Learning Library

← Back to Library

IBM and Salesforce Unite on Generative AI

Key Points

  • Malcolm Gladwell introduces the “Smart Talks with IBM” podcast season, which spotlights visionary “New Creators” using artificial intelligence as a transformative, game‑changing multiplier for business.
  • IBM’s long‑standing “better together” partnership with Salesforce has expanded into a new collaborative effort focused on generative AI, highlighting how both giants are combining forces to accelerate AI adoption.
  • Susan Emerson explains Salesforce’s evolution from the Einstein analytics platform to a dedicated generative‑AI team, stressing that the technology is unlocking broader opportunities for data‐driven decision‑making across enterprises.
  • Matt Candy describes IBM’s multi‑decade investment in a generative‑AI stack built for the enterprise, positioning the company to help clients and partners integrate these capabilities into existing technical ecosystems.
  • The discussion emphasizes practical applications of generative AI—particularly in customer service—and outlines how businesses can leverage the combined IBM‑Salesforce expertise to create more intelligent, data‑centric interactions with their clients.

Sections

Full Transcript

# IBM and Salesforce Unite on Generative AI **Source:** [https://www.youtube.com/watch?v=MaIFF5h2_UM](https://www.youtube.com/watch?v=MaIFF5h2_UM) **Duration:** 00:39:45 ## Summary - Malcolm Gladwell introduces the “Smart Talks with IBM” podcast season, which spotlights visionary “New Creators” using artificial intelligence as a transformative, game‑changing multiplier for business. - IBM’s long‑standing “better together” partnership with Salesforce has expanded into a new collaborative effort focused on generative AI, highlighting how both giants are combining forces to accelerate AI adoption. - Susan Emerson explains Salesforce’s evolution from the Einstein analytics platform to a dedicated generative‑AI team, stressing that the technology is unlocking broader opportunities for data‐driven decision‑making across enterprises. - Matt Candy describes IBM’s multi‑decade investment in a generative‑AI stack built for the enterprise, positioning the company to help clients and partners integrate these capabilities into existing technical ecosystems. - The discussion emphasizes practical applications of generative AI—particularly in customer service—and outlines how businesses can leverage the combined IBM‑Salesforce expertise to create more intelligent, data‑centric interactions with their clients. ## Sections - [00:00:00](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=0s) **AI Collaboration Between IBM & Salesforce** - The episode introduces IBM’s long‑standing partnership with Salesforce, highlighting their new joint generative‑AI initiative and how it can transform customer service and business operations. - [00:04:36](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=276s) **AI Simplifies Everyday Decisions** - Matt describes using an AI image‑recognition tool on holiday to decipher a parking sign, avoiding a fine and illustrating how AI reduces friction for consumers. - [00:08:23](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=503s) **From Rules to Generative AI** - The speakers contrast simple rule‑based automation (e.g., an “umbrella” rule for England) with generative AI that can navigate multiple systems to fulfill requests, and note that companies are currently focused on grasping the technology’s capabilities, risks, and governance frameworks before broader adoption. - [00:11:43](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=703s) **Scaling Generative AI in Enterprises** - The speaker explains how embedding generative AI into a sales‑service platform can boost workflow efficiency, emphasizing outcome‑driven use cases and the shift from early experimentation to organization‑wide scaling with proper guardrails and frameworks. - [00:15:01](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=901s) **Scaling Generative AI: ESG and Governance** - The speakers highlight that while data quality, access, and security are essential for expanding generative AI, organizations must also embed ESG concerns—particularly carbon emissions—and robust governance on bias, fairness, and security into their scaling strategies. - [00:18:08](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=1088s) **AI Integration Across Enterprise Platforms** - The speakers describe how IBM watsonx and Salesforce’s AI combine to surface back‑office data (supply chain, finance) within customer‑facing workflows, illustrating a partnership‑driven “one plus one equals three” opportunity and outlining the strategic steps for defining value, use cases, and implementation order. - [00:22:30](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=1350s) **Generative AI Transforming Sales Operations** - The speaker outlines how generative AI can automate personalized customer outreach, reduce administrative friction for sales teams, and shift from click‑based to conversational interfaces, emphasizing the importance of an IBM‑Salesforce ecosystem partnership for scaling these solutions. - [00:25:49](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=1549s) **Balancing Explainability and LLM Transparency** - Salesforce stresses explainable AI and robust governance while tackling the opaque nature of generative LLMs through prompt auditing, result tracking, and updated ethical safety frameworks. - [00:29:10](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=1750s) **Predicting Generative AI's Near‑Future Shift** - Jacob and Susan speculate that in the coming years generative AI will overhaul existing workflows—from banking to salesforce onboarding—replacing manual processes and manuals with intuitive, AI‑driven experiences. - [00:32:45](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=1965s) **Creativity, AI, and Human‑Centered Strategy** - Matt and Susan discuss how they blend generative AI with creative thinking in their work, emphasizing people‑first adoption and innovative product development at Salesforce. - [00:35:52](https://www.youtube.com/watch?v=MaIFF5h2_UM&t=2152s) **Diverse Collaboration Fuels Innovation** - They discuss how cross‑industry teamwork and intentional unplugged, unstructured time unleash fresh ideas, enhancing creativity and delivering better outcomes for clients. ## Full Transcript
0:00Malcolm Gladwell: Hello, hello. Welcome to Smart  Talks with IBM, a podcast from Pushkin Industries, 0:08iHeartRadio and IBM. I’m Malcolm Gladwell. This season, we’re continuing our conversations 0:14with New Creators— visionaries who are creatively  applying technology in business to drive change, 0:21but with a focus on the transformative power  of artificial intelligence and what it means 0:27to leverage AI as a game-changing  multiplier for your business. 0:32Today’s episode highlights the power  of collaboration. IBM has long been a 0:37supporter of the “better together” mindset  and embraces partnerships. They have been 0:42working together with Salesforce for more than  two decades, but have recently launched a new 0:48collaborative effort surrounding generative AI. Pushkin’s very own Jacob Goldstein sat down 0:55with Matt Candy and Susan Emerson. Matt is the  global managing partner of generative AI at IBM 1:02Consulting, helping clients and partners around  the world embrace this new era of technology. 1:09And Susan is a Senior Vice president for  Salesforce dedicated to AI, analytics and data. 1:16They discussed the historic collaboration  between the two tech giants, explored the 1:21opportunity AI presents for customer service,  and walked through how businesses can use 1:27generative AI to interface with clients. Okay, let’s get to the conversation. 1:34Jacob: Thank you, guys, for coming this  morning. So I'm interested in how you both 1:41came to generative AI—or maybe it sort of came  to you in the way it sort of came to all of us. 1:45But how did you arrive at  working on generative AI? 1:49Susan: As part of my remit at Salesforce over  the years, I've brought a lot of analytics and 1:54data and machine-learning products to life  under the Einstein brand at Salesforce. So 2:01as we pivoted Salesforce into taking  advantage of the generative-AI moment, 2:07it was natural that I became part of the advance  team leveraging generative AI. And it's—it's 2:14become interesting. What I see as I speak with  customers—the moment that everyone is facing 2:21in terms of how they incorporate generative AI  into their businesses, their workforces and their 2:28technical stacks—it's actually opening up a lot  of doors to the utility of analytics, data and AI. 2:36So it's been this big pull-through in terms  of incorporating not just generative AI, 2:42but a larger conversation around how we  become all better using data in our day jobs. 2:50Jacob: So that's a great frame for sort of what's  going on at Salesforce with generative AI. Matt, 2:56tell us a little bit about, you know, how that  fits with the way IBM is approaching the space. 3:02Matt: So I guess—probably three sides to that  question. And so there's the technology side to 3:08it. So IBM has a technology organization. And  so, you know, we are building, and have been 3:14over many years—decades, in fact—IBM has been  working in this space, a generative-AI stack, 3:21that allows organizations to adopt generative-AI  technology, uh, aimed at enterprise and 3:28business use within their organizations. So then within the consulting business, 3:33you know, we have 160,000 people who work every  day with clients across every industry—regulated 3:39industries, government organizations. And so this,  you know, is a really important technology that 3:46those companies are going to be using to drive the  next level of transformation in their enterprises’ 3:52processes, and the types of experiences they  build for their customers. And so, you know, 3:57we work extensively with partners—uh,  technology such as Salesforce, AWS, 4:01Microsoft, as well as our own technology. And then finally, I guess, the third angle 4:08is the work that we've got to do to reinvent the  business of consulting. And so if I think about, 4:13you know, consulting and systems integration, you know, ultimately 4:16we are knowledge workers, right? And so  from an industry perspective, I think, 4:21you know, our industry is—same as many others—is  gonna, is gonna go—undergo a level of disruption 4:28caused by this technology, but therefore  that will also create a huge opportunity 4:32for us as well. So those three aspects, Jacob. 4:36Jacob: Great. So, so that's the point of view sort of from your companies in your work. I'm curious to talk for a moment about AI from the point of 4:45view of consumers and employees kind of out in  the world today. So just to start with consumers: 4:52when I'm just out as a person, as a consumer  in the world, how am I experiencing AI today? 4:58Matt: I'll give you a great little use case,  actually. I was on holiday, uh, three weeks ago, 5:03in Tenerife in Spain. And, uh, I was trying to  find somewhere to park the car with the family 5:08for dinner that evening. And, uh, I found,  I found this area, next to this, uh, kind of 5:15shopping center. And there was this sign there. And, uh, I couldn't quite work out if it was 5:20saying I could park there or not. And so I took a  photo of the sign and I uploaded it to an AI tool, 5:28and I said, “What does this mean?” And it  basically explained to me what the sign 5:31was saying, and basically told me that  I shouldn't be parking there. And so I 5:34drove on and I found some—somewhere else to park. But yeah, that, that allowed me, in under sixty 5:42seconds to probably avoid a hundred- euro fine by  parking the car there. So—just a simple example, 5:49but I think the ability that these tools have to  take friction out of our daily lives, you know, 5:55and to be able to make, just, things that we do in  our everyday life simple and more frictionless—um, 6:02you know, that's how I look at, how “Matt  the consumer” is going to benefit from 6:06some of this type of technology. 6:08Susan: And from my perspective, it's also a travel story. I spend a lot of time on  the road for, for work, but recently had to send 6:18my sister and her family to a destination they'd  never been to for a wedding. And, it was really 6:25quick and easy to use some generative tools to  come up with a whole plan for them, because they 6:30love to hike and to be outdoors and to hike in  areas that aren't overly crowded with, um, people. 6:37And so gen AI very quickly gave me an itinerary  of a bunch of terrific hikes for them, uh, 6:42for a destination. So, things like that. 6:45Jacob: Great, and then what about the, the effect of AI and of automation more generally on, on employees, on the workforce? 6:54Susan: Well, there's so many dimensions to take  that from. Generative AI really can up-level a 7:00workforce in all sorts of ways by providing these  consistent ways to engage with technology with 7:06these natural-language experiences. So I think  it changes everything from—it finds us content, 7:12it generates us content, it makes it easier to  work with our systems of engagement and operation. 7:19And for many organizations, uh, it's,  uh, it can be a, a lifting factor in 7:24terms of bringing a more consistent workforce  experience, because these tools can just be 7:29ever present in our systems of, of work. 7:32Matt: I'll give you a little example. Here in IBM, we have something called Ask HR. And so that's our conversational AI interface 7:42that we use to interact with HR services. And  94 percent of every employee interaction now 7:49happens without human intervention, through that  interface. But you would never know that. And so 7:55if I think about, you know, our HR processes, you  know, we have this amazing conversational- based, 8:01uh, AI that we use for all of our HR  interactions, and we surface that through 8:06Slack. And so Slack becomes the front door for  how we access, a lot of these different enterprise 8:12processes and capabilities and how we surface AI. In fact, I'm taking a flight shortly back to the 8:17UK and our Ask HR boss is reminding me that it's  raining in the UK and I should take an umbrella. 8:23Susan: Isn’t it always, like, raining in England? Matt: Yeah, I don't think there's any AI needed 8:30for that. I think that's just a hard-coded “If England, then take umbrella.” That's right. 8:35Susan: That’s just a rule.  That's just a rule, right? 8:38Matt: Right, and you're able to converse, and, you  know, “I need to book a holiday.” “I need to move 8:41somebody between managers.” “I need to figure  out the policy on this.” And the AI basically 8:47navigates across the different systems to  be able to help get that information—to 8:52summarize it back, to be able to carry out the  transactions that I need to be carried out. 8:56And it just, just removes all of that complexity  and makes it easier to get things done. 9:01Jacob: Uh, when you are working with  companies to implement generative AI now, 9:11what do you find tends to be their primary focus? Susan: I mean, I speak with a lot of customers 9:16each week, and for the last several months,  most organizations have just been reorienting 9:23themselves in terms of “Where are we in this  moment? What is this technology capable of? 9:29What are the risks and governance and frameworks  that I need to establish in order to engage and 9:36talk to everyone—talk to my vendors, talk to  my cloud providers, talk to my consultants, 9:42talk to academics, and generally get your sea legs  under them?” And the—sort of the unstructured, 9:49hands-on-keyboards fiddling which technology seems  to be moving towards. “Let's get some points on 9:55the board. Let's turn this stuff on and go.” So  that's what I've been seeing in terms of, uh, you 10:01know, the work within the Salesforce ecosystem.  Matt, you've got a larger aperture as well. 10:06Uh, what, uh, what are you seeing? 10:08Matt: Yeah, so I, I—I definitely agree. I think, you know, there's been lots of getting sea legs, experimentation, just trying to build knowledge, 10:17being able to try and build almost a, uh,  you know, internal organizational point 10:23of view and reference framework. You know,  I've seen lots of what I would refer to as 10:27“random acts of AI,” you know, in, in, in  terms of, uh, in terms of experimentation, 10:33but I think, I think people—now looking into 2024  and—this is all about, now, adoption and scaling. 10:39What's become really clear is organizations  have started to realize: this is going to be 10:43a very multimodal world that they're  going to live in. There is no one AI 10:47that is the answer for their organization. And so there's—they're going to have lots of 10:52different generative-AI models and technologies  that are going to sit in the organization, 10:57servicing different use cases, different  domain areas, different products and 11:02services. And so therefore having to figure  out how they're going to navigate and manage 11:08this kind of open world that they're going  to be sitting in, and the decisions that 11:12they're going to have to make around that. I think the second thing that I've seen,—that 11:17people are now becoming very clear that  this needs to be what I would refer to as 11:21“use-case led and outcome focused.” and so really needing to start with thinking about 11:27the business outcome and the problem  that, you know, we're trying to solve, 11:32and therefore “How do I use generative AI as  part of the mechanism to solve that problem?” 11:38And I think, you know, what, what Susan and the  Salesforce team do is an amazing example of that. 11:43You know, they've got this incredible platform  and engine that allows, you know, companies to 11:48transform their sales and service processes and  to be able to put data in the hands of users, 11:52to be able to make better decisions, et cetera. And so now by weaving generative AI into that 11:58platform, we're going to be able to make those  processes and workflows even more efficient, 12:01right? So it's generative AI plus all of these  other amazing things that are there, but it will 12:06be led through business outcomes and it will  be led through the use case and the business 12:11problem or workflow that we're trying to improve. And then I think the third thing is shifting from 12:16this experimentation to scale. You know—I think  everybody's really early in this journey, but 12:21what's become clear is that, you know, everybody  now needs—realizes and is starting to lay down 12:28these, these ground rules, the guardrails, the  frameworks to allow them to scale this across 12:35the organization. So, you know, I think, I think  we're in for an exciting, exciting time in 2024. 12:41Jacob: So now that we're getting to this  moment, what are the key things companies have 12:46to figure out about scaling generative AI? 12:50Susan: I would put that in , kind of two categories. And following what—on what Matt was saying in terms of use-case defined and outcome 12:59led—100 percent on that in terms of starting  with a hypothesis of value while at the same 13:04time people are getting, uh, you know, closer to  the technology, to know what their bounds are. 13:10But the biggest, you know, set of conversations  is in the enterprise area in terms of embarking 13:17and using with generative AI—how to do it in  ways, that is safe for—uh, use of data that 13:24is safe around, not just, the larger topic  of generative AI and hallucinations, which, 13:31which are fun to talk about in the media, but— 13:34Jacob: It’s a fun word, right? If it was called something other than “hallucinations,” people wouldn't   talk about it as much. It was just mistakes. 13:41Susan: Yeah, that's right—just things that aren't factually true. We’ve been doing a lot  of work at Salesforce around using, you know, 13:48dynamic and structured—grounding the data so  we can give very strong and not naive prompt 13:54instructions to LLMs to get return on that. So, so just to summarize, uh: top of mind for 13:59organizations using, you know, large language  models is using their data in ways that are 14:05safe. Trusted. Not exposed. And reducing  the opportunity for hallucinations and 14:11maximizing relevant content. 14:13Jacob: Great. So, so Matt, Susan was talking about, you know, both what organizations are, are concerned with as they 14:19scale generative AI and how Salesforce is  working to sort of address those concerns. 14:24What are you seeing at IBM? 14:27Matt: Yeah, so I think—certainly from a, from a “scaling of generative AI” perspective,  this topic of governance, and how organizations 14:37are going to have to govern all of these models  that sit, sit with—inside, how they manage, 14:41kind of, bias, fairness, model drift. You know, if you think about the data 14:47that's gone into a model and the output it gives  to start with—not because the model changes, 14:53but because the context of the world  moves on. And so being able to kind of 14:56manage this model drift is going to be a really  important thing. I think data really matters. 15:01And so quality, access, security, uh, around data  within the enterprise is going to be critical to 15:07scaling generative AI. And the other one I think  that's going to be really important, and I think 15:11many organizations haven't even got there yet in  their thinking, is around the ESG implications. 15:16So: carbon. You know, the use of this technology  does not come without a cost of carbon. 15:22Jacob: Carbon meaning it's very energy intensive. 15:24Matt: Correct. Yeah. The training of the models. And so thinking about carbon disclosures and thinking about where I'm infusing it into my 15:33business and how much I'm using it and what the  carbon cost of that is. As I think about the—you 15:39know, my own organizational responsibilities to  reduce carbon, I think, you know, there's all of 15:45these things that I think are going to become  important factors as people are thinking about 15:49the scaling implications of this technology. 15:51Malcolm Gladwell: AI is already making new experiences possible, but we must be mindful in how we integrate this new technology as we 15:59continue scaling generative AI. Matt touched upon  some crucial aspects from an IBM perspective: 16:06governance, bias, fairness, and security are  all key considerations when organizations 16:11aim to expand their use of generative AI. The  environmental aspect is especially important, 16:18and it’s refreshing to hear leading thinkers like  Matt and Susan highlight these issues. As this 16:25technology continues to evolve, these factors are  becoming increasingly important for organizations 16:31to address. The historic collaboration between  IBM and Salesforce is helping to remedy issues 16:38companies face when scaling AI. 16:42Jacob: So IBM and Salesforce recently announced a new collaborative project 16:47around generative AI. Tell me more about that. 16:50Matt: We've been partners for over two decades now, IBM and Salesforce. And so within our consulting business, we work with Salesforce 17:00technology to help our clients implement that  technology to transform their businesses. 17:03We've got a huge practice—over 12,000 people with  certifications—around Salesforce platforms. And 17:10so, you know, with Susan and her team  and the broader team in Salesforce, 17:14we're infusing more capability into the platform  around generative AI. Then our mission is really 17:19simple. It's to help clients who are using the  Salesforce platform adopt those capabilities 17:25to help them get more benefit within their  organization. You know, we’re also a significant 17:29user of Salesforce technology within IBM. We're one of Salesforce's largest customers 17:34globally. And so, you know, as we continue to  transform our own sales and service processes 17:40within IBM, then our use of the generative-AI  capabilities that they're infusing into Sales 17:45Cloud, Service Cloud, Slack, et cetera will be  something that will become really important to 17:50us driving productivity within the company. And then the other thing that I would say is, 17:55you know—as I think about the work  that we do with clients, you know, 17:58as they're implementing and on their generative-AI  journeys, you know—they're going to utilize and 18:03leverage the Salesforce capabilities within the  platform, and their generative-AI technologies. 18:08But then you start thinking about processes  and workflows that run beyond the walls of CRM, 18:13right? That run into supply chain and into  the finance area of the organization. And 18:18so there is work that we're doing with clients  where we're using IBM's watsonx platform to be 18:24able to help get access to—to generate insights  from data sources that sit in all of these kind 18:30of back-office areas of the enterprise, and to be  able to get that data across the Salesforce into 18:36these customer-interaction points and into the  employees who are servicing those customers, using 18:42Salesforce's AI and generative-AI technology. So  there's a kind of “one plus one equals three” kind 18:48of, you know, “better together,” and being—being  able to bring our technologies together in service 18:53of these clients’, problems as you think about  these processes that run across their enterprise. 18:59So, so, yeah, so, so huge, huge  opportunity in what we're doing 19:02together in the market to help clients. 19:05Susan: Yeah, and just building on that, uh, it is a huge moment for, for organizations and for technology companies like Salesforce. 19:12And we couldn't be happier to have partnerships  like we have with IBM. Like, the range of thought 19:18leadership that is appropriate at the moment  is everything from “What is that hypothesis 19:25of value?” and “What are those use cases?”  and “What is the order of operation in terms 19:29of approaching it just in terms of focus?” But then things that would help organizations 19:35assess their AI readiness and then their  approach. Like, you know, we talked earlier 19:40about frameworks and guardrails. Uh, you know,  “What are use cases that we're comfortable with, 19:45given the state of the technology, that face  employees or face customers?” So creating 19:49these much larger roadmaps in terms of how to  approach this over a series of initiatives. 19:56The way it can fundamentally change the way we  engage with technology and what that means for 20:03the, you know, training and change management and  use cases that fundamentally shift how you engage 20:11with systems like Salesforce. There's just a massive opportunity for us together. 20:16Jacob: So you're talking in sort of general terms. I'm interested in, you know,  thinking in particular about the way generative 20:24AI can essentially lead to better business  outcomes, right? Like, what does that look like? 20:29How do you measure it? You know, there's a certain  bottom-line question there, right? Like, how does 20:34AI make businesses work better, and in what ways? 20:37Susan: You know, as consumers of products and services, we, we all love and respect great service, you know, in terms of getting timely, 20:44quick answers, resolving issues  quickly—all those, those types of things. 20:48And from the perspective of, of using generative  and predictive capabilities for agents who are 20:55interacting with customers, there is just a whole  ton of opportunity to take friction out of the 21:01process in terms of finding answers, resolving  issues, in terms of using these generative 21:05capabilities that will bring, you know, answers  and content to the fingertips more easily to the 21:11human agents that are working with customers. Now, taking that to the next step for 21:16organizations, uh, when they're ready to  move into more customer-facing automation, 21:22that's yet another channel, as a consumer, we'll  all enjoy with the brands and the products and 21:26the services that we want in terms of fast answers  and resolutions to customers. And as we all know, 21:32great customer experience yields return business. Now on the sales side, you know, maybe a different 21:39example—and these are areas where I think the  capability of predictive and generative go very 21:45well together in terms of focusing on business  outcomes. And a classic example would be, 21:51you know, predictions that help us understand  customer health. You know, “Is this customer 21:56engaged?” “Is this customer at risk?” Predictions  that help us understand the next best product or 22:02next best conversation—these all help focus a  sales team's time on a customer or a territory, 22:11and so that deep focus “puts all the wood  behind an arrow,” so to speak, in terms of 22:16where we should be engaging. And those types  of driven sales organizations that have these 22:23capabilities just lead to better performance  and outcomes and customer experience, too. 22:30Now, let's also layer in generative capabilities,  uh, where we're using the generative capabilities 22:36to assist and augment a sales team, where  we're using the power of generative for 22:40everything like generating, uh, personalized  and relevant customer- interaction content. 22:46For example, leveraging our customer data—like  engagement history, product purchases, 22:51service history—to create an email or a campaign.  And, uh, the scale of automation has just never 22:57been possible before. And, you know, maybe even  taking this one step further with generative, 23:02where we take all the administrative friction out  of the day job and doing things for sales teams, 23:07like summarizing their calls or  creating a meeting plan for them. 23:11And, you know, very broadly speaking, using  generative AI to change the interaction 23:15mode with systems like Salesforce  from clicks and, and training, uh, 23:20where people have to focus on the process,  to more-conversational user experiences, 23:25which are much more engaging and easier to use.  So all of this together is just incredible and 23:31transformational, uh, and makes, uh,  all businesses and people work better. 23:36Jacob: I just want to spend one more moment  on the partnership between IBM and Salesforce 23:42and generative AI. And there's this phrase that's  interesting to me. It's “ecosystem partnership.” 23:49That, I think, is relevant here. So what is an  “ecosystem partnership” and why is it, you know, 23:54helpful in, in creating scalable AI solutions? 23:58Matt: This idea of being open, I think is probably one of the most important premises for us as technology companies, for us as consultancies 24:10and system integrators. And for our clients to  think about the, the sources of value that can 24:15be created through taking an open approach is  hugely important. So if I think about—for us, 24:22“ecosystem” means making sure that we have  all of the different partnerships that we 24:28need with technology providers, with service  providers—that we can bring to our clients the 24:36right set of capabilities to solve the problem  that they've got, and not thinking that just, 24:41you know, what we have in house or what we have  with just one other partner that we work with, 24:46you know, is, is, is the right thing. And so, you know, I think every problem 24:50that our clients have is solved through a  range of technologies that come together in 24:55service of creating that business outcome. 24:58Jacob: I want to touch briefly on ethics and governance. Something like 80 percent of CEOs see explainability, ethics, bias, 25:08trust as major concerns on the road to  AI adoption. And so I'm, I'm curious how 25:15business leaders navigate these things, and in  particular how Salesforce and IBM are building 25:22these concerns into how they work with customers. 25:25Susan: You know, we've been incorporating predictive machine learning into our, our products since mid–last decade, and at that time we started 25:35with all of our ethics and governance, work at  that time in terms of frameworks for engaging with 25:41AI in ethical and safe ways, and have a lot of  guidance for customers in terms of those programs. 25:49The machine-learning, focus that we've had  at Salesforce has always been deeply focused 25:53on explainability. So if we're making,  you know, predictive recommendations 25:59to explain how we got to that, you know,  whether—that's something that a user sees as 26:04they're engaging with it, so they have full  trust, uh, in terms of interacting with it, but 26:09also for the practitioners who are building it. So we have this, like, long-standing vibe and 26:15capability with our predictive, side of the house,  and on the generative side of the house, you know, 26:21the state of the marketplace right now is—LLMs  for most people are, are largely black boxes, 26:28uh, in terms of not fully interpretable in  terms of how they come up with their content. 26:33Now, that said, there is a lot that you  can do in terms of audit, in terms of, 26:39you know, transparency, in terms of “What are  the prompts that are being submitted to these 26:44LLMs?” “What do these LLMs provide back in  terms of return?” And then “What did the 26:50human do to change it, use it, or adjust it?” So we've been updating all of our ethics and, 26:56and governance frameworks now, I guess I would  call it, with safety components as well in terms 27:01of how to work with data, in safe ways and  with these transparent governance models. 27:06Matt: Yeah. So, I mean, this is an area that IBM  has been kind of working on for many years. And 27:12so, you know, our AI ethics board that we have  internally kind of governs and, and provides 27:17frameworks and guidance for everything that we do  in the company. There's a lot of work that we do 27:22to help our clients and organizations establish  their strategies for AI governance, as well as 27:28their own internal policies, models, approaches,  ethics boards, et cetera. And so, you know, 27:35helping them put in place these ground rules and  guardrails, organizational process, uh, changes, 27:42et cetera, I think is a really important part  of this scaling discussion that we were having 27:47earlier as, as, as people are going to be kind of  rolling out more of this technology internally. 27:52And then I think there's, a lot that organizations  are going to have to do to think about— especially 27:57in the generative world—around all of the  different types of models that they're using, 28:02models that they're training and tuning  and building, and how they manage all of 28:06those for explainability and bias drift  and, actually, regulatory requirements. 28:12Like if you—if you think about what's, what's  happening around the world as different 28:16countries—uh, the EU AI Act—you know, there's  lots of different regulatory requirements that are 28:22going to be coming in. And so for multinational  companies operating across multiple countries, 28:29how they're going to be—have to make sure that  they're, they're complying with all of not only 28:33their own internal policies, but the requirements of the country, 28:39as well as, potentially, industry  regulatory requirements as well. 28:44And so there's a lot that we are doing and going  to be doing in, helping them manage complexity. 28:50But IBM has a very firm view that we believe  that this is all about regulating AI risk, 28:55not AI algorithms, and so focusing on,  precision regulation. So, you know, 29:01use the, use the bodies and—regulatory bodies  that are out there to provide the, the control, 29:08as opposed to trying to regulate the technology. 29:10Jacob: So generative AI is changing kind of absurdly quickly, right? A year and a half ago, we wouldn't have been having this conversation. 29:17We're here today. Everything's happening now.  I'm curious what you both think about— about the 29:22near-term future of generative AI, right? If we  came back in a year, or let's say two years from 29:27now, if we came back two years from now to talk  about the work you're doing in generative AI, 29:32what would we be talking about? 29:34Susan: I use this example sometimes. I have three kids, and I don't think any of them  have ever gone into a bank to deposit a check, 29:45right? They pull out their mobile phone and they scan the check with the camera, and they're done. 29:50Jacob: I'm surprised that they even know what a check is, for the record, but yes. 29:53Susan: Right. Well, yeah, uh, sometimes their parents give them one. Like, they get direct deposit. But anyway, like, this experience of, 30:02like, “What do you mean I go into a branch and  cash a check? I just do this with my mobile 30:07phone.” And I, I think a little bit of, of it that  way in terms of the systems that we use at work. I 30:13can imagine explaining to my, my kids like, “Oh,  yeah, at Salesforce, you know, back when someone 30:18had their first day on the job, you know, as  a, as a service agent or as a salesperson, 30:24they would have tabs on the screen and they  would be trained where to click and they'd have 30:28documented processes in manuals, and it showed  them where to get from point A to point B.” 30:35And as the clock turns forward, they're just  interacting with a natural-language prompt. But, 30:41it just kind of fundamentally  changes the way we'll be able 30:45to interact with our systems of record at work. 30:48Jacob: It’ll be just much more conversational. Instead of clicking through something,  you'll just basically have a conversation. 30:53Susan: Much more conversational. 30:55Matt: Yeah, this is the biggest paradigm shift in how we interact with technology, I think, since the graphical user interface. 31:02And it's going to enable us to almost put aside  all of that complexity within organizations 31:08around system silos, process silos, flows,  because you're just going to layer this, just, 31:14simple, natural-language interface over all of  that complexity. Yeah. It's going to amplify, 31:21I think, the potential of every person on every  team in a way that we've never been able to see 31:26before. And the other thing that I think, as you  project forward in a couple of years—and Susan, 31:31just picking up on the point that you  talked about, about banking, you know, 31:34I think there's a wonderful little example. If you think back to the ’70s and the ’80s, 31:38when the ATM, kind of, cash machines  were rolling out—and at that time it 31:44wasn't really a reaction that was one  of awe or appreciation for convenience, 31:48but people were concerned that we were  automating away the bank-teller jobs, 31:53right? But now when you think about it, what  actually happened was, this technology allowed 31:58the banks to scale their branch networks, more  branches than ever before, more bank tellers than 32:04ever before. Bank-teller employment and salaries  increased, even though we automated a lot of work, 32:10because when they weren't having to spend  their time counting cash out for people, 32:14they were able to do more-valuable things, right? And new types of financial products and services 32:18and mortgages. And so if I think back to that,  in the ’70s and ’80s, and then I project to 32:24where we are today, we're just going to  unleash this creativity and potential for 32:29employees and enterprises by freeing up the  time that they're spending on things that, 32:33you know—they can do far more value-added tasks. And so I think—we're going to be amazed, I think, 32:38around what, what happens and what companies and  people are going to be able to do as we give them 32:43the time and space to be able to do that. 32:45Jacob: Great. So just to close, I want to talk about how both of you use creativity in your own work. 32:52Just to start with you, Matt: I know that you love to combine creativity  and technology through design. Do you use 33:01generative AI in your own creative process? 33:03Matt: Yeah, so I, I, I'm a firm believer that this combination of experience and AI is going to be the thing that makes a difference. Like, 33:12these large language models and this technology  has been around actually for a number of years. 33:17And it's only at the point, late 2022,  where open AI wrapped a digital experience 33:23around this and put it in the hands of  people, that suddenly the transformative 33:27power of this technology was realized. And so I think the way that we surface 33:31these capabilities and put them in the hands  of people, to be able to adopt it in a really 33:37frictionless way, is, is the thing that's going to  be, hugely important to the adoption and scaling 33:42of this. So I, I think the most important  thing for companies to do is to make people, 33:47not technology, central to their strategy. 33:50Jacob: Just to go more broadly into your work, Susan—I mean, I know that you have, have launched Salesforce's AI products into 33:58the market and that, you know, a lot  of those have been built—obviously, 34:01given Salesforce’s business—around helping  people build stronger customer relationships. 34:07Right? And so I'm curious: what  creativity did you bring to that work? 34:11Susan: Some of the products that I work with  at Salesforce—they're, they're deeply visually 34:15focused. And my personal perspective is,  is that the world can be really noisy. Uh, 34:21we're just inundated with all sorts of demands  on our time through so many channels, right? 34:28Like the phone is firing off, you're getting  instant messages, you're getting Slack messages, 34:32you're getting, you know, DMS, you're  getting emails, your phone is ringing, 34:37there's processes that are bearing down on you. And if we can use really good design to filter 34:44out and essentially weed the garden— because, you  know, we have this, this phrase at Salesforce, 34:49is “Everything—if everything's important,  nothing's important.” So using really good design 34:54to create the user experience in Salesforce just  brings stuff to life in the most powerful way. 35:02So I always think of it from that perspective.  Like if I'm going to put this on a screen, and, 35:06and, and Salesforce—what did I not put on?  Is this the most important thing? And is this 35:12the thing that's going to align everyone to the  larger initiative of the firm? So, so it's that 35:17kind of design thinking that I use probably every  moment of the day, whether I'm building a demo or 35:24talking to an executive at a company in terms  of—as I see a vision for how they might deploy 35:28our products, to actual product development. 35:32Jacob: And then just to kind of bring together these two themes we've been talking about: on the one hand, this sort of—ecosystem partnerships, 35:39and on the other hand, creativity. I mean, can you talk a little bit 35:42about how working with, working with partners  can foster a different kind of creativity? 35:49Susan: More perspectives are always  better than few perspectives. 35:52Matt: I completely agree. I think the more minds, the 35:55more perspectives, the more experiences, I think  about some of the best sessions, best workshops, 36:03best work we do with clients—it's when you've  got people not just from one industry but from 36:10many industries, because actually the adjacencies  and the things that are happening in these other 36:15spaces trigger new thoughts and new ideas. And  so, you know, I think the richness that we get 36:22when we partner with Salesforce together around  helping clients transform their front office, 36:26their sales, service, marketing processes,  we all bring these unique experiences. 36:31And I think that just opens  the aperture to better, better 36:34outcomes and better perspectives for our clients. 36:37Susan: Well, you know, you've been asking these questions about, like the use of, of tech and AI and creativity all sort of in the same sentence. 36:45And one of the things that I also think of  is—in terms of remaining deeply creative—is 36:51the actual process of unplugging for all—from  all that stuff. So taking a trial run with no 36:58earphones in your head for me is always a really  good way of unleashing and unbridling a lot of, 37:06you know, creative spirit. Just that, that downtime and 37:10the unstructured time where your brain can  just run free, actually, not assisted by any 37:14kind of device in my head or in my face. So— 37:18Jacob: I think, with that praise of unplugged time, we should say goodbye and let's unplug. It was lovely to talk with you guys. It was 37:25really interesting to learn about your work  and the relationship between the companies. 37:28So thank you for your time. 37:32Matt: Thank you, Jacob. 37:33Thank you. Malcolm Gladwell: A huge thanks is due to  Jacob, Matt and Susan for illuminating the possibilities of generative AI. 37:39This technology has great promise for creating new experiences in the future but requires the scaling capabilities made possible 37:48by partnerships like IBM and Salesforce. As our conversation with Susan and Matt 37:54illustrated, we’re at an exciting phase  of adoption. Most companies have moved 38:00beyond experimentation and are now prioritizing  scaling. The key areas of focus for organizations 38:07now include managing multiple AI models, as  well as thinking about specific use cases 38:13and desired outcomes. However, this scale is  difficult for companies to do on their own. 38:19To unlock the real potential of generative AI  in transforming experiences, they’ll require 38:25the scaling capabilities made possible  by partnerships like IBM and Salesforce. 38:31This conversation showed the promise of  teamwork. When massive companies combine 38:36their brainpower to push forward technology,  their collaborative efforts have the potential 38:42to revolutionize industries. One quick programming note: 38:47we will be taking a little time off, and will be  returning in just a few weeks with a new episode. 38:54Smart Talks with IBM is produced by Matt  Romano, Joey Fischground, David Zha, and 38:59Jacob Goldstein. We’re edited by Lidia Jean Kott. Our engineers are Jason Gambrell, Sarah Bruguiere, 39:06and Ben Tolliday. Theme song by Gramoscope. Special thanks to Andy Kelly, Kathy Callaghan, 39:14and the EightBar and IBM teams, as well as the  Pushkin marketing team. Smart Talks with IBM is a 39:20production of Pushkin Industries and Ruby Studio  at iHeartMedia. To find more Pushkin podcasts, 39:27listen on the iHeartRadio app,  Apple Podcasts, or wherever you 39:31listen to podcasts. I’m Malcolm Gladwell. This is a paid advertisement from IBM.