娇色导航

Keith Shaw: Hi everybody, welcome to DEMO, the show where companies come in and show us their latest products and platforms. Today, I’m here with Shibani Ahuja. She is the Senior Vice President for Enterprise IT Strategy at Salesforce. Welcome to the show, Shibani. Shibani Ahuja: Thank you!

Thrilled to be here. Keith: All right, we don’t have to explain what Salesforce is—CRM, cloud, all that. So, what are you showing us today?

Shibani: I’m going to show you two things. First is Agentforce. We’ve evolved from chatbots to copilots, and now we’re entering the era of agentic, autonomous digital labor.

It's a hard concept to grasp, so we’ve prepared a couple of demos to help explain what Agentforce is and how it drives digital labor. Second, we’re introducing our Agentic Maturity Model.

It’s an in-house framework we use with customers to help them understand where they are today and where they want to be with autonomous agents.

Everyone's head is spinning — people are asking, “Where do I even start?” And when they try to jump to Level 4 outcomes and ROI, it gets really complex.

Keith: People hear about agents and immediately jump to the end goal. Or they start at the beginning and ask, “How do I get from here to there?” Shibani: Exactly.

We want our customers to start at Level 1. Even starting there helps build the foundation for an agentic architecture. Sure, keep Level 4 in mind—with multi-agent orchestration—but start where you can demonstrate real value on a specific use case. Prove the technology works.

Keith: Now, with enterprises using Salesforce, does Agentforce and the Agentic Maturity Model apply to everyone, or just specific groups?

Shibani: You might assume this is just for marketing or sales, but the Salesforce platform goes far beyond that. We’re building use cases across IT, finance, HR—even compliance. This is an enterprise-wide platform now.

Keith: What would companies do without this? Would they just be building all this on their own?

Shibani: Yes, and that would be complex. A lot of companies say, “We’ve got our own LLM,” but an LLM alone isn’t enough — it’s just one piece. The challenge is the ongoing management: fine-tuning, refining, monitoring risks.

A platform approach gives you a unified, extensible foundation with open APIs that can connect across systems.

Keith: All right, we’re going to jump into the demo now. You’ve got a little surprise for us? Shibani: Yes!

My partner in crime — Mike from Slalom — is here.

Keith: Let’s bring on Mike Jortberg, Senior Global Sales Director at Slalom. Welcome, Mike! Mike Jortberg: Thanks, Keith. Keith: Let’s jump into the demos. You've got some cool use cases.

Mike: We’ve got four. As Shibani mentioned, the maturity model spans from basic to advanced. We’ll follow that format. First: What does a sales rep do to prepare for a meeting?

Then we’ll look at email correspondence, followed by property claims in insurance, and finally customer service in travel and hospitality. Salespeople don’t always have time to prepare thoroughly. What we’ve built is a simple account summary — a quick snapshot of everything going on with a customer.

Instead of digging through 10 screens, it’s all right there. It turns a 30-minute task into a one-minute prep — and they’re doing something they’ve never done before.

Mike: When a CEO gets an urgent email about a customer, this tool helps staff immediately understand the issue. That’s a huge time saver. Second use case: email. We’re drafting an email to you, Keith, using prebuilt instructions and CRM data.

AI is merging CRM and context to create a personalized draft, even in multiple languages like French. Customers have 500+ templates and send 10,000+ emails a month. We’ve seen 85% time compression.

Keith: The personalization alone—knowing my products, interests, and then writing in French—that would take forever manually. Mike: Exactly.

With Outlook or other copilots, it’s still alt-tab, copy-paste all day. Here, AI + CRM + data are unified. That’s a big deal. Next, we’re switching to the insurance industry — property claims analysis. External systems contain files like police reports and policies. These get summarized automatically.

Comparing them manually takes time — four-page documents, two monitors, lots of scrolling. We show the power of AI by asking it for a summary and to find discrepancies between the report and claim. That’s often where fraud is discovered.

Keith: Does that happen often? Mike: Yes.

It’s a common concern in insurance.

Mike: Now we see the discrepancy: the customer claimed all windows were broken, but the police report says only one. We write that discrepancy directly into Salesforce automatically.

Mike: Next up: Level 3 of the maturity model. We’re moving into customer service for travel and hospitality.

Mike: This time, we’ve got public site interactions—passenger issues with their itinerary. On the left, call center software; on the right, customer self-service. The passenger submits their booking info, and Agentforce kicks off a resolution process, including loyalty points, case creation, and system integration.

Keith: And this includes sentiment analysis too? Mike: Yes.

Let’s run another example. The customer complains about a bad smell in the room. Sentiment analysis detects frustration from a high-value guest. We authenticate, kick off a service case, and offer a perk like a private cabana with a QR code.

Keith: So the agents are interacting with internal and external systems—even third-party companies? Mike: Exactly.

Refunds through Stripe, loyalty systems, booking tools—this is the blend of AI, data, and CRM. And it delivers “white glove” service without the high cost or manual time.

Keith: What’s great is that these are real, practical examples. They show what’s possible.

Mike: When working with customers, we focus on high-volume tasks. Saving two minutes per task 10,000 times a month adds up. We've seen 85% compression in things like post-flight customer service emails — 30 minutes down to two. Keith: Thanks, Mike.

We’re bringing Shibani back for a couple of final questions.

Keith: If someone’s watching this and thinking, “I want to try this out,” where can they go? Shibani: We want as many hands on Agentforce as possible — testing, building, deploying. We’ve got Agentforce World Tours happening globally.

Check the Salesforce website, show up at an event — there will be demos, booths, and more.

Keith: And they can meet you there? Shibani: Absolutely! Keith: Shibani, thanks again for being on the show. Shibani: My pleasure. Thank you! Keith: That’s all the time we have for today’s episode. Be sure to like, subscribe, and drop your thoughts in the comments.

Join us every week for new episodes of DEMO. I’m Keith Shaw — thanks for watching!