Chapters
Shownotes Transcript
Any time when technology has an impact, even if it's a positive impact, it translates into change. Some people are scared of AI because they've read those scary headlines and now they're like, "Oh my God, are robots gonna kill humanity?" Well, not necessarily, at least not yet.
I also think there's so much value for organizations who are thinking about implementing AI. It's a journey. It's a transformational journey. It's not all at once. So delivering capability in consumable chunks so people can be comfortable with it, get used to it, understand how to interact with it and understand how it fits into the overall work ecosystem is the best way to approach it.
Hello, everyone, and welcome to Experts of Experience. I'm your host, Lauren Wood. Today, I'm joined by Irina Goodman, the global leader of professional services for AI at Salesforce, to talk about all things AI, automation, data, and implementation. Irina, how are you? I'm doing well. How are you, Lauren? I'm very well. So I want to get into...
everything that your team does. And I think the best way to start with that is to really understand how is your team supporting your customers when it comes to really transitioning to the AI products and services that Salesforce offers?
Sure, I'll be happy to tell you. So when Salesforce made a serious commitment into AI as a platform, it made an equal serious commitment into customer success. Hence, my organization was stood up to support the sellers and implementers within Salesforce, as well as our customers. My team is truly global. We are in 12 countries, actually.
And we're helping our customers through the ideation, through the kind of brainstorming effort, as well as the actual implementation when it comes of enabling AI capability within Salesforce.
So tell me a little bit about the steps because you're starting with the ideation phase. And so what I'm hearing in that is this is not an off-the-shelf solution, as I'm sure anyone who has tried to do implementation with AI knows. It is extremely malleable. But tell me a little bit about how you take the clients through everything from that ideation to implementation. I know that's a lot, but even just high level, if you can kind of talk us through what those steps look like. Sure.
Sure. Now, this is a great question. And I'm glad you pointed out like off the shelf capability, which some people can say, yeah, sure. We can just turn on email generation or case summary. And yeah, it will work. However, the reason we take customers through ideation is to ensure that solution actually solves customer problem and brings value. And to start with that, regardless AI or any other implementations,
it is helpful to understand customer vision, goals, and objective. What do they actually want to achieve from a business perspective?
Once we understand that, it is also important to know, well, what are the obstacles? What are the gaps? What is staying in the way of achieving those objectives? Once those two components are in place, we can demonstrate how Salesforce AI, as well as other capabilities, are able to close those gaps.
And then we can overlay that capability over the use cases and the needs that customers have. So the best way for us to do it is what we call is an AI day workshop.
When we actually educate customer in the context, right, because you hear, you read, but it's really helpful to know what are the capabilities that are available in specific Salesforce cloud today and near term. So once that context is set, we can say, okay, knowing what you just learned, let's talk about your needs, your pain point, your obstacles, and kind of
overlay it with the capabilities that you've just seen to see how it can bring maximum value. I really like how you tied this in in the beginning to Salesforce's commitment to customer success, because the way that you're approaching this is very much from the start a consultative relationship. And like you said, you want to make sure that the product and the solutions that they're purchasing actually solve their problems, which is
For anyone who's in customer success, you know the pains of when a client is actually not having their problem solved with your product. We need to make sure that there's that alignment from the start. And I also think there's so much value for organizations who are thinking about implementing AI to do
a workshop to understand what are we really trying to solve here? Because there's so many possibilities when it comes to what AI can solve, but really understanding what is the problem that we have and maybe that our customers have that we're trying to solve so that we can really
target the right thing and not just have fancy tech loaded on top of fancy tech that actually just makes things more complicated, which I know I've experienced in the past as a leader where we're like, oh no, now we just have so many tools. So I really appreciate that approach. I'm curious to know how you work with your clients to manage change, because there's one thing in terms of addressing the problem, finding the solution, doing the implementation, but
This is a pretty big technological shift that a lot of companies are approaching as they implement AI. And I'm curious to know how much of your work ends up coming into really that change management for their teams. Yeah, absolutely. So if any typical, and I'm going to put typical in quotes,
technology implementation project, people sometimes kind of push back on change management or strategy. Oh, it's optional. We're just getting rid of our old technology stack. It really becomes table stakes when it comes to generative AI.
And this is due to the nature of the technology we're referring to, because generative AI is very much interactive, right? It's almost conversational with the users. So it interacts with its user base very different from traditional change. And sometimes we can say it almost augments or changes the behavior and productivity of people.
Anytime when technology has an impact, even if it's a positive impact, it translates into change. Some people are scared of AI because they've read those scary headlines and now they're like, oh my God, are robots going to kill humanity? Well, not necessarily, at least not yet.
Other people saying, okay, I've tried this tech. It's not as scary, but is it going to replace me? It is definitely going to replace me, which is also, you know, potential non-founded fear. So in order to address all those items, it is important to recognize that change management is an intricate part of implementing AI technology. Mm-hmm.
The first step that we start with is to demystify those questions, those unfounded or somewhat founded, so to speak, mysteries and assumptions.
For example, what is the impact to the workforce? Well, the research shows that your top performers are going to have marginal benefit from AI. I don't know, 10%. However, your average performance, and I don't like to use the word average, but your average performance would have a huge productivity improvement, 50% to 60%.
So now what's happening, the cheaper workforce all of a sudden is becoming very productive. It is an opportunity to potentially redirect, repurpose your top performance. Maybe they'll be your control group. Maybe they will transition into the new role such as prom builder, etc. So there's inadvertently a change that comes.
could be introduced to the work environment, merely to the positive productivity impact that AI has onto the workforce. So education and explaining to people the change that AI is going to bring is step one. Also demonstrating to people what's in it for them. That yes, even though it will gear productivity improvement or other changes, it also opens up other opportunities for people to pursue.
And the last piece that I would say is delivering capability into consumable chunks. It's a journey. It's a transformational journey. It's not all at once. So delivering capability in consumable chunks so people can be comfortable with it, get used to it, understand how to interact with it and understand how it fits into the overall work ecosystem is the best way to approach it. Thank you.
Thank you so much for sharing that. I think these are such important points for anyone thinking about, and I'm pretty sure everyone is, thinking about implementing AI solutions in your business that, yes, it's going to help you
do things faster, more efficiently in a more cost-effective way. But there is this critical moment of transition that needs time and attention. Because like you said, this is not just swapping tools, one tool for a better tool. This is a new way of working. This is a whole new way of looking at how we do things in a different way. And everybody is going to go through a transition. At the end of the day, I was listening to a podcast the other day about this.
At the end of the day, when we go through change, we grieve. Like there's an actual emotional response. Even if it's good change, we're still grieving the comfort of the old way. And...
I think it's such an important factor when we talk about AI that there's actually a human experience that happens as we adopt new technology. And that human experience is resistance to change, even if it's good change. So I really appreciate you sharing all of those tips and, and,
methods for getting people on board. And there's a lot of things I want to double click on there, but we'll come back to it. What I want to understand is a little bit about what are the AI products that you're supporting your customers to transition to? Good question, because
We used to say that computer is doing something, it must be AI, right? If something happens automatically or automagically, it must be AI. But actually, there are a lot of flavors of AI. Obviously, generative AI and conversational AI, co-pilot is the talk of the town. And yes, Salesforce does have this capability, and I'll talk a little bit about them. However, it's not the only capability that we're limited to.
For instance, Salesforce had been proficient for years, close to 10 years, actually, in predictive AI, which is very powerful and is able to bring to light helpful analytics and insights without which people will be driving blind, so to speak.
We also support automation, robotic process automation, flow automation, which is actually a point zero of AI, the most rudimentary form of AI. So when we talk about implementing AI, we'll look at the broad spectrum of what is actually needed and using the right technology for the right problem. In fact, one of the common barriers of going from pilot to production is
is using the right technology to solve a problem because people just trying to get that Gen AI checkmark. So the other kind of important component, the way Salesforce went to market with AI is that we embedded this capability into flow of work.
So customers don't need to get used to a new platform. We don't need to implement new tools for them. It's actually the same tools, the same Salesforce clouds that they've been using day in and day out that have been augmented. Why?
powered, enhanced with generative AI capability, giving that automation, that kind of automatic assistant within their flow of work. And by combining generative AI with predictive AI, you're actually able to get more powerful prediction based on the custom prompt that you put together. Mm-hmm.
Super quick for everyone. Could you define the difference between predictive AI and generative AI, just so that everyone really understands? Predictive AI would be analyzing the data and making prediction or recommendation based on the data. Mm-hmm.
It has a lot of models, a lot of data that it uses to make those predictions. So for example, if we're talking about sales process, it would analyze all the attributes that user will define that are important and make recommendation on leads prioritization and scoring.
So that would be an example of predictive AI. Generative AI is the most common example is chat GPT, right? That generates responses based on the questions that you ask using natural language model. Great. And so another question just kind of in the details about the Salesforce products. So
For existing customers, they can opt to add on an AI layer, or are you rolling this out for existing customers just for everyone? How is it kind of working in terms of the rollout of your AI capabilities? So from a kind of technicality point of view, there is an Einstein skew that customers need to purchase in order to have that capability available within their cloud.
And my team can help with configuring it based on customer specification. Got it. Great. What are some of those configurations? If you can give us some examples of how one tool might be used in different ways. One of the things that AI is great at is summarization. So the generative AI...
summaries can be applied to summarize. Call summaries or case summaries, expediting performance. Imagine a customer service representative working on a customer case and then finishing that work needed to write an extensive summary notes. Well, now with a click of a button, all they have to do is review, which takes significantly faster. So from
15, 20 minutes of writing the summary, we got it down to five minutes of review. Another example is what we refer to as generative responses based on knowledge. There's a capability to store knowledge articles within Salesforce Cloud.
If that knowledge is properly organized, so my team will work with the customer to make sure that the knowledge is properly organized, knowledge articles properly indexed, that is accessible and available for AI. Based on the questions agents get,
they could get generative responses based on those knowledge articles. So again, instead of doing extensive search, reading the articles, figuring out which information is relevant with the click of a button, you get that response presented to you. We always have human in the loop or human at the helm. The human makes ultimate decision. However, the tools and technology make it
easy for the human, for the user to have that information available to them and presented to them in consumable manner.
I'm so excited because these are things that would take us so long. Even two years ago, I was running a customer service team and creating our knowledge base because we were operating without one essentially. And putting that all together is a lot of work, but then to use it, it's a lot of work. And what AI is solving is the use of that information can be at
the touch of a button or at your fingertips so quickly. I'd love to double click on what you were saying about to properly index that information, just to kind of paint the picture for people of like, what does this really entail? What's ahead of them if they're thinking about implementing something like this? So I think it's a perfect segue to talk about data.
You probably have heard the saying, your AI is only as good as your data. When I talk at the conference and I do a poll and I say, hey, raise your hand if you feel that your data is perfect and you could be a poster child for data cleanliness. Not only I don't get any hands raised,
I, exactly, the reaction is the same as you had. People just start laughing. So I said, well, wait, if AI is only as good as your data, should we just all pack and go home? And the answer is no. There's always enough clean data
to get started, and which is what we want our customers to do, to start to try it out, kind of to dip your toes in the water to see if you're like temperature. If we go back to change management, to start getting that consumable bite of experiences with AI. However, it does not take the
data cleansiness and organization of the data away. So you do need to, as a company, you do need to have data strategy. Going back to knowledge, as you know, in a lot of customers, not every article is neatly stacked and organized within Salesforce.
Customers have a lot of knowledge outside of Salesforce. Sometimes that knowledge is in PDF attachments or videos. Or if we're talking to a technical company, sometimes it's like hundreds of pages in manual. So that's where it's AI plus data, right? So that's where data cloud comes into play, which allows to...
deal with that unstructured data allows to bring that outside data into a safe environment protected by one of the capabilities that Salesforce has referred to the stress layer, meaning it protects customer data when it used by AI.
So with AI plus data cloud really opens huge opportunities for our customers, open the door of having that data outside of Salesforce, but having that available for AI to use to make those recommendation and predictions. And then when they upload it,
How do you sort it? How do you make sure you just like putting files in there? How are you tagging things? Is it a lot of work? Like, what is that? What does that kind of look like? So without getting too technical, but it is based on the business rules and specific criteria that are
our team will work with the customers on to define it and how to. So it would be some of the solution architects gathering those requirements and then configuring the system per those specifications to make it easy for us, to make it easy for us. Totally. There's a little bit of work so that we can really get to it. And what about training the models?
How do you go about doing that, especially for specific use cases? And I especially think about the use case of the response, of pre-writing a response for a customer service associate based on what the customer was asking and using all of that knowledge base to create the right answer.
But there's still tone of voice. There's still specific elements of how we want the AI to speak on our behalf. And how do you go about really training it to follow those cultural norms that the team has? That's where our prompt studio comes into play, where customers are able to set up custom prompts and fine tune those prompts.
to get the right response, the right tone that they're looking for. We do have a prompt library that we come with. However, that's again where the partnership comes in, where we would work with a specific customer to help build those prompts to make sure that the tone, the right template,
the right response is taken into account based on how you engineer and write the plan. Great. And that's a whole new skill set. It's a whole new job. Totally. And I wanted to come back to what you had mentioned about the fear of losing jobs. I think that's a huge topic. And in a way, the answer is yes, there are some jobs that will be less needed or won't have as much
to them, or maybe will disappear altogether. But at the same time, new jobs will emerge. And we've seen this throughout history, throughout technological innovations or revolutions such as AI. And so I'd love to hear your opinion on that topic. What do you say to people when they say, what about the job loss? Right. So, well, who thought that helicopter pilots will be great drone operators? And drone operators
as a profession didn't even exist certain number of years ago. Neither did prompt engineering. What AI is trying to do, so first of all, it's not possible, at least not yet, for AI to completely replace workforce. It's not what's there for, right? It's there to automate manual repeatable tasks.
It is to help humans with presenting the data and information faster and better to them. Even if it's code generation, it's still based on the question, based on the prompt, providing information back. So yes, certain jobs will be optimized and maybe will not be as needed. But those are probably the jobs that may be not as exciting,
And it will open up, as you said, opportunities for people to transition and explore new jobs, new territories. When we were talking about productivity before, maybe those top performers in the call centers can now transition into more of a supervisory role.
Or maybe they will be more of a prompt engineers because they know those responses in and out, and they can help to write prompts, write questions in exactly the best way to get the most accurate results. Or maybe there will be a control group that will check and validate that AI is actually bringing the right responses.
So every change comes with opportunities. And even though certain professions might go away, I think it will open significantly more doors for opportunities and new roads to take.
We know that data-driven companies are outperforming their competition. That is a fact. But connecting all of your data is not easy. And thankfully, Salesforce is here to help. With their data cloud, you can unlock value from your trapped data. So head on over to salesforce.com slash products slash data to learn more. And I think it's really important for anyone who is having this concern to think about what has happened in the past.
in the past. If we look back on history, someone mentioned to me the other day, it's like when the steam engine was first invented, right? Or crash print. Yeah, exactly. So we had manual labor creating commodities and then we created machines so that we could do it at scale. And that created jobs for people to actually manage those machines and create those machines. And time has progressed and look at where we are now, right? The internet, computers.
We are not going to be without jobs. It's just that jobs will change. We're going to go through that phase of change and it's happening quickly. It's happening faster than it has in the past. So we need to buckle up. And I think just as you were saying everything about, you know, how this is really impacting, especially the customer service industry, I'm thinking about how working in a call center just got a lot more interesting. Yeah.
You know, and I having managed customer service teams in the past, it can be difficult to keep those people engaged because they're just sending the same responding to hundreds of tickets every day and it becomes really monotonous. But what if we're actually putting our minds together to think about how can we do this in a more scalable way and in a way that creates an even better customer experience? Because now we just freed up time to focus on what matters most, which is actually are our customers satisfied? Yeah.
And I truly believe that customer service is broken and I could go on a whole tangent about this. I really think that AI is coming to save the day and helping us to actually create good experiences in a customer service environment. So, and that's both for the employees and the customers.
So it's very exciting. I also think that there's a mindset here that we need to have. And I really wanted to just double click on something that you said in that AI is not going to eliminate entire workforces. And I have heard people say, we're not going to need customer service associates anymore. We're just going to be able to have AI and it's going to be automated and AI voice is going to work. What do you say to that?
Well, I think we will always need a human at the loop or human at the helm because at the end of the day, AI is very capable, but it's not smart.
And this is an important thing to understand, that it is a very capable technology. But it's not a magic wand. It's more as a wizard of Oz behind the curtains. And that wizard of Oz is the human making the final decision.
So especially with the customer-facing applications, I don't think we're there yet to just let AI run wild and you can read some of those headlines. I think it will be irresponsible
to remove the human from that loop. The final decision, the ultimate validation has to be with the human, with the user. And AI is there to help and assist our customer success representative to make their job easier, effective, faster, but not necessarily place them. I heard the term the other day, AI-adjacent.
instead of like AI led. And I think it's, it's really human led where AI, AI is a partner. AI is our bionic arm. It is not replacing the whole human. Like, yes, robots are being created, but that's a different thing for now. It's,
We are adjacent to AI. We are using AI as a tool just as we use the printing press and the steam engine to help us be more efficient. So I really think that's important because I operate in the startup world a lot and I hear founders saying, we're not going to need people to do this. AI is just going to solve it for us. And I think that even if that is possible, which at the moment it is not,
Even if that is possible, is it good for our business? Exactly. And not just from a cost perspective, from a trust perspective, which in my opinion is worth more than saving some expenses. From a trust perspective and also from a customer experience perspective. Totally. It also depends on the industry. If you are working as a support group,
for mental health or for addict recovery or for some other kind of very sensitive type of group, would you want a machine? They would not want machine to answer that. Yes, there might be some canned questions, some information that will be helpful to be provided in a timely manner.
But at the end of the day, it is such a sensitive conversation that will require human and emotion and understanding having that EQ to actually provide that experience that it's called for. And I think we have all been, I'm sure everyone listening has experienced a bot experience that just left you way more frustrated than you were in the first place. Yeah.
Just clicking zero to get to the operator. Exactly, exactly. And I mean, I think this is where I think that customer service is broken because when you call a phone number and you have to be waiting on hold and you keep pressing the numbers and you say, operator, operator, operator, speak to an agent, speak to an agent, like, please just get me through. We can solve that with AI, but we still need to have humans overseeing and being able to jump in when needed. And I think there's such that critical moment of,
training the AI to know when does a human need to be involved here? Is that something that you're working with clients on? So we are one of the teams that I have is actually responsible AI strategy team, which is a critical team because as we talked about in the very beginning,
There are a lot of human consideration and ethical consideration when it comes to AI. And one of the standard conversations that we have with our customers is risk assessment. Not to scare anyone. No, it's not a scary technology, but it is a technology that requires to have responsible use as a foundation in design of that technology. Mm-hmm.
So when we have our AI implementation conversation, having that specific conversation around impact to workforce, AI risk assessment, how it fits into the overall ecosystem and what will be the change of human experience.
That is a critical component of our engagement. And that is led by the experts in my team who are specializing in responsible AI usage. Great. I'm so happy to hear that that's a part of it. And I think that that responsibility is such a key component, again, that anyone rolling out AI needs to be thinking about in AI.
And how they can, you know, really be responsible for the human experience that is going to take place on the other side of that AI. Absolutely. What are some of the ethical considerations that you really think about that are really important to be considered in AI?
utilizing AI? So I want to start with some kind of foundation. So first of all, Salesforce already took care of some stuff. We have what we refer to as I mentioned, trust layer. So it is at the platform layer where our AI capability were developed
They were developed on the platform that is based on trust, meaning customer data is protected by design. If some prompts go outside of Salesforce, there's a zero retention policy. Information is masked.
And when a response comes back into Salesforce, first of all, it checks for toxicity. We don't want anything bad coming in. And only then information is back, kind of de-masked and put back before it's presented back in the system. So that's at the platform level. However, there's a kind of principle and practices that people need to take into account when using AI.
And this is, there are some implied biases. For example, zip code could be a proxy for race.
There are some other kind of little triggers like that. So just making users aware of what data they're using and how it could impact the responses that they get back because AI by itself, it's not biased. We're putting those biases advertly or inadvertently into our prompt, into our data. And being aware of that and making informed decision is very important. Mm-hmm.
It also depends on the industry. If you are selling socks, if your customers get socks their own color, yes, might be a, you know, fashion police might not be happy, but it's not going to make or break a deal. If you're selling medical equipment and AI sends a wrong article with wrong recommendation, that might cost a lot. Yeah.
So risk assessment also depends on the industry. There are certain government regulations like PMI are in place that we need to consider when we're working with healthcare, when we're working with financial services versus when we're working with like some retail. So all of those ethical consideration and risk assessment also vary on the industry and the risk of exposure and the risk of kind of impact.
Mm-hmm. That's great. And I'm happy to hear that those things are in place to make sure that those risks don't become reality. Nobody wants to be a headline for a wrong reason. Completely, completely. I mean, and I think that's where a lot of the fear is coming from. When I speak to my friends about AI, some folks are like, oh, this is great. Can't wait. Making life easier. And other people are like, I'm scared.
I'm scared of what's happening with my data. I'm scared of who this bot is speaking to me or what they're going to do with this information. When they see mistakes, we are so much more likely to
a company, I think, if it's an AI mistake instead of just a human mistake, because we can understand the human mistake and we can't understand the AI mistake or the hallucinations that are happening. So what is the importance of a good AI implementation when it comes to the overall business success
success and image of that business? Well, I think it's all those things that you mentioned. First of all, trust, that is the ultimate kind of beacon. And if we are not responsible and not taking that seriously, then it's very easy to violate that trust with the customer. It's very hard to get it back. Yeah. Yeah.
And that's why, as I said, when we work on the implementation, there is a set of workshops, there is a set of mechanisms that actually target it specifically to identify those potential areas of exposure so we can help customers mitigate it ahead of time that they don't get into the situation. And it's also important for Salesforce. We're saying that trust is our number one value and we're living by that.
I said nobody wants to be a headline for the wrong reason. Well, neither we want Salesforce's screenshot to be displayed somewhere with something inappropriate. So we are very, very serious and we take that accountability very seriously. And we've developed a set of mechanism, tools, and techniques to help our customers implement AI responsibly.
Great. I'm curious to know, kind of on the flip side, in terms of the possibilities of AI, have you had any favorite projects or use cases that you've thought is really at the forefront of utilizing this technology? Sure. Yes. So I'd love to share one customer that we're currently working with.
And it all started with a what-if conversation. Imagine that your salesperson just had a meeting with an important customer, leaving the restaurant, turning on his or her phone and saying, hey, just had a meeting with John Smith.
He is interested in XYZ opportunity. We were at this restaurant and would love to have a follow-up conversation. And then AI picks it up and said, John Smith, new content, interested in XYZ, new opportunity. Next action, follow up with John Smith tomorrow.
And then all that conversation turns into Salesforce opportunity in Sales Cloud with all that information appropriately entering contact field and next recommended, next best action, etc. One of the feedback that we got that salespeople do not like entering nodes. I'm very aware. Yeah.
Especially when they're after dinner and then getting in the car, you have to remember what you said, getting back to your computer, having that voice-to-text conversation via mobile interface.
keeps your system up to date, keeps your data up to date, and keeps that human up to date with the recommendation, next batch action. So this is the conversation. This is the ultimate implementation that we will be achieving for this given customer. However,
What I loved about this program, that it opened significantly broader opportunity. As we started talking, we've discovered that our team can do so much more for this customer. We've learned that we can actually help them optimize entire sales process.
We've learned that their leads, they're kind of saying, yeah, sometimes I start here with my lead, sometimes I start there. Well, by enabling our already available capability, predictive AI with lead prioritization scoring, we're going to give their sellers a systematic data-based approach to target their leads, to allow them to focus on the top priority, highest possibility leads, and to
versus kind of being all over the place or randomly selecting. So then we learned that they have a lot of external sources that they're basing their decision on. Well, by bringing data cloud in and by using our integration platform called MuleSoft, we're actually able to integrate that disparate external sources and make it available in data cloud
for Salesforce system decision-making. Mm-hmm.
And along the way, we're going to start enabling those capabilities. So yes, maybe voice-to-text will be a later implementation. But as we implement those incremental improvements, we'll be able to enable call summary for our sellers and some of the other capabilities. So what makes this project exciting for me is that we actually started with this aspirational conversation
But it's only a small part of a significantly broader solution. And we're partnering with this particular customer to truly reimagine their sales process
to help them improve their profitability by using a full power of Salesforce with AI in the mix. And this is amazing to see this transformation and to see the value which I... Oh my gosh. Very rewarding. Yes, I can understand that. Just as you're saying, I'm like, hallelujah. As a...
I mean, as a customer success and customer experience leader, I am always on the sales team to input their notes and
And it is always a struggle. I can't tell you how many sales teams I have trained on how to input notes correctly because at the end of the day, it's good for them, but it's really good for my team. It's really good for the team who's taking those customers post-sale to understand what happened pre-sale that we can then build on and how can we grow this relationship from that moment. I think another interesting thing that you said is lead prioritization.
And again, selfishly thinking about how this works for customer service or sorry, customer success and customer service is when we find out, oh, here's the customers that we really find are easy to retain. There's a really good fit for their needs. These are the types of customers that we want more of. We can help the sales team to automatically prioritize those attributes and
And focus on those customers that we're consistently defining the ICP. And we need to make sure that that's a part of the sales process. And we're not just bringing in leads that are only going to last so long. We want to focus on retention and customers that will actually grow with us. So...
These are really exciting advancements. I'm so happy to hear that that's what you've been working on. So starting to close things out a little bit, I'm curious to know what advice do you have with leaders who are prepping for AI implementation? We've talked about a lot of things, but if you can kind of dwindle it down into kind of some core bite-sized pieces, what would you say? No, absolutely. So I would talk about a systematic approach. First of all, don't go for a checkbox.
And sometimes I hear, oh, we already have Gen AI, we enabled XYZ capability, checkbox is done. But is it really generating value? Customers are under a lot of pressure from CEOs, CIOs to get going with AI. Mm-hmm.
But I think my advice is that don't be deterred by that. Truly start with what is the business goal? What are the objectives, truly business objectives that you're trying to achieve within a given year or a few years ahead? And what are the obstacles? What problems do we need to solve?
And then second, having that the right technology, even if it's an automated technology of any AI flavor to solve that problem. And then have, yes, do have an ultimate roadmap for both data and AI, but approach it as a sidestep, not as a straight line. Go slowly in manageable kind of
bytes on the maturity ladder of both AI and data so that you can get incremental value along the way. The best way to ensure that company continues to invest in AI, believe in its power, is to show that incremental value along the way.
Another important aspect, actually a few of them, is setting in the foundation. Understanding what underlying technology you need to make that ultimate goal happen. Maybe as a customer, you do need to find your data solution or your integration solution. You can't make real-time decision based on yesterday's data.
So even though you have enough good data to get started, what is that ultimate data strategy, data roadmap, and putting that foundational technology in along the way to get ready for the ultimate solution? And last but not least, definitely not least when we talk about change management and governance. I would especially recommend it for large enterprise customers.
we talk to them about standing up a dedicated AI practice or dedicated AI center of excellence to have the right foundation and the right mechanism to have that governance and decision making, propagation of knowledge in place. So for somebody who is looking to
stand up this program enterprise-wide and maybe beyond Salesforce, we talk about standing up COE or practice as part of the foundational elements. And that would also help with governance and as I said, and change management.
Great. Thank you. There's a lot to think about. But it's also, this is an investment for now and into the future. And it's really a mindset shift. We're approaching business differently and we're using a technology that is changing rapidly. So how do we really implement things in a way that acknowledges that inevitable change? Yeah. That's great. And
When you think about driving adoption within your clients and getting people to actually adopt this new technology, do you have any tips or advice for leaders on how to get their teams on board? Yes. I don't really believe in do-it-yourself projects for a reason, because on the surface, everything looks like a turn switch. And yes, some of the capability is truly out of the box.
However, we as humans are not going to say, oh, I don't know how to use it or maybe I need more training. We're going to say this technology is no good. It doesn't do what I need it to do. And then it becomes shelfware. So I feel that proper implementation, proper training, proper change management and enablement
is key to adoption. Companies invest a lot of money in technology, and it will be a shame for it to be a shelfware. I feel that it's as equally important to invest into proper implementation. We want to make sure that our teams are using this on a daily basis. The potential is so huge, but we need to be using it to really learn. And that means we need
We need to enable people to actually use it the right way completely. I'm curious to know what you personally use to keep yourself up to date on all things AI. I mean, I've mentioned a couple of times it is changing so fast. And I'm curious to know what resources, blogs, podcasts, thought leaders you follow or listen to to help you stay up to date.
So I'm lucky in the regards that my entire team is dedicated to AI. And everyone reads something and then shares with the rest of the group. There are some obvious sources such as like McKinsey Digital or some of the respectable institutions such as Harvard and MIT. But
As I said, it's also my team is another source. I have an international team, global team that everyone have their favorite source.
And by using all those sources and sharing this information back, it provides such vast opportunity to stay up to date. And then we can use Slack AI to help us summarize it. Great. I think it's such a good idea to have an AI knowledge group within your work team. Or I mean, I'm even thinking maybe I should do this with like other consultants that I work with because, yeah,
there's just so much out there. So I love that, that you're really supporting one another. You've been in this space for a while and I'm curious to know, what are your predictions for the future? Like what can we expect in the next five, 10 years when it comes to AI and how it evolves? It is a really hard question because it is such a
fast-paced changing technology that has ability to learn and improve itself. I know people are saying that the next horizon will be autonomous agents when AI is talking to other AI.
So according to academics and technology, that is the future. I think that given the nature of technology, some of the progress and capabilities will be balanced by ethics, risk assessment, and human in the loop.
So even if we get to some of the autonomous capabilities, I think we will recognize that there is absolute need for balance. It's just moving so quickly and it could go. Everyone's like, is this going to go in the wrong direction? But I think keeping humans in the loop is going to be the essential piece of making sure that the way that this progresses is actually in our best interest.
So my last question for you is, what is one piece of advice that you think every customer experience leader should hear? With regards to AI, I would say don't be afraid to experiment, to get started.
But be mindful what this technology is for. As we talked about before, right, the AI is capable, not smart. It's not an individual. It's a tool. So treat it for what it is. Understand the impact that it has on the overall ecosystem.
and build it in into that environment, knowing what it is and how it acts and interacts with other components. Take it as an incremental journey and onward and forward. Thank you for saying that. This is really something we need to test and learn and consistently reiterate on. And that's very, very helpful advice.
Well, thank you so much, Irina, for coming on the show. This has been so fascinating to dive into the world of AI and implementing it within our teams, how to do it correctly, and the way that Salesforce has really been pioneering incredible technological advancements that are truly helping so many teams around the world operate better, faster, and stronger. So thank you for the work you do, Irina, and thanks for coming on the show. Thank you for having me.
Salesforce Data Cloud is a hyperscale data platform built right into Salesforce. Data Cloud helps you to better know your customers, act on AI in the flow of work, and enrich your data with open access. To learn more, visit salesforce.com slash products slash data.