Latest news with #AIagents
Khaleej Times
a day ago
- Business
- Khaleej Times
Would you let an AI agent decide your crypto strategy?
As I've moved further into crypto, I've finally started to understand digi-currency humour. There's so much I don't grasp in this space that it's a noteworthy triumph. My favourite giggle of late comes from a meme featuring a therapist and patient, captioned: 'Is the AI agent in the room with us now?' People keep saying that artificial intelligence (AI) agents are one of the narratives of the upcoming — supposedly — crypto bull run. I'm still not sure I understand any of it, but I'm going to give it a try. So, what is an AI agent? In simple terms, AI agents are autonomous systems running on software that can perceive their environment, process information, make decisions, and take actions to achieve specific goals – often on behalf of a human, company, or industry. Unlike static, pre-programmed traditional software, AI agents are adaptive. Like humans — eek! — they evolve through learning and experience. Like clockwork Think Siri or Alexa, but on steroids. An AI agent doesn't just remind you to buy groceries — it actually orders them each week. It would go beyond suggesting investments to rebalancing your portfolio. I suppose one could manage that tiresome school WhatsApp chain for you too. In business, AI agents are already proving useful in everything from customer service and supply chain management to IT support and risk analysis. The next frontier? AI agents working with other AI agents. This emerging area — called multi-agent systems, agent-to-agent collaboration, or more dramatically, a swarm — involves coordinated networks of AI entities completing tasks together, often without human intervention. Yikes is right. And while most people are getting used to this new way of things through mainstream tools like Anthropic's Claude or Microsoft's Co-Pilot, radical innovation is happening in the decentralised world, too, away from traditional tech. And that's why I am starting to believe that when it comes investing in this new space, tokenised AI projects are where it's at. It's really interesting and, like everything in crypto, very speculative. Once you start exploring this space, you'll find everything from productivity agents to niche AI-driven trading bots. I've already experimented with (and immediately lost money) investing in transaction-based projects like SoloChain and The SWARM, from CryptoSlam. Projects like ai16z use agents to make real-time investment decisions. On Solana, Goatseus Maximus offers what is considered the first AI-managed venture capital platform. Virtuals Protocol provides a platform for developers to create and deploy AI agents across various Web3 applications. Bittensor offers an entire decentralised marketplace for AI models, built on its very own blockchain. One of the buzziest new players is a trading assistant called MIND of Pepe, which has been trained on more than 60,000 crypto tweets and 3,000 market insights, according to The Cryptonomist. Of course, aside from investing, we all still need to get over the biggest hurdle in adopting AI agents, especially for financial or high-stakes tasks. And that is trust. Will the agent do what it's supposed to? Safely and securely? What could go wrong with a swarm? I'm sure you are feeling a tonne of resistance. I am too. But the truth is, we are already using AI agents as they become integrated into almost everything we do. Over the past few weeks, I've been working with Underlord, a new AI tool from the editing platform Descript. It's been helping me produce the Hotflash inc podcast — removing filler words, laying music under ads, adding in sound effects as quickly as I can write 'add a whoosh sound here', and, when I get around to it, adding B-roll and graphics to video. I just have to type in what I need it to do, and it does it. I can refine and correct as needed. It's genuinely next-level. I'm already wondering how much more smoothly my platform would run, swarm-style, if Underlord could talk with a counterpart connected to my YouTube, Instagram, and Substack accounts. Market share One thing is for sure: there is not only a productivity leap here but also huge market potential. According to citing Bitget CEO Gracy Chen, tokens tied to AI agents are expected to grow from $15 billion to $60 billion (Dh55 billion to Dh220 billion) in market capitalisation this year alone. My take here is, feel the fear and do it anyway. Every generation has its technological reckoning. The printing press. The combustion engine. The telephone. The Internet. Smartphones. AI agents are just what's next. Like all that came before, they're arriving whether we're ready or not. I'm not here to tell you to invest. I'm suggesting that it's time to understand where this innovation is happening, who's building it — and how we might thoughtfully participate. If only there was an AI agent for that.
Geeky Gadgets
2 days ago
- Business
- Geeky Gadgets
AI Agents : What Experts Predict for the Future of Business
What if the next major innovation in your industry wasn't a product or a service, but an autonomous system working tirelessly behind the scenes? AI agents are no longer a futuristic concept—they're here, reshaping industries and redefining the way we work. From streamlining logistics to transforming healthcare diagnostics, these specialized systems are unlocking new levels of precision and efficiency. Over the course of 72 hours, leading experts gathered to discuss the fantastic potential of AI agents, revealing insights that could shape the future of business, technology, and even careers. The question isn't whether AI agents will impact your field—it's how soon and how profoundly. Tina Huang distills the most compelling takeaways from those discussions, offering a roadmap to understanding the rise of vertical AI agents, the skills needed to build them, and the trends driving their evolution. You'll discover how these systems are tailored to specific industries, why skills like prompt engineering and evaluations are critical for developers, and how no-code tools are lowering barriers to entry. Whether you're a professional exploring AI integration or an enthusiast curious about its potential, these insights will deepen your understanding of a rapidly advancing field. The future of AI isn't just about innovation—it's about adaptation, and this exploration will leave you questioning how prepared you are for what's coming. AI Agents: Key Insights How AI Agents Are Shaping Industries AI agents are becoming integral to modern business operations, driving efficiency and innovation across sectors. Industry leaders, including Google, have identified 2025 as a pivotal year for the widespread adoption of agent-driven products. These autonomous systems are automating tasks, streamlining workflows, and allowing businesses to achieve greater precision and scalability. A particularly notable development is the rise of vertical AI agents. These specialized systems are designed to address the unique needs of specific industries or tasks, much like vertical SaaS (Software as a Service). By focusing on niche markets, vertical AI agents are unlocking significant potential for businesses to optimize operations and reduce costs. Healthcare: Vertical AI agents assist in patient scheduling, analyze medical data, and support diagnostic processes, improving both efficiency and accuracy. Vertical AI agents assist in patient scheduling, analyze medical data, and support diagnostic processes, improving both efficiency and accuracy. Finance: These agents enhance operations by automating fraud detection, managing portfolios, and providing real-time insights. These agents enhance operations by automating fraud detection, managing portfolios, and providing real-time insights. Logistics: AI agents optimize supply chains, streamline inventory management, and improve delivery efficiency. By targeting specific domains, vertical AI agents empower businesses to deliver tailored solutions, reduce operational costs, and enhance overall performance. Essential Skills for Building AI Agents Creating effective AI agents requires a blend of technical expertise and strategic thinking. Two skills, in particular, stand out as critical for developers aiming to build robust and scalable systems: Prompt Engineering: This skill involves designing precise and adaptable prompts that guide AI agents to perform tasks effectively. Developers must define roles, tasks, inputs, outputs, and constraints, followed by iterative testing to refine the agent's behavior. For example, a customer service AI agent may require prompts that ensure empathetic yet efficient responses to user inquiries. This skill involves designing precise and adaptable prompts that guide AI agents to perform tasks effectively. Developers must define roles, tasks, inputs, outputs, and constraints, followed by iterative testing to refine the agent's behavior. For example, a customer service AI agent may require prompts that ensure empathetic yet efficient responses to user inquiries. Writing Evaluations (Evals): Evals are structured assessments that measure an AI agent's performance across dimensions such as task completion, reasoning quality, tool usage, robustness, and efficiency. By identifying weaknesses and making sure reliability, evals prepare AI agents for real-world deployment. Mastering these skills enables developers to create AI agents that are not only functional but also reliable and adaptable to diverse use cases. AI Agent Insights From 72 Hours of Expert Panels Watch this video on YouTube. Browse through more resources below from our in-depth content covering more areas on AI Agents. Career Opportunities in AI Agent Development The rapid growth of AI agents has given rise to new career opportunities, including the emerging role of 'Agent Engineer.' This multidisciplinary position combines expertise in prompt engineering, software development, product knowledge, and machine learning. As businesses increasingly adopt AI-driven solutions, the demand for skilled professionals in this field is surging. To excel as an agent engineer, focus on the following areas: Stay Updated: Keep track of advancements in AI frameworks, tools, and methodologies to remain competitive. Keep track of advancements in AI frameworks, tools, and methodologies to remain competitive. Gain Hands-On Experience: Work with automation technologies and experiment with building prototypes to refine your skills. Work with automation technologies and experiment with building prototypes to refine your skills. Develop Core Expertise: Hone your abilities in crafting effective prompts and conducting thorough evaluations to ensure the reliability of AI agents. By cultivating a diverse skill set and staying informed about industry trends, you can position yourself as a valuable contributor to the evolving AI landscape. Practical Steps for Building AI Agents Developing AI agents requires a structured and iterative approach. Whether you are building a simple chatbot or a complex vertical AI system, the following steps can guide you through the process: Identify Tasks: Analyze workflows to pinpoint repetitive or time-consuming tasks that can be automated effectively. Analyze workflows to pinpoint repetitive or time-consuming tasks that can be automated effectively. Decompose Processes: Break down tasks into smaller, manageable components and map out the necessary steps for automation. Break down tasks into smaller, manageable components and map out the necessary steps for automation. Create a Prototype: Develop an initial version of the AI agent, focusing on core functionalities and testing its performance. Develop an initial version of the AI agent, focusing on core functionalities and testing its performance. Iterate and Refine: Use feedback and testing results to improve the agent's accuracy, efficiency, and reliability. AI agents offer numerous advantages, including 24/7 availability, cost efficiency, and personalized user experiences. For instance, a retail AI agent could analyze customer preferences to recommend products, enhancing both customer satisfaction and sales. By following these steps, you can create AI solutions that deliver tangible value to businesses and users alike. Emerging Trends in AI Agent Development The future of AI agents is being shaped by several fantastic trends that are expanding their capabilities and accessibility: Voice Agents: Conversational interfaces are becoming more intuitive, allowing seamless interactions in areas such as customer service. These systems reduce wait times and enhance user satisfaction by providing quick and accurate responses. Conversational interfaces are becoming more intuitive, allowing seamless interactions in areas such as customer service. These systems reduce wait times and enhance user satisfaction by providing quick and accurate responses. No-Code Tools: The rise of no-code platforms is providing widespread access to AI development, allowing domain experts to prototype and deploy solutions without extensive programming knowledge. This trend is fostering innovation by allowing a broader range of professionals to contribute to AI advancements. While coding expertise remains valuable, these trends are lowering barriers to entry, making AI development more accessible and inclusive across industries. Insights from Industry Leaders Experts predict that vertical AI agents will surpass traditional SaaS solutions in market impact, fundamentally transforming how businesses operate. To remain competitive in this rapidly evolving field, continuous learning is essential. Focus on developing skills in areas such as prompt engineering, evaluations, and coding to stay ahead of the curve. By embracing these insights and trends, you can position yourself at the forefront of AI innovation, contributing to solutions that drive efficiency, scalability, and value across industries. Media Credit: Tina Huang Filed Under: AI, Top News Latest Geeky Gadgets Deals Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, Geeky Gadgets may earn an affiliate commission. Learn about our Disclosure Policy.
Geeky Gadgets
2 days ago
- Geeky Gadgets
The Hidden Flaws of Multi-Agent AI That You Need to Know
What if the very systems designed to transform problem-solving are quietly failing behind the scenes? Multi-agent AI, often hailed as the future of artificial intelligence, promises to tackle complex challenges by allowing multiple AI agents to collaborate like a well-oiled machine. But here's the catch: these systems are far less reliable than they appear. Research shows that failure rates in multi-agent AI can soar as high as 80%, with agents miscommunicating, misaligning, or even spiraling into infinite loops. These aren't just growing pains—they're systemic flaws that could derail the technology's potential. And yet, amidst the excitement surrounding AI's rapid advancements, these critical issues remain largely overlooked. Synsation provides more insights into the hidden vulnerabilities of multi-agent AI systems, from their troubling behavioral biases to their glaring safety gaps. You'll uncover why agents often mimic the worst traits of human group dynamics—like conformity and polarization—and how these tendencies can lead to disastrous outcomes in high-stakes environments. We'll also examine the paradox of autonomy: why these systems, despite their promise of reducing human workload, often demand constant oversight and intervention. The deeper you look, the clearer it becomes that multi-agent AI's greatest strength—its collaborative nature—might also be its Achilles' heel. So, what does this mean for the future of AI? Let's unravel the complexities. Challenges in Multi-Agent AI Understanding Multi-Agent AI Systems Multi-agent AI systems consist of multiple AI agents, each assigned a specific role or responsibility. These roles may include planning, executing tasks, or verifying outcomes. Frameworks such as Crew AAI, Autogen, Langraph, and NATN have been developed to assist collaboration among agents, aiming to harness collective intelligence for solving complex problems. In theory, this collaborative approach should outperform single-agent systems. However, in practice, multi-agent AI systems often fall short of expectations. They frequently exhibit high failure rates and unpredictable behaviors, raising concerns about their readiness for real-world applications. These shortcomings highlight the need for a deeper understanding of their limitations and the development of more robust solutions. High Failure Rates: A Systemic Problem Research indicates that multi-agent AI systems experience failure rates ranging from 60% to 80% across various frameworks. These failures are categorized under the MAST taxonomy, which identifies three primary areas of concern: Specification Issues (42%): Agents often fail due to hardcoded responses, infinite loops, or an inability to recognize task completion. These issues stem from poorly defined parameters and rigid programming. Agents often fail due to hardcoded responses, infinite loops, or an inability to recognize task completion. These issues stem from poorly defined parameters and rigid programming. Inter-Agent Misalignment (37%): Miscommunication, role misinterpretation, or conflicting actions between agents frequently lead to inefficiencies and errors. This misalignment undermines the intended collaboration. Miscommunication, role misinterpretation, or conflicting actions between agents frequently lead to inefficiencies and errors. This misalignment undermines the intended collaboration. Verification Failures (21%): Many systems lack robust mechanisms to verify task completion, allowing errors to go unnoticed and uncorrected. These systemic issues underscore the fragility of current multi-agent systems, particularly in dynamic or high-stakes environments. Without addressing these challenges, their reliability and practical utility will remain limited. Multi-Agent AI Failure Rates Watch this video on YouTube. Master Multi-Agent AI systems with the help of our in-depth articles and helpful guides. Behavioral Biases: When AI Mimics Human Flaws One of the more surprising findings in multi-agent AI research is the emergence of human-like group behaviors among AI agents. While this phenomenon is intriguing, it can negatively impact performance and decision-making: Conformity: Neutral agents often align with the majority or more 'intelligent' agents, even when the majority's conclusions are incorrect. This behavior reduces the system's ability to challenge flawed decisions. Neutral agents often align with the majority or more 'intelligent' agents, even when the majority's conclusions are incorrect. This behavior reduces the system's ability to challenge flawed decisions. Group Polarization: During debates or decision-making processes, agents in the majority may escalate their positions to extremes, amplifying biases and reducing objectivity. These tendencies raise significant concerns about the ability of multi-agent systems to make balanced, unbiased decisions. In scenarios requiring impartiality or nuanced judgment, such behaviors could lead to suboptimal or even harmful outcomes. Safety Mechanisms: Falling Short Safety remains a critical concern for multi-agent AI systems. Testing across 16 popular agents on 2,000 tasks revealed that none achieved a safety score above 60%. Common safety failures include: Overconfidence: Agents often exhibit excessive confidence in their decisions, leading to errors that could have been avoided with more cautious approaches. Agents often exhibit excessive confidence in their decisions, leading to errors that could have been avoided with more cautious approaches. Rule-breaking: Some agents disregard safety protocols or ethical guidelines, compromising the integrity of the system. Some agents disregard safety protocols or ethical guidelines, compromising the integrity of the system. Inability to recover: When errors occur, many systems fail to recover effectively, resulting in cascading failures that exacerbate the problem. These limitations not only restrict the applicability of multi-agent systems but also pose significant risks in high-stakes environments such as healthcare, finance, or autonomous vehicles. In these contexts, errors can have severe and far-reaching consequences. The Burden of Human Oversight Despite their autonomous design, multi-agent AI systems often require substantial human intervention. Tasks such as monitoring, correcting errors, and guiding agents can be labor-intensive, undermining the efficiency gains these systems are intended to provide. This reliance on human oversight highlights the need for smarter frameworks that enable better delegation, error recovery, and trust-building. Without addressing these shortcomings, multi-agent systems risk becoming a burden rather than a solution. The promise of reducing human workload will remain unfulfilled unless these systems can operate with greater autonomy and reliability. Charting a Path Forward To overcome the challenges of multi-agent AI systems, a multifaceted approach is essential. Key areas for improvement include: Defining clearer roles: Assigning specific, well-defined roles to agents can reduce misalignment and improve collaboration. Assigning specific, well-defined roles to agents can reduce misalignment and improve collaboration. Developing robust verification processes: Implementing mechanisms to verify task completion can minimize errors and enhance reliability. Implementing mechanisms to verify task completion can minimize errors and enhance reliability. Enhancing collaboration frameworks: Improving the underlying frameworks can make agent interactions more efficient and effective. Improving the underlying frameworks can make agent interactions more efficient and effective. Mitigating behavioral biases: Addressing conformity and group polarization can ensure more balanced and objective decision-making. Addressing conformity and group polarization can ensure more balanced and objective decision-making. Strengthening safety mechanisms: Introducing stricter safety protocols and recovery strategies can prevent critical failures and reduce risks. While the current limitations of multi-agent AI systems are significant, they also present opportunities for innovation. Continued research and development could transform these systems into reliable tools capable of solving complex problems across industries. However, until these challenges are addressed, their real-world applications will remain constrained. By focusing on these areas, researchers and developers can unlock the full potential of multi-agent AI systems, paving the way for their successful deployment in diverse and demanding environments. Media Credit: Synsation Filed Under: AI, Top News Latest Geeky Gadgets Deals Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, Geeky Gadgets may earn an affiliate commission. Learn about our Disclosure Policy.
Bloomberg
2 days ago
- Business
- Bloomberg
Marc Benioff's Race to One Billion AI Agents
Emily Chang meets Salesforce Co-Founder & CEO Marc Benioff to discuss the rise of AI agents and what it's like to run a very memeable global tech company. (Source: Bloomberg)
Zawya
3 days ago
- Business
- Zawya
Salesforce launches Agentforce 3 to solve the biggest blockers to scaling AI agents: Visibility and control
New Agentforce Command Center provides a complete observability solution for optimizing AI agents — enabling leaders to manage, track, and scale how AI agent activity enhances human productivity Agentforce 3 enables seamless agent interoperability with built-in support for open standards like Model Context Protocol (MCP). Through the expanded AgentExchange, customers will be able to access plug-and-play services from over 30 partners including AWS, Box, Cisco, Google Cloud, IBM, Notion, PayPal, Stripe, Teradata, WRITER, and more Agentforce 3 delivers an updated Atlas architecture — from reasoning, to performance, to trust — for unmatched enterprise readiness, including lower latency, greater accuracy, enhanced resiliency, and support for natively hosted LLMs such as Anthropic UAE - Salesforce (NYSE: CRM), the world's #1 AI CRM, has announced Agentforce 3: a major upgrade to its digital labor platform that gives companies the visibility and control to scale AI agents without compromise. As enterprise adoption accelerates, the real blocker has become clear: teams can't see what agents are doing — or evolve them fast enough. Agentforce 3 changes that. Built on learnings from thousands of Agentforce deployments since its initial launch in October 2024, Agentforce has helped customers deliver undeniable value. Agentforce 3 equips leaders to monitor, improve, and scale their AI workforce with confidence. With a new Command Center for complete observability, built-in support for Model Context Protocol (MCP) for plug-and-play interoperability, and over 100 new prebuilt industry actions to speed time to value, Agentforce 3 helps companies scale what works, fix what doesn't, and unlock the full potential of agentic AI — with clarity, control, and speed. Why it Matters: AI agent adoption is surging. According to a soon to be released Slack Workflow Index, AI agent usage is up 233% in six months, and over that same period, 8,000 customers have signed up to deploy Agentforce. But until now, agent platforms have lacked the tooling, governance, and observability needed to scale enterprise-wide. Agentforce 3 closes this gap — delivering the complete visibility, secure tool integration, and enterprise-grade controls organizations need to make agent velocity their competitive advantage. 'With Agentforce, we've unified agents, data, apps, and metadata to create a digital labor platform, helping thousands of companies realize the promise of agentic AI today,' said Adam Evans, EVP & GM of Salesforce AI. 'Over the past several months we've listened deeply to our customers and continued our rapid pace of technology innovation. The result is Agentforce 3, a major leap forward for our platform that brings greater intelligence, higher performance, and more trust and accountability to every Agentforce deployment. Agentforce 3 will redefine how humans and AI agents work together — driving breakthrough levels of productivity, efficiency, and business transformation.' Agentforce Command Center — Monitor, Measure, and Optimize Your AI Agents: As AI agents take on routine tasks and begin collaborating more closely with human teammates, teams need a new observability layer built for the era of digital labor. Agentforce Command Center is that layer: a complete observability solution that gives leaders a unified pane of glass to monitor agent health, measure performance, and optimize outcomes. Built into Agentforce Studio, it completes the agent lifecycle with powerful tools to understand and refine agents at scale. Uncover patterns across interactions to optimize your agents: Command Center empowers teams to analyze every AI agent interaction, drill into specific moments, understand trends in usage, and see AI-powered recommendations for tagged conversation types to continuously improve your Agentforce. Track agent health and intervene in real time: Get live, detailed analytics for latency, escalation frequency, and error rates, plus real-time alerts when the unexpected happens, so teams can act fast and keep agents running smoothly. Understand what's working, and where to improve: Command Center offers detailed dashboards that track agent adoption, feedback, success rates, cost, and topic performance — so teams can see what's gaining traction and where to improve. See what your agents are doing — in the tools your teams already use: Agentforce captures all agent activity in a native, extensible session-tracing data model in Data Cloud — powering analytics, monitoring, and real-time alerting. Built on the OpenTelemetry standard, these agent signals integrate seamlessly with tools your teams already use, including Datadog, Splunk, Wayfound, and other monitoring partners for end-to-end visibility across your existing stack. Deliver a configurable Command Center for every team: Monitor AI agents alongside human teammates — right in the flow of work. Starting with Service Cloud, agent activity will surface in real-time wallboards so contact center supervisors can track performance and escalate fast. And over time, every department will have a Command Center purpose-built for optimizing their agents. Build and test agents fast with AI-assisted development tools: In Agentforce Studio, use natural language to generate topics, instructions, and test cases. Testing Center simulates behavior at scale with state injection and AI-driven evals — so you can pressure-test your agents before going live. Enabling Secure Enterprise Connectivity with MCP and A2A Support: AI agents can't drive impact if they can't take action using the tools your business relies on. As open standards like Model Context Protocol (MCP) gain traction, they bring new opportunities for interoperability, but also challenges around governance, identity, and control. Agentforce 3 solves this by pairing open connectivity with enterprise-grade trust — giving agents native access to the tools they need, without compromising on control. MCP support built natively into Agentforce: Agentforce will include a native MCP client, enabling Agentforce agents to connect to any MCP-compliant server — no custom code required. Like a 'USB-C for AI,' this enables access to enterprise tools, prompts, and resources — governed by your existing security policies. Turn APIs into MCP servers instantly with MuleSoft: Leveraging new MCP connectors, MuleSoft converts any API and integration into an agent-ready asset, complete with security policies, activity tracing, and traffic controls — empowering teams to orchestrate and govern multi-agent workflows. Easily host and manage custom MCP servers with Heroku: Heroku Managed Inference and AppLink make it fast and easy to deploy, register, maintain, and connect your custom MCP servers. With Heroku's secure infrastructure and DevOps automation, developers can bring trusted custom actions to Agentforce with less friction. Expanding the AgentExchange — a Leading AI Agent Ecosystem — to Make Agent-First Tools Easy to Discover and Use: As organizations embed AI agents across their teams, AgentExchange already makes it easy to equip them with the trusted tools they need to drive value fast with partner-built agent actions and templates. Now, customers will be able to discover MCP servers from 30+ partners that provide trusted access to new third-party tools and resources and seamlessly connect them to Agentforce through a secure AI agent gateway. Launch MCP partners include AWS, Box, Cisco, Google Cloud, IBM, Notion, PayPal, Stripe, Teradata, WRITER and more. Customers will be able to leverage MCP Servers through AgentExchange from leading partners, including: AWS: Agentforce will enable agents to unlock critical insights from vast repositories of unstructured data stored on AWS, leveraging a suite of MCP servers such as the Amazon Bedrock Data Automation MCP Server and Amazon Nova Canvas MCP Server. This will allow Agentforce to intelligently process and summarize documents, extract key information from images, transcribe and analyze audio recordings, identify crucial moments in video files, query enterprise knowledge bases, perform semantic code repository searches, and interact with databases through natural language interfaces via Aurora PostgreSQL/MySQL MCP Servers — all enhancing customer service, compliance checks, and operational reporting. Box: Agentforce can help retrieve the latest signed NDA from a Box MCP service, including a summary of key terms. Agentforce can then post a Slack update to the relevant teams that includes the AI summarization, the contract in PDF form, and recommended next steps. Google Cloud: Customers can unlock new Agentforce use cases by connecting to key Google and Google Cloud services via MCP. For example, developers will be able to seamlessly connect Agentforce to the Google Maps Platform and generative AI models like Veo, Chirp, Imagen, and Lyria, while also enabling agentic interactions with Google Cloud Databases. This can enable Agenforce actions that span location-based services, creative AI, and transactional data. PayPal: Customers can now power a full range of agentic commerce capabilities by connecting Agentforce to PayPal's MCP server. They can list products, place orders, process payments, dispute claims, track shipping, manage subscriptions, issue refunds, and more with PayPal, making integrations simple and easy. Stripe: Agentforce enables AI agents to directly handle a wide range of Stripe functionalities via its MCP server, including real-time product lookups, quick payment link generation, subscription setup, and faster dispute management. WRITER: Through a secure connection to WRITER's MCP server, Agentforce can call WRITER's enterprise-grade AI agents to automate content generation, knowledge retrieval, and compliance checks all within Salesforce workflows. Enhancing the Agentforce Architecture for Unmatched Enterprise Readiness: Underpinning every new capability in Agentforce 3 is an enhanced Atlas architecture, providing an enterprise-ready foundation through lower latency, greater accuracy, global availability, and additional options for control through new LLMs hosted on Salesforce infrastructure. Expanded LLM choice with hosted Anthropic: Agentforce can now use Anthropic's Claude Sonnet model hosted via Amazon Bedrock within the Salesforce trust boundary to meet the needs of customers in high compliance industries. As part of this expanded relationship, Anthropic will work with Salesforce to empower customers in regulated industries to scale Agentforce adoption with Claude. Later this year, Salesforce will also allow customers to use Google's Gemini in Agentforce, solidifying Agentforce's position as the leader in trusted, flexible AI agents. Faster performance and response streaming: Experience a more responsive Agentforce with 50% lower latency since January 2025. Response streaming is also generally available in this release, so users can see answers appear in real time. Greater accuracy through web search, inline citations: Agentforce 3 features trusted, expanded grounding through the addition of web search as a data source, allowing agents to go beyond internal data to answer requests as well as inline citations that provide references to the grounding sources used in responses. More languages, more geographies: Agentforce 3 features an expanded global footprint, deploying to Canada, the U.K., India, Japan, and Brazil to serve AI agent traffic within those regions. This release also adds GA support for six new languages, including French, Italian, German, Spanish, Japanese, and Portuguese — with more than 30 additional languages rolling out in the coming months. Enhanced resiliency with automatic model failover: To ensure agents are always on, Agentforce now enables automatic, latency-based failover — dynamically shifting traffic between model providers in case of performance degradation or outages. Bringing Agentforce to Public Sector customers with FedRAMP High Authorization: Agentforce is now authorized and generally available in Government Cloud Plus, which enables public sector customers to bring Agentforce to their missions with the highest standards of trust, security, and compliance. Unmatched Industry Readiness Out-of-the-Box Delivers the Fastest Path to Value from AI Agents: Agentforce customers in every industry can deliver value from AI agents fast through more than 200 pre-built industry actions — half of them new this summer — from patient scheduling, to advertising proposal generation, to vehicle servicing, and more. Agentforce 3 also introduces simplified, flexible pricing with new Agentforce for Sales, Service, and Industry Cloud SKUs with per-user pricing and unlimited usage of actions for employee-facing agents to help teams get started fast and scale boldly. Deploy and Scale Agentforce with Confidence Through Trusted Experts: The Salesforce partner ecosystem, including Accenture, Deloitte Digital, NeuraFlash, PwC, and more, has aided customers across thousands of Agentforce deployments, providing 272,000 certified specialists to deliver strategic guidance and facilitate the successful activation of AI agents that help drive business impact while maintaining governance. Available Today: Agentforce 3 Agentforce adoption analytics Testing Center enhancements 100+ new, pre-built industry actions New Agentforce add-on SKUs with unlimited employee action usage Heroku managed MCP server hosting Increased speed and response streaming Web Search for Agentforce Data Libraries Agentforce for Government Cloud Plus with FedRAMP High authorization Expanded global availability (Canada, U.K., India, Japan, Brazil) and language support (French, Italian, German, Spanish, Japanese, Portuguese) In Pilot or Beta Today: Anthropic Claude models hosted within the Salesforce trust boundary — generally available in July MuleSoft MCP and A2A support — generally available in July Heroku AppLink — generally available in July Session Tracing Data Model — generally available in August Agent health monitoring — generally available in August To Be Released Soon: Agentforce native MCP support – July Agentforce Command Center and Agentforce Studio app – August About Salesforce Salesforce empowers companies of every size and industry to connect with their customers in a whole new way through the power of AI + data + CRM.
