zero'.
“our help.salesforce.com, actually external facing as well as a whole slew of internal agents in terms of the human teams”
Niles describes designing an AI agent around its role, task, guardrails, data and knowledge, and the channels where it works, which he compares to a human job.
“what's the roles? What's the capabilities? Like the task? What should they be able to do? What are the guardrails? Where are the channels, right, sort of in which they work?”
Niles says enterprise-grade agentic AI needs transparency, explainability and auditability, meaning being able to see what occurred and why.
“We're going to need transparency. We're going to need explainability. We're going to need auditability, right? Meaning what occurred and why.”
Niles says guardrails for AI agents come in preventative, runtime and self-improving kinds.
“Some are preventative. Some are runtime guardrails. Some are like self-improving guardrails.”
Niles says Salesforce's current agent systems hand work from agent to agent, and can chain several agents before returning to a human.
“It's actually agent to agent. Right? And you can have agent to agent to agent, right? And then sort of go back to the human.”
Niles says a person managing a fleet of AI agents needs a command center, a theme of the Agentforce 3.0 launch.
“You have your own agentic span of control. You need a command center.”
Niles says he oversees Salesforce's Office of Ethical and Humane Use of technology.
“the office of OEHU, Office of Ethical Humane Use, is an office I oversee”
Niles says Salesforce issued its first-ever Trusted AI and Agents Impact Report.
“we've issued our inaugural, it's the first ever Trusted AI and Agents Impact Report.”
Niles says Salesforce has millions of 'Trailblazers' worldwide through its Trailhead learning platform and wants millions of 'agent blazers'.
“for years, we've had this thing called trailblazers. Trailblazers, you know, and we have millions around the world.”
Niles says Salesforce has about 75,000 to 77,000 employees.
“We've got 75, 77 ,000 people at the company.”
Niles says the global AI summit series started at Bletchley Park in the UK with a focus on safety.
“They started in Bletchley Park, right? You know, in the UK around safety.”
Niles says the AI summit in Paris focused on AI action, and that a later one is planned in India.
“Then in Paris, right? We're talking about AI action. Then there'll be one in India that we're working on.”
Niles says Salesforce runs an annual intern program called Future Force.
“every year we do a future force, which is we bring the pleasure of having these interns come in”
Niles says he used to say that trust begins long before the prompt and continues long after the output.
“In the generative AI era, I used to have a phrase that trust begins long before the prompt and continues long after the output.”
Niles says Salesforce Ventures partners with and invests in companies as part of Salesforce's ecosystem approach.
“We have Salesforce Ventures where we just have the privilege of partnering with a lot of these companies. We're investing in these companies.”
Niles says Salesforce's vision is for Agentforce to help small and medium-sized businesses succeed rather than fail.
“what if instead of most small businesses failing what if more small businesses succeeded”
Niles recommends The Future Self, a book by a professor friend of his.
“a book I've been rereading, and it's called The Future Self. It's a friend of mine, a professor.”