From smarter search engines to AI-powered creative workflows, vectors are the bridge between unstructured data and actionable intelligence. By fully leveraging this technology, businesses can transform their approach to data and secure their place in a rapidly evolving, data-centric world.
Salesforce Agentforce: Redefining Autonomous AI for the Enterprise Salesforce has unveiled Agentforce, a groundbreaking addition to its platform that enables businesses to create and deploy autonomous AI agents capable of performing tasks across any business function. Unlike traditional chatbots or copilots, Agentforce leverages advanced reasoning capabilities to independently resolve customer cases, qualify sales leads, optimize marketing campaigns, and more. By automating actions triggered by data changes, business rules, pre-built automations, or API signals, Agentforce empowers companies like OpenTable, Saks, and Wiley to scale their workforce and enhance customer experiences. “Agentforce is redefining what’s possible in business and beyond,” said Marc Benioff, Chair and CEO of Salesforce. “This technology is ushering in an era of AI abundance and limitless workforces, augmenting employees, building deeper customer relationships, and driving unprecedented growth and profitability.” Built on a Metadata-Driven Architecture Agentforce builds on Salesforce’s metadata-driven platform, ensuring rapid deployment and customization without extensive manual effort. “Agentforce taps into the trusted Salesforce platform, seamlessly integrating enterprise data, workflows, and AI models to deliver actionable intelligence,” said Srini Tallapragada, President and Chief Engineering Officer at Salesforce. Addressing Market Inefficiencies Salesforce research identified significant inefficiencies Agentforce is designed to address: Marc Benioff emphasized, “The agent revolution is as transformative as the cloud, social, and mobile revolutions, setting a new standard for operational efficiency.” Rapid Adoption and Industry Use Cases At Dreamforce 2024, customers created over 10,000 autonomous agents in three days, with adoption accelerating to 20,000 agents globally through Salesforce World Tour events. The platform’s natural language interface enables business users—not just IT teams—to build and customize agents. Industry Applications: Ecosystem and Partner Support The launch is bolstered by partnerships with global systems integrators, including Accenture, Deloitte, IBM, PwC, and Slalom. These collaborations have accelerated emerging technology rollouts by 31%, according to Salesforce. Additionally, the new Agentblazer Community provides a hub for knowledge sharing and integrations via the Salesforce AppExchange. Pricing and Future Development Agentforce is now generally available, with pricing starting at $2 per conversation, including volume discounts. Customers can explore the platform through the Salesforce Foundations trial, offering 1,000 conversation credits for free. Future development plans focus on enabling multi-agent scenarios for complex workflows and expanding autonomous capabilities across business processes. “This is just the first chapter,” said Clara Shih, CEO of Salesforce AI. “The innovation we’ll see from this technology will transform businesses in ways we’ve only begun to imagine.” What This Means for Businesses Agentforce represents Salesforce’s most ambitious platform expansion since Einstein AI’s debut in 2016. Its consumption-based pricing model aligns with Salesforce’s broader shift toward pay-as-you-use services, mirroring strategies with Data Cloud and Einstein GPT. For enterprises, the platform offers a low-risk trial to assess its potential for driving operational improvements. The challenge will be transitioning from pilot projects to large-scale deployments while maintaining the claimed accuracy and efficiency advantages. With over 20,000 agents created and a robust partner ecosystem, Salesforce is well-positioned to lead the autonomous AI revolution. However, the ultimate success of Agentforce will depend on its ability to consistently deliver measurable business value beyond the initial wave of adoption. Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more The Salesforce Story In Marc Benioff’s own words How did salesforce.com grow from a start up in a rented apartment into the world’s Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more
From Science Fiction to Reality: AI’s Game-Changing Role in Service and Sales AI for service and sales has reached a critical tipping point, driving rapid innovation. At Dreamforce in San Francisco, hosted by Salesforce we explored how Salesforce clients are leveraging CRM, Data Cloud, and AI to extract real business value from their Salesforce investments. In previous years, AI features branded under “Einstein” had been met with skepticism. These features, such as lead scoring, next-best-action suggestions for service agents, and cross-sell/upsell recommendations, often required substantial quality data in the CRM and knowledge base to be effective. However, customer data was frequently unreliable, with duplicate records and missing information, and the Salesforce knowledge base was underused. Building self-service capabilities with chatbots was also challenging, requiring accurate predictions of customer queries and well-structured decision trees. This year’s Dreamforce revealed a transformative shift. The advancements in AI, especially for customer service and sales, have become exceptionally powerful. Companies now need to take notice of Salesforce’s capabilities, which have expanded significantly. Agentforce – AI’s New Role in Sales and Service Some standout Salesforce features include: At Dreamforce, we participated in a workshop where they built an AI agent capable of responding to customer cases using product sheets and company knowledge within 90 minutes. This experience demonstrated how accessible AI solutions have become, no longer requiring developers or LLM experts to set up. The key challenge lies in mapping external data sources to a unified data model in Data Cloud, but once achieved, the potential for customer service and sales is immense. How AI and Data Integrate to Transform Service and Sales Businesses can harness the following integrated components to build a comprehensive solution: Real-World Success and AI Implementation OpenTable shared a successful example of building an AI agent for its app in just two months, using a small team of four. This was a marked improvement from the company’s previous chatbot projects, highlighting the efficiency of the latest AI tools. Most CEOs of large enterprises are exploring AI strategies, whether by developing their own LLMs or using pre-existing models. However, many of these efforts are siloed, and engineering costs are high, leading to clunky transitions between AI and human agents. Tectonic is well-positioned to help our clients quickly deploy AI-powered solutions that integrate seamlessly with their existing CRM and ERP systems. By leveraging AI agents to streamline customer interactions, enhance sales opportunities, and provide smooth handoffs to human agents, businesses can significantly improve customer experiences and drive growth. Tectonic is ready to help businesses achieve similar success with AI-driven innovation. Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more The Salesforce Story In Marc Benioff’s own words How did salesforce.com grow from a start up in a rented apartment into the world’s Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more
The first notable change in the field of language models is the significant expansion of context window sizes and a reduction in token costs. For instance, Anthropic’s largest model, Claude, has a context window exceeding 200,000 tokens, while recent reports indicate that Gemini’s context window can reach up to 10 million tokens. Under such circumstances, Retrieval-Augmented Generation (RAG) may no longer be necessary for many tasks, as all required data can be accommodated within the expanded context window. Several financial and analytical projects have already demonstrated that tasks can be solved without needing a vector database as intermediate storage. This trend of reducing token costs and increasing context window sizes is likely to continue, potentially decreasing the need for external mechanisms in LLMs, although they are still relevant for the time being. If the context window remains insufficient, methods for summarization and context compression have been introduced. LangChain, for example, offers a class called ConversationSummaryMemory to address this challenge. pythonCopy codellm = OpenAI(temperature=0) conversation_with_summary = ConversationChain( llm=llm, memory=ConversationSummaryMemory(llm=OpenAI()), verbose=True ) conversation_with_summary.predict(input=”Hi, what’s up?”) Knowledge Graphs As the volume of data continues to grow, navigating through it efficiently becomes increasingly critical. In certain cases, understanding the structure and attributes of data is essential for effective use. For example, if the data source is a company’s wiki, an LLM might not recognize a phone number unless the structure or metadata indicates that it’s the company’s contact information. Humans can infer meaning from conventions, such as the subdirectory “Company Information,” but standard RAG may miss such connections. This challenge can be addressed by Knowledge Graphs, also known as Knowledge Maps, which provide both raw data and metadata that illustrates how different entities are interconnected. This method is referred to as Graph Retrieval-Augmented Generation (GraphRAG). Graphs are excellent for representing and managing structured, interconnected information. Unlike vector databases, they excel at capturing complex relationships and attributes among diverse data types. Creating a Knowledge Graph The process of creating a knowledge graph typically involves collecting and structuring data, which requires expertise in both the subject matter and graph modeling. However, LLMs can automate a significant portion of this process by analyzing textual data, identifying entities, and recognizing their relationships, which can then be represented in a graph structure. In many cases, an ensemble of vector databases and knowledge graphs can improve accuracy, as discussed previously. For example, search functionality might combine keyword search through a regular database (e.g., Elasticsearch) and graph-based queries. LangChain can also assist in extracting structured data from entities, as demonstrated in this code example: pythonCopy codedocuments = parse_and_load_data_from_wiki_including_metadata() graph_store = NebulaGraphStore( space_name=”Company Wiki”, tags=[“entity”] ) storage_context = StorageContext.from_defaults(graph_store=graph_store) index = KnowledgeGraphIndex.from_documents( documents, max_triplets_per_chunk=2, space_name=space_name, tags=[“entity”] ) query_engine = index.as_query_engine() response = query_engine.query(“Tell me more about our Company”) Here, searching is conducted based on attributes and related entities, instead of similar vectors. If set up correctly, metadata from the company’s wiki, such as its phone number, would be accessible through the graph. Access Control One challenge with this system is that data access may not be uniform. For instance, in a wiki, access could depend on roles and permissions. Similar issues exist in vector databases, leading to the need for access management mechanisms such as Role-Based Access Control (RBAC), Attribute-Based Access Control (ABAC), and Relationship-Based Access Control (ReBAC). These access control methods function by evaluating paths between users and resources within graphs, such as in systems like Active Directory. To ensure the integrity of data during the ingestion phase, metadata related to permissions must be preserved in both the knowledge graph and vector database. Some commercial vector databases already have this functionality built in. Ingestion and Parsing Data needs to be ingested into both graphs and vector databases, but for graphs, formatting is especially critical since it reflects the data’s structure and serves as metadata. One particular challenge is handling complex formats like PDFs, which can contain diverse elements like tables, images, and text. Extracting structured data from such formats can be difficult, and while frameworks like LLama Parse exist, they are not always foolproof. In some cases, Optical Character Recognition (OCR) may be more effective than parsing. Enhancing Answer Quality Several new approaches are emerging to improve the quality of LLM-generated answers: While these advancements in knowledge graphs, access control, and retrieval mechanisms are promising, challenges remain, particularly around data formatting and parsing. However, these methods continue to evolve, enhancing LLM capabilities and efficiency. Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more Alphabet Soup of Cloud Terminology As with any technology, the cloud brings its own alphabet soup of terms. This insight will hopefully help you navigate Read more
Guide to RAG (Retrieval-Augmented Generation) Retrieval-Augmented Generation (RAG) has become increasingly popular, and while it’s not yet as common as seeing it on a toaster oven manual, it is expected to grow in use. Despite its rising popularity, comprehensive guides that address all its nuances—such as relevance assessment and hallucination prevention—are still scarce. Drawing from practical experience, this insight offers an in-depth overview of RAG. Why is RAG Important? Large Language Models (LLMs) like ChatGPT can be employed for a wide range of tasks, from crafting horoscopes to more business-centric applications. However, there’s a notable challenge: most LLMs, including ChatGPT, do not inherently understand the specific rules, documents, or processes that companies rely on. There are two ways to address this gap: How RAG Works RAG consists of two primary components: While the system is straightforward, the effectiveness of the output heavily depends on the quality of the documents retrieved and how well the Retriever performs. Corporate documents are often unstructured, conflicting, or context-dependent, making the process challenging. Search Optimization in RAG To enhance RAG’s performance, optimization techniques are used across various stages of information retrieval and processing: Python and LangChain Implementation Example Below is a simple implementation of RAG using Python and LangChain: pythonCopy codeimport os import wget from langchain.vectorstores import Qdrant from langchain.embeddings import OpenAIEmbeddings from langchain import OpenAI from langchain_community.document_loaders import BSHTMLLoader from langchain.chains import RetrievalQA # Download ‘War and Peace’ by Tolstoy wget.download(“http://az.lib.ru/t/tolstoj_lew_nikolaewich/text_0073.shtml”) # Load text from html loader = BSHTMLLoader(“text_0073.shtml”, open_encoding=’ISO-8859-1′) war_and_peace = loader.load() # Initialize Vector Database embeddings = OpenAIEmbeddings() doc_store = Qdrant.from_documents( war_and_peace, embeddings, location=”:memory:”, collection_name=”docs”, ) llm = OpenAI() # Ask questions while True: question = input(‘Your question: ‘) qa = RetrievalQA.from_chain_type( llm=llm, chain_type=”stuff”, retriever=doc_store.as_retriever(), return_source_documents=False, ) result = qa(question) print(f”Answer: {result}”) Considerations for Effective RAG Ranking Techniques in RAG Dynamic Learning with RELP An advanced technique within RAG is Retrieval-Augmented Language Model-based Prediction (RELP). In this method, information retrieved from vector storage is used to generate example answers, which the LLM can then use to dynamically learn and respond. This allows for adaptive learning without the need for expensive retraining. Guide to RAG RAG offers a powerful alternative to retraining large language models, allowing businesses to leverage their proprietary knowledge for practical applications. While setting up and optimizing RAG systems involves navigating various complexities, including document structure, query processing, and ranking, the results are highly effective for most business use cases. Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more Alphabet Soup of Cloud Terminology As with any technology, the cloud brings its own alphabet soup of terms. This insight will hopefully help you navigate Read more
Significant Changes in Context Windows and Token Costs The first significant change in the AI world is the substantial increase in the context window size and the decrease in token costs. For example, the context window size of the largest model, Claude from Anthropic, exceeds 200,000 tokens. According to the latest news, Gemini’s context window can reach up to 10 million tokens. Under these conditions, Retrieval-Augmented Generation (RAG) may not be required for many tasks, as all necessary data can fit into the context window. This shift has been observed in several financial and analytical projects where tasks were completely solved without using a vector database as intermediate storage. The trend of reducing token costs and increasing context window sizes is likely to continue, diminishing the need for external mechanisms for large language models (LLMs). However, they remain necessary for the time being. Advanced RAG. If the context size is still insufficient, different methods of summarization and context compression have been devised. LangChain has introduced a class aimed at this: ConversationSummaryMemory. pythonCopy codellm = OpenAI(temperature=0) conversation_with_summary = ConversationChain( llm=llm, memory=ConversationSummaryMemory(llm=OpenAI()), verbose=True ) conversation_with_summary.predict(input=”Hi, what’s up?”) Knowledge Graphs As the amount of data LLMs must navigate grows, the ability to navigate this data becomes increasingly important. Without the ability to analyze the data structure and other attributes, it’s impossible to use them effectively. For example, suppose the data source is a company’s wiki with a page containing the company’s phone number, but this isn’t explicitly indicated anywhere. How does the LLM understand that this is the company’s phone number? It doesn’t, which is why standard RAG won’t provide any information about the company’s phone number (as it sees no connection). A person can understand that this is the company’s phone number from the convention of how the data is stored (i.e., from the structure or metadata). For LLMs, this problem is solved with Knowledge Graphs with metadata (also known as Knowledge Maps), which means the LLM has not only the raw data but also information about the storage structure and the connections between different data entities. This approach is also known as Graph Retrieval-Augmented Generation (GraphRAG). Graphs are excellent for representing and storing heterogeneous and interconnected information in a structured form, easily capturing complex relationships and attributes among different types of data, which vector databases struggle with. Example of a Knowledge Graph Creating a Knowledge Graph typically involves collecting and structuring data, requiring a deep understanding of both the subject area and graph modeling. This process can largely be automated with LLMs. Thanks to their understanding of language and context, LLMs can automate significant parts of the Knowledge Graph creation process. By analyzing textual data, these models can identify entities, understand their relationships, and suggest how best to represent them in a graph structure. Advanced RAG This ensemble of a vector database and a knowledge graph generally improves accuracy and often includes a search through a regular database or by keywords (e.g., Elasticsearch). Knowledge Graph Retriever Example For example, a user asks a question about the company’s phone number. If this is done in code, the entities from the question can be formatted in JSON or using with_structured_output from LangChain. These entities are then searched for in the Knowledge Graph. How this is done depends on where the graph is stored. pythonCopy codedocuments = parse_and_load_data_from_wiki_including_metadata() graph_store = NebulaGraphStore( space_name=”Company Wiki”, tags=[“entity”] ) storage_context = StorageContext.from_defaults(graph_store=graph_store) index = KnowledgeGraphIndex.from_documents( documents, max_triplets_per_chunk=2, space_name=”Company Wiki”, tags=[“entity”] ) query_engine = index.as_query_engine() response = query_engine.query(“Tell me more about our Company”) The search differs from a vector database search in that it searches for attributes and related entities, not similar vectors. Returning to the initial question, if the wiki structure was transferred to the graph correctly, the company’s phone number would be added as a related entity in the graph. The data from the graph and the vector database search are then passed to the LLM to generate a complete answer. Challenges and Solutions Access Control Access to data may not be uniform. In the same wiki, there may be roles and permissions, and not every user can see all information. This problem exists for both graph and vector database searches, requiring access management. Role-Based Access Control (RBAC), Attribute-Based Access Control (ABAC), and Relationship-Based Access Control (ReBAC) are common methods. Permissions and categories are also forms of metadata, which must be preserved at the data ingestion stage in the knowledge graph and vector database. When searching in the vector database, it is necessary to check whether the role or other access attributes match what the user has access to. Some commercial vector databases already include this functionality. Data embedded in the LLM during training relies on the LLM’s reasonableness, which is not recommended. Ingestion and Parsing Data needs to be inserted into the graph and the vector database. For the graph, the format is critical as it reflects the data structure and serves as metadata. Parsing data, especially from PDFs, can be challenging. Frameworks like LLama Parse attempt this with varying degrees of success. However, OCR or recognizing a document image can sometimes be easier. Improving Answer Advanced RAG Several approaches aim to improve answer quality beyond using knowledge graphs: Corrective Retrieval Augmented Generation (CRAG) CRAG addresses incorrect RAG results by automating correction processes. LangGraph can implement this approach, which essentially forms a state machine. Self-RAG Self-reflective RAG fine-tunes the LLM to generate self-reflection tokens in addition to regular ones, helping build a state machine for better results. HyDe HyDe (Hypothetical Document Embeddings) modifies the usual RAG retrieval process by using the LLM to generate a response and then searching the vector database with that response. This is useful when users’ questions are too abstract and require more context. These methods, including CRAG, Self-RAG, and HyDe, provide various ways to enhance the performance of LLMs and improve the quality of their answers. Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM
Best Practices for Building a RAG LLM Chatbot In today’s competitive landscape, a robust Retrieval-Augmented Generation (RAG) Large Language Model (LLM) chatbot can provide a significant edge. However, implementing such a system involves multiple steps and challenges. Below is a guide to help you navigate this journey effectively. 1. Define Clear Objectives Start by identifying your goals: A clear purpose will drive your design and implementation decisions. 2. Prepare Your Data Your chatbot’s success relies heavily on the quality and structure of your data. Assess Your Data Format: Enrich Your Data: 3. Select the Right Platform Choose a platform based on your goals and data type. Common architectures include: Standard RAG with Vector Databases: Relational Databases: Hybrid Text Search: Graph Databases: 4. Fine-Tune the Model While RAG often enriches data without altering the base model, fine-tuning may be essential in specific scenarios, such as: You can fine-tune open-source models directly or use OpenAI’s APIs for custom training. OpenAI also offers domain-specific GPTs, which can save development time. 5. Master Prompt Engineering Crafting precise prompts is critical to guide LLM responses. Prompts generally include: Example for Elasticsearch: javascriptCopy codeYour task is to create a valid Elasticsearch DSL query. Given the mapping: “`{mapping}“`, translate the query: “`{query}“` into JSON. – Use only fields from the mapping. – Ensure case-insensitivity and support fuzzy matches. – Compress the JSON output, removing spaces. For advanced prompt strategies, see Modern Advances in Prompt Engineering. 6. Iterative Testing Testing machine learning models is inherently complex. Key strategies: 7. Optimize the Front-End Collaborate with UI/UX teams to integrate LLM capabilities seamlessly. Consider: 8. Avoid Common Pitfalls Looking Ahead The rise of LLMs has transformed how we think about search and chatbot applications. While we’re in the early stages, the possibilities are immense. Businesses that adapt and strategically integrate these technologies will unlock innovative opportunities and thrive in the evolving landscape. Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more Top Ten Reasons Why Tectonic Loves the Cloud The Cloud is Good for Everyone – Why Tectonic loves the cloud You don’t need to worry about tracking licenses. Read more
Deep Dive Summer 24 Release Get ready, Salesforce fans! The Summer ’24 release is here, and it’s like Christmas morning for tech geeks. We’re talking about new features, enhancements, and improvements that will make you wonder how you ever lived without them. This Tectonic insight is your ultimate guide to all the exciting updates, changes, and key considerations for this release. So hang on tight to your keyboards and let’s dive into the Christmas treat bag of goodies coming your way! Key Highlights – Deep Dive Summer 24 Release What’s New in Einstein AI? 1. Einstein for Flow Meet your new best friend for building Salesforce workflows, Salesforce Flow. Just describe what you need in plain English, and Einstein will whip up the workflow for you. For example, say “Notify sales reps when a lead converts,” and boom, it’s done. Automation just got a whole lot easier and way cooler. How to: Einstein for Flow makes complex processes feel like a walk in the park, letting you deliver solutions faster than you can say “workflow.” Considerations: 2. Einstein for Formulas No more tearing your hair out over formula syntax errors. Einstein for Formulas will not only tell you what’s wrong but also suggest fixes, saving you from endless hours of debugging. How to: Einstein for Formulas cuts down errors and speeds up formula creation, making your life exponentially easier. Like easier squared. Easier to the nth degree. Considerations: UI/UX Enhancements 1. Add New Custom Fields to Dynamic Forms-Enabled Pages Say goodbye to limitations! You can now add new custom fields directly to Dynamic Forms-enabled pages, aligning fields with your ever-changing business needs. Considerations: 2. Use Blank Spaces to Align Fields on Dynamic Forms-Enabled Pages Finally, a way to make your Dynamic Forms pages look neat and tidy with blank spaces for perfect alignment. Considerations: 3. Set Conditional Visibility for Individual Tabs in Lightning App Builder Now you can make specific tabs visible based on user profiles, record types, or other criteria. Customization just got a whole lot more precise. Considerations: 4. Create Rich Text Headings in Lightning App Builder Make your headings pop with bold, italic, and varied font sizes. Your Lightning pages are about to get a visual upgrade. Considerations: Flow Updates 1. Automation Lightning App A one-stop shop for managing and executing all your automation tools and processes. Considerations: 2. Lock and Unlock Records with Action Gain more control over your processes by locking records during critical stages and unlocking them when done. Considerations: 3. Check for Matching Records (Upsert) When Creating Records Avoid duplicates by checking for existing records before creating new ones. One can never have too many de-dupe tools. Considerations: 4. Transform Your Data in Flows (Generally Available) Now generally available, perform calculations, data transformations, and more with the Transform element in Flow Builder. Considerations: Admin Enhancements 1. Field History Tracking Manage tracked objects and fields more efficiently with a centralized page in “Setup.” Considerations: 2. See What’s Enabled in Permission Sets and Permission Set Groups (Generally Available) Enhanced permission set viewing improves visibility and control over security configurations. Considerations: 3. Get a Summary of User’s Permissions and Access Quickly view user permissions, public groups, and queues from the user’s detail page. Help and Training Community: Salesforce is simplifying Permission Set management by phasing out Profiles. Data Cloud Vector Database Vector search capabilities allow the creation of searchable “vector embeddings” from unstructured data, enhancing AI applications’ understanding of semantic similarities and context. Considerations: Deep Dive Summer 24 Release The Salesforce Summer ’24 release is packed with features designed to enhance your Salesforce experience. From a sleek new interface to powerful automation tools, enhanced analytics, and expanded integration options, this release aims to elevate workflow efficiency and data protection. Jump into the exciting updates, and let’s make automation simpler and more user-friendly together! Like1 Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more The Salesforce Story In Marc Benioff’s own words How did salesforce.com grow from a start up in a rented apartment into the world’s Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more
The AI tech stack integrates various tools, libraries, and solutions to enable applications with generative AI capabilities, such as image and text generation. Key components of the AI stack include programming languages, model providers, large language model (LLM) frameworks, vector databases, operational databases, monitoring and evaluation tools, and deployment solutions. The infrastructure of the AI tech stack leverages both parametric knowledge from foundational models and non-parametric knowledge from information sources (PDFs, databases, search engines) to perform generative AI functions like text and image generation, text summarization, and LLM-powered chat interfaces. This insight explores each component of the AI tech stack, aiming to provide a comprehensive understanding of the current state of generative AI applications, including the tools and technologies used to develop them. For those with a technical background, this Tectonic insight includes familiar tools and introduces new players within the AI ecosystem. For those without a technical background, the article serves as a guide to understanding the AI stack, emphasizing its importance across business roles and departments. Content Overview: The AI layer is a vital part of the modern application tech stack, integrating across various layers to enhance software applications. This layer brings intelligence to the stack through capabilities such as data visualization, generative AI (image and text generation), predictive analysis (behavior and trends), and process automation and optimization. In the backend layer, responsible for orchestrating user requests, the AI layer introduces techniques like semantic routing, which allocates user requests based on the meaning and intent of the tasks, leveraging LLMs. As the AI layer becomes more effective, it also reduces the roles and responsibilities of the application and data layers, potentially blurring the boundaries between them. Nevertheless, the AI stack is becoming a cornerstone of software development. Components of the AI Stack in the Generative AI Landscape: The emphasis on model providers over specific models highlights the importance of a provider’s ability to update models, offer support, and engage with the community, making switching models as simple as changing a name in the code. Programming Languages Programming languages are crucial in shaping the AI stack, influencing the selection of other components and ultimately determining the development of AI applications. The choice of a programming language is especially important for modern AI applications, where security, latency, and maintainability are critical factors. Among the options for AI application development, Python, JavaScript, and TypeScript (a superset of JavaScript) are the primary choices. Python holds a significant market share among data scientists, machine learning, and AI engineers, largely due to its extensive library support, including TensorFlow, PyTorch, and Scikit-learn. Python’s readable syntax and flexibility, accommodating both simple scripts and full applications with an object-oriented approach, contribute to its popularity. According to the February 2024 update on the PYPL (PopularitY of Programming Language) Index, Python leads with a 28.11% share, reflecting a growth trend of +0.6% over the previous year. The PYPL Index tracks the popularity of programming languages based on the frequency of tutorial searches on Google, indicating broader usage and interest. JavaScript, with an 8.57% share in the PYPL Index, has traditionally dominated web application development and is now increasingly used in the AI domain. This is facilitated by libraries and frameworks that integrate AI functionalities into web environments, enabling web developers to leverage AI infrastructure directly. Notable AI frameworks like LlamaIndex and LangChain offer implementations in both Python and JavaScript/TypeScript. The GitHub 2023 State of Open Source report highlights JavaScript’s dominance among developers, with Python showing a 22.5% year-over-year increase in usage on the platform. This growth reflects Python’s versatility in applications ranging from web development to data-driven systems and machine learning. The programming language component of the AI stack is more established and predictable compared to other components, with Python and JavaScript/TypeScript solidifying their positions among software engineers, web developers, data scientists, and AI engineers. Model Providers Model providers are a pivotal component of the AI stack, offering access to powerful large language models (LLMs) and other AI models. These providers, ranging from small organizations to large corporations, make available various types of models, including embedding models, fine-tuned models, and foundational models, for integration into generative AI applications. The AI landscape features a wide array of models that support capabilities like predictive analytics, image generation, and text completion. These models can be broadly classified into closed-source and open-source categories. Closed-Source Models have privatized internal configurations, architectures, and algorithms, which are not shared with the model consumers. Key information about the training processes and data is withheld, limiting the extent to which users can modify the model’s behavior. Access to these models is typically provided through an API or a web interface. Examples of closed-source models include: Open-Source Models offer a different approach, with varying levels of openness regarding their internal architecture, training data, weights, and parameters. These models foster collaboration and transparency in the AI community. The levels of open-source models include: Open-source models democratize access to advanced AI technology, reducing the entry barriers for developers to experiment with niche use cases. Examples of open-source models include: Open-Source vs. Closed-Source Large Language Models (LLMs) AI engineers and machine learning practitioners often face the critical decision of whether to use open- or closed-source LLMs in their AI stacks. This choice significantly impacts the development process, scalability, ethical considerations, and the application’s utility and commercial flexibility. Key Considerations for Selecting LLMs and Their Providers Resource Availability The decision between open- and closed-source models often hinges on the availability of compute resources and team expertise. Closed-source model providers simplify the development, training, and management of LLMs, albeit at the cost of potentially using consumer data for training or relinquishing control of private data access. Utilizing closed-source models allows teams to focus on other aspects of the stack, such as user interface design and data integrity. In contrast, open-source models offer more control and privacy but require significant resources to fine-tune, maintain, and deploy. Project Requirements The scale and technical requirements of a project are crucial in deciding whether to use open-
In 2023, Algo Communications, a Canadian company, faced a significant challenge. With rapid growth on the horizon, the company struggled to train customer service representatives (CSRs) quickly enough to keep pace. To address this, Algo turned to an innovative solution: generative AI. They needed to Adopt a Large Language Model. Algo adopted a large language model (LLM) to accelerate the onboarding of new CSRs. However, to ensure CSRs could accurately and fluently respond to complex customer queries, Algo needed more than a generic, off-the-shelf LLM. These models, typically trained on public internet data, lack the specific business context required for accurate answers. This led Algo to use retrieval-augmented generation, or RAG. Many people have already used generative AI models like OpenAI’s ChatGPT or Google’s Gemini (formerly Bard) for tasks like writing emails or crafting social media posts. However, achieving the best results can be challenging without mastering the art of crafting precise prompts. An AI model is only as effective as the data it’s trained on. For optimal performance, it needs accurate, contextual information rather than generic data. Off-the-shelf LLMs often lack up-to-date, reliable access to your specific data and customer relationships. RAG addresses this by embedding the most current and relevant proprietary data directly into LLM prompts. RAG isn’t limited to structured data like spreadsheets or relational databases. It can retrieve all types of data, including unstructured data such as emails, PDFs, chat logs, and social media posts, enhancing the AI’s output quality. How RAG Works RAG enables companies to retrieve and utilize data from various internal sources for improved AI results. By using your own trusted data, RAG reduces or eliminates hallucinations and incorrect outputs, ensuring responses are relevant and accurate. This process involves a specialized database called a vector database, which stores data in a numerical format suitable for AI and retrieves it when prompted. “RAG can’t do its job without the vector database doing its job,” said Ryan Schellack, Director of AI Product Marketing at Salesforce. “The two go hand in hand. Supporting retrieval-augmented generation means supporting a vector store and a machine-learning search mechanism designed for that data.” RAG, combined with a vector database, significantly enhances LLM outputs. However, users still need to understand the basics of crafting clear prompts. Faster Responses to Complex Questions In December 2023, Algo Communications began testing RAG with a few CSRs using a small sample of about 10% of its product base. They incorporated vast amounts of unstructured data, including chat logs and two years of email history, into their vector database. After about two months, CSRs became comfortable with the tool, leading to a wider rollout. In just two months, Algo’s customer service team improved case resolution times by 67%, allowing them to handle new inquiries more efficiently. “Exploring RAG helped us understand we could integrate much more data,” said Ryan Zoehner, Vice President of Commercial Operations at Algo Communications. “It enabled us to provide detailed, technically savvy responses, enhancing customer confidence.” RAG now touches 60% of Algo’s products and continues to expand. The company is continually adding new chat logs and conversations to the database, further enriching the AI’s contextual understanding. This approach has halved onboarding time, supporting Algo’s rapid growth. “RAG is making us more efficient,” Zoehner said. “It enhances job satisfaction and speeds up onboarding. Unlike other LLM efforts, RAG lets us maintain our brand identity and company ethos.” RAG has also allowed Algo’s CSRs to focus more on personalizing customer interactions. “It allows our team to ensure responses resonate well,” Zoehner said. “This human touch aligns with our brand and ensures quality across all interactions.” Write Better Prompts – Adopt a Large Language Model If you want to learn how to craft effective generative AI prompts or use Salesforce’s Prompt Builder, check out Trailhead, Salesforce’s free online learning platform. Start learning Trail: Get Started with Prompts and Prompt Builder Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more The Salesforce Story In Marc Benioff’s own words How did salesforce.com grow from a start up in a rented apartment into the world’s Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more
Salesforce Unveils Data Cloud Vector Database with GenAI Integration Salesforce has officially launched its Data Cloud Vector Database, leveraging GenAI to rapidly process a company’s vast collection of PDFs, emails, transcripts, online reviews, and other unstructured data. Gen AI Unleased With Vector Database. Rahul Auradkar, Executive Vice President and General Manager of Salesforce Unified Data Services and Einstein Units, highlighted the efficiency gains in a one-on-one briefing with InformationWeek. Auradkar demonstrated the new capabilities through a live demo, showcasing the potential of the Data Cloud Vector Database. Enhanced Efficiency and Data Utilization The new Data Cloud integrates with the Einstein 1 platform, combining unstructured and structured data for rapid analysis by sales, marketing, and customer service teams. This integration significantly enhances the accuracy of Einstein Copilot, Salesforce’s enterprise conversational AI assistant. Gen AI Unleased With Vector Database Auradkar demonstrated how a customer service query could retrieve multiple relevant results within seconds. This process, which typically takes hours of manual effort, now leverages unstructured data, which makes up 90% of customer data, to deliver swift and accurate results. “This advancement allows our customers to harness the full potential of 90% of their enterprise data—unstructured data that has been underutilized or siloed—to drive use cases, AI, automation, and analytics experiences across both structured and unstructured data,” Auradkar explained. Comprehensive Data Management Using Salesforce’s Einstein 1 platform, Data Cloud enables users to ingest, store, unify, index, and perform semantic queries on unstructured data across all applications. This data encompasses diverse unstructured content from websites, social media platforms, and other sources, resulting in more accurate outcomes and insights. Auradkar emphasized, “This represents an order of magnitude improvement in productivity and customer satisfaction. For instance, a large shipping company with thousands of customer cases can now categorize and access necessary information far more efficiently.” Additional Announcements Salesforce also introduced several new AI and Data Cloud features: Auradkar noted that these innovations enhance Salesforce’s competitive edge by prioritizing flexibility and enabling customers to take control of their data. “We’ll continue on this journey,” Auradkar said. “Our future investments will focus on how this product evolves and scales. We’re building significant flexibility for our customers to use any model they choose, including any large language model.” For more insights and updates, visit Salesforce’s official announcements and stay tuned for further developments. Like1 Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more The Salesforce Story In Marc Benioff’s own words How did salesforce.com grow from a start up in a rented apartment into the world’s Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more
Salesforce World Tour Highlights: Data Cloud Vector Database and Hyperforce At the Salesforce World Tour on June 6, 2024, at the Excel Centre in east London, the focus was on advancements in the Data Cloud and Slack platforms. The event, sponsored by AWS, Cognizant, Deloitte, and PWC, showcased significant innovations, particularly for GenAI enthusiasts. Data Cloud Vector Database and Hyperforce. Vector Database in Data Cloud A key highlight was the announcement of the general availability of a Vector Database capability within the Data Cloud, integrated into the Einstein 1 Platform. This capability enhances Salesforce’s CRM platform, Customer 360, by combining structured and unstructured data about end-users. The Vector Database collects, ingests, and unifies data, allowing enterprises to deploy GenAI across all applications without needing to fine-tune an off-the-shelf large language model (LLM). Addressing Data Fragmentation Salesforce reports that approximately 80% of customer data is dispersed across various corporate departments in an unstructured format, trapped in PDFs, emails, chat conversations, and transcripts. The Vector Database unifies this fragmented data, creating a comprehensive profile of the customer journey. This unified approach not only improves customer engagement but also enhances organizational agility. By consolidating data from all corporate silos, companies can quickly and efficiently address issues such as product recalls and returns. Hyperforce: Enhancing Data Residency and Compliance During the keynote, Salesforce emphasized the importance of personalization in customer engagement and the benefits of deploying GenAI in customer-facing sectors. The event highlighted the need to overcome the fear and mistrust of GenAI and showcased how enterprises can enhance employee productivity through upskilling in GenAI technologies. One notable announcement was the general availability of Hyperforce, a solution designed to address data residency issues by integrating all Salesforce applications under the same compliance, security, privacy, and scalability standards. Built for the public cloud and composed of code rather than hardware, Hyperforce ensures safe delivery of applications worldwide, offering a common layer for deploying all application stacks and handling data compliance in a fragmented technology landscape. Salesforce AI Center The Salesforce AI Center was also introduced at the event. The first of its kind, located in the Blue Fin Building near Blackfriars, London, this center will support AI experts, Salesforce partners, and customers, facilitating training and upskilling programs. Set to open on June 18, 2024, the center aims to upskill 100,000 developers worldwide and is part of Salesforce’s $4 billion investment in the UK and Ireland. Industry Reactions and Future Prospects GlobalData senior analyst Beatriz Valle commented on Salesforce’s continued integration of GenAI across its portfolio, including platforms like Tableau, Einstein for analytics, and Slack for collaboration. According to Salesforce, the Data Cloud tool leverages all metadata in the Einstein 1 Platform, connecting unstructured and structured data, reducing the need for fine-tuning LLMs, and enhancing the accuracy of results delivered by Einstein Copilot, Salesforce’s conversational AI assistant. Vector databases, while not new, have gained prominence due to the GenAI revolution. They power the retrieval-augmented generation (RAG) technique, linking proprietary data with large language models like OpenAI’s GPT-4, enabling enterprises to generate more accurate results. Competitors such as Oracle, Amazon, Microsoft, and Google also offer vector databases, but Salesforce’s early investments in GenAI are proving fruitful with the launch of the Data Cloud Vector Database. Data Cloud Vector Database and Hyperforce Salesforce’s AI-powered integration solutions, highlighted during the World Tour, underscore the company’s commitment to advancing digital transformation. By leveraging GenAI and innovative tools like the Vector Database and Hyperforce, Salesforce is enabling enterprises to overcome the challenges of data fragmentation and compliance, paving the way for a more agile and competitive digital future. Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more The Salesforce Story In Marc Benioff’s own words How did salesforce.com grow from a start up in a rented apartment into the world’s Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more
Lessons Learned in the First Year of Developing AI Agents In the first year of working on AI agents, valuable insights emerged from direct collaboration with engineers and UX designers, as they iterated on the overall product experience. The objective was to create a platform for customers to use standard data analysis agents and build custom agents tailored to specific tasks and data structures relevant to their business. This platform integrates connectors to databases like Snowflake and BigQuery with built-in security, supports RAG over a metadata layer describing database contents, and facilitates data analysis through SQL, Python, and data visualization tools. Feedback on the effectiveness of these developments came from both internal evaluations and customer insights. Users from Fortune 500 companies utilize these agents daily to analyze their internal data. Key Insights on AI Agents Additional Insights Further insights on code and infrastructure include: These lessons underscore the importance of focusing on reasoning, iterative improvements to the agent-computer interface, understanding model limitations, and building robust supporting infrastructure to enhance AI agent performance and user satisfaction. Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more Top Ten Reasons Why Tectonic Loves the Cloud The Cloud is Good for Everyone – Why Tectonic loves the cloud You don’t need to worry about tracking licenses. Read more
As AI continues to advance, designers, builders, and creators are confronted with profound questions about the future of applications and how users will engage with digital experiences. AI Design Beyond the Chatbot. Generative AI has opened up vast possibilities, empowering people to utilize AI for tasks such as writing articles, generating marketing materials, building teaching assistants, and summarizing data. However, alongside its benefits, there are challenges. Sometimes, generative AI produces unexpected or biased responses, a phenomenon known as hallucination. In response, approaches like retrieval augmented generation (RAG) have emerged as effective solutions. RAG leverages a vector database, like SingleStore, to retrieve relevant information and provide users with contextually accurate responses. AI Design Beyond the Chatbot Looking ahead, the evolution of AI may lead to a future where users interact with a central LLM operating system, fostering more personalized and ephemeral experiences. Concepts like Mercury OS offer glimpses into this potential future. Moreover, we anticipate the rise of multimodal experiences, including voice and gesture interfaces, making technology more ubiquitous in our lives. Imran Chaudhri’s demonstration of a screen-less future, where humans interact with computers through natural language, exemplifies this trend. However, amidst these exciting prospects, the current state of AI integration in businesses varies. While some are exploring innovative ways to leverage AI, others may simply add AI chat interfaces without considering contextual integration. To harness AI effectively, it’s crucial to identify the right use cases and prioritize user value. AI should enhance experiences by reducing task time, simplifying tasks, or personalizing experiences. Providing contextual assistance is another key aspect. AI models can offer tailored suggestions and recommendations based on user context, enriching the user experience. Notion and Coda exemplify this by seamlessly integrating AI recommendations into user workflows. Furthermore, optimizing for creativity and control ensures users feel empowered in creation experiences. Tools like Adobe Firefly strike a balance between providing creative freedom and offering control over generated content. Building good prompts is essential for obtaining quality results from AI models. Educating users on how to construct effective prompts and managing expectations regarding AI limitations are critical considerations. Ultimately, as AI becomes more integrated into daily workflows, it’s vital to ensure seamless integration into user experiences. Responsible AI design requires ongoing dialogue and exploration to navigate this rapidly evolving landscape effectively. Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more Top Ten Reasons Why Tectonic Loves the Cloud The Cloud is Good for Everyone – Why Tectonic loves the cloud You don’t need to worry about tracking licenses. Read more
Salesforce Unveils Summer 2024 Release with Generative AI at the Forefront Salesforce has announced its Summer 2024 release, featuring generative AI (GenAI) as a key highlight. Set to be generally available on June 17, 2024, this release promises enhanced productivity and access to large language models (LLMs) on an open platform. Read on to see why we call Summer 24 the AI release. Key Features of the Summer 2024 Release 1. Bring Your Own LLM Expansion 2. Slack AI 3. Zero Copy Integration with Amazon Redshift 4. Vector Database 5. Data Cloud for Commerce 6. Digital Wallet Enhanced Security with Einstein Trust Layer The Einstein Trust Layer ensures enhanced protection for customer and company data, making the new features more secure. Upcoming Pre-Summer Releases In addition to the major features coming in June, Salesforce has already introduced several innovations: Unified Knowledge Solution Salesforce and Vonage Partnership Conclusion Salesforce’s Summer 2024 release is packed with generative AI enhancements, robust integrations, and new tools aimed at boosting productivity, security, and data insights. With features gradually rolling out and pre-summer innovations already available, Salesforce continues to lead in delivering cutting-edge AI solutions to its users. Like Related Posts Salesforce OEM AppExchange Expanding its reach beyond CRM, Salesforce.com has launched a new service called AppExchange OEM Edition, aimed at non-CRM service providers. Read more The Salesforce Story In Marc Benioff’s own words How did salesforce.com grow from a start up in a rented apartment into the world’s Read more Salesforce Jigsaw Salesforce.com, a prominent figure in cloud computing, has finalized a deal to acquire Jigsaw, a wiki-style business contact database, for Read more Health Cloud Brings Healthcare Transformation Following swiftly after last week’s successful launch of Financial Services Cloud, Salesforce has announced the second installment in its series Read more
