Podcast Summary
From AI models in customer service to leading Salesforce's AI efforts: Clara Shi's background in AI and ML, gained from her experience at Hearsay Social and Service Cloud, prepared her for her current role as CEO of Salesforce AI. She saw the potential of AI models in customer service and expanded Salesforce's efforts to include AI Copilot and agent platform, as well as partnerships and ecosystems.
Clara Shi's background in AI and ML, gained from her experience at Hearsay Social and Service Cloud, prepared her well for her current role as CEO of Salesforce AI. When she joined Service Cloud, they were using early AI models for customer service, and as open AI models improved, she and her team saw the potential for these models to be a core part of Service Cloud. They began experimenting with prototypes, such as Service GPT, and when ChatGPT was launched, Salesforce saw the opportunity to apply large language models to every cloud, leading to the expansion of AI efforts across Salesforce. Clara now leads these efforts, including the development and implementation of AI Copilot and agent platform, as well as partnerships and ecosystems. Her role has evolved in response to the rapid innovation in generative AI and Salesforce's commitment to staying at the forefront of this technology in the enterprise.
Salesforce's Approach to AI and ML: Flexibility, Empowerment, and Comprehensive Solutions: Salesforce is providing a flexible platform for AI and ML services, allowing customers to choose between in-house, third-party, and external models. They are also collaborating with customers and offering integrations with third-party providers to cater to diverse needs.
Salesforce is committed to creating a common platform for AI and ML services, offering a range of solutions from model development to fine-tuning, and providing customers with choice between in-house, third-party, and external models. They are taking an open architecture approach to serve their diverse customer base, which includes large enterprises, SMBs, and those who prefer managed solutions. Salesforce's efforts include developing their own models, collaborating with customers, and offering integrations with third-party providers such as Anthropic, Cohere, OpenAI, and Google Vertex. They are also exploring agent-based platforms and copilots, recognizing the nascent stage of these technologies but seeing great potential for their evolution. Salesforce's strategy is to offer flexibility, empower customers, and provide a comprehensive AI/ML solution.
Salesforce's AI Expansion: Customizable Prompts, Copilot, and Einstein Studio: Salesforce is enhancing its AI capabilities with customizable prompts, Copilot platform, and Einstein Studio, enabling enterprises to integrate AI into their service, sales, and marketing processes.
Salesforce is rapidly expanding its AI capabilities, integrating AI features into every existing Salesforce Cloud and providing tools for customers to customize and build upon these features. This includes prompt templates for service, sales, and marketing, as well as the Copilot platform with its components: prompt builder, action builder, and Einstein Studio. The prompt builder allows customers to customize templates, point them to different models, and ground them in unique data. Action builder empowers the copilot with agent powers, enabling the use of workflows, integrations, and sharing rules. Einstein Studio allows customers to train or fine-tune their own predictive or generative models using their Salesforce data. Salesforce has already launched pilots for prompt builder and is seeing incredible feedback, showcasing the speed at which these advancements are being made. Despite the industry's recent focus on generative AI, Salesforce has been making strides in AI for some time, and the adoption rate by enterprises is a topic of interest.
Enterprises are in the experimentation phase of AI adoption, consolidating data is crucial: Enterprises are experimenting with AI, but data consolidation is essential for wider adoption, Salesforce's data cloud growth highlights this trend
While there is already significant adoption of generative AI in various industries and use cases, especially in customer service, the majority of enterprises are still in the experimentation phase. The key challenge lies in bringing all the disparate data sources together to effectively power these generative use cases. Salesforce, for instance, has recently introduced 0 ETL data sharing partnerships with BigQuery, Databricks, and Snowflake to help customers consolidate their data. From a broader perspective, the enterprise adoption of AI is in its second or third inning, with a few pioneering companies demonstrating substantial business process transformations. However, most enterprises are still in the process of organizing their data, making it a crucial first step towards wider AI adoption. Salesforce's data cloud is growing rapidly as a result, marking a significant milestone in the AI journey for the enterprise sector.
Starting AI implementation with organized data: Organizing data is crucial for implementing AI in an organization. It enables internal tooling, efficiency gains, and external customer use. Collaborative efforts with other teams can lead to improved AI end-products and enhanced user experiences.
Implementing AI in an organization starts with getting data in order. Once the data is organized, it can be used for internal tooling and efficiency gains, or prototyped for external use. The next step is to integrate AI into end products or services for customers. This process can be collaborative and involve educating other teams about the potential of AI. For instance, in the customer service world, having all knowledge articles consolidated using data cloud can lead to better chatbot responses. AI is transforming software development by allowing dynamic handling of user experiences, reducing the need for hard-coded branches and screens. Salesforce's approach to AI has been a collaborative effort, with many ideas coming from across the organization. AI is revolutionizing the way we interact with software, and it's important to remember that it's not just about making things more efficient internally, but also about enhancing the end-user experience. Although Salesforce's earlier acquisitions may not have been explicitly focused on AI, they are now being leveraged to drive AI capabilities forward.
Generative AI transforms enterprise software UX: Generative AI enhances UX in enterprise software through real-time data visualization, personalized customer service, and efficient interactions.
The integration of generative AI and user experience (UX) in enterprise software, such as Einstein GPC's generative canvas and Slack, is revolutionizing the way users interact with data and customer service. From a product UX standpoint, generative canvas allows for real-time visualization and updating of data from various sources like Salesforce reporting and Tableau. It's a significant departure from the hard-coded and hardwired components of the past. Additionally, the day-to-day experience of users, particularly customer service representatives, is being transformed. For instance, Gucci's service representatives are benefiting from retrieval augmentation, which arms them with the right brand storytelling and troubleshooting to provide better customer service. The result is a decrease in average handle time and an opportunity for deeper, more personalized conversations with customers. Overall, these advancements represent a shift towards more intuitive and efficient interactions with enterprise software.
Shifting focus from roles to customer needs and managing unstructured data for AI products: Businesses are prioritizing customer needs over traditional roles, managing unstructured data for AI applications, and embracing AI's effectiveness to revolutionize enterprise software.
Businesses are shifting their focus from traditional department roles to understanding and addressing customer needs, empowering individuals with the necessary knowledge to do so. This includes managing enterprise data for AI products, particularly dealing with unstructured data. Some unstructured data, like PRDs or service articles, can be used directly, while other forms, like transcripts, require additional processing. The future of enterprise software is expected to be significantly impacted by AI, with potential changes in business models and user interactions. The most unexpected thing emerging from generative AI is its effectiveness, and it's predicted that it will continue to revolutionize enterprise software in the coming years, much like the introduction of cloud technology did.
Balancing the shift to cloud and AI: While some apps move to cloud and AI for flexibility, others stay on-premise for determinism. Engineers, PMs, and designers prescribe why and what, while AI handles how. High costs of AI are a challenge, but demonstrating value and ROI can help. Optimism exists for net new capabilities and productivity gains, with costs decreasing over time.
While some applications are ideal for moving to the cloud and leveraging AI for on-demand access and flexibility, others require deterministic decision-making and may remain on-premise. The role of software engineers, product managers, and designers is shifting towards prescribing the why and what, and allowing AI to determine the how. However, the cost of AI products, which can be significant due to compute requirements, is a challenge. Salesforce aims to strike a balance by covering costs while keeping pricing understandable for customers. The key is to demonstrate value and ROI, such as reducing average handle time and driving sales conversion uplift. Despite the complexity, there is optimism due to the potential for net new capabilities and productivity gains, as well as decreasing costs over time as AI technology improves. An extreme example of this is RunwayML's involvement in the movie "Everything Everywhere All at Once," which showcases the potential of AI to create unique and valuable experiences.
Generative AI revolutionizes film industry with smaller teams and cost savings: Generative AI enables smaller teams to create complex special effects, leading to significant cost savings for studios. Foundational models and domain-specific startups are promising areas for innovation and tooling improvements are needed for effective usage.
Generative AI is revolutionizing the film industry by enabling smaller teams to create complex special effects, leading to significant cost savings. This was highlighted in the example of a movie that required only 7 people on the video editing team instead of the traditional 700. From a business standpoint, this is an efficient and cost-effective solution for studios and companies like Runway, which are capitalizing on this technology. For startups, there are several areas where generative AI can be applied. Foundational models and domain-specific startups are particularly interesting, as they have the potential to address a wide range of needs and industries. Additionally, there is a need for better tooling and applications to help organizations effectively utilize this technology. Salesforce and other large companies have made strides in this area, but there is still much to be discovered and developed. As a founder, focusing on these areas could lead to innovative and successful businesses.
