News from the AI & ML world
DeeperML - #businesses
|
@cloud.google.com
//
References:
AI & Machine Learning
, www.tomsguide.com
Google Cloud is enhancing its text-to-SQL capabilities using the Gemini AI model. This technology aims to improve the speed and accuracy of data access for organizations that rely on data-driven insights for decision-making. SQL, a core component of data access, is being revolutionized by Gemini's ability to generate SQL directly from natural language, also known as text-to-SQL. This advancement promises to boost productivity for developers and analysts while also empowering non-technical users to interact with data more easily.
Gemini's text-to-SQL capabilities are already integrated into several Google Cloud products, including BigQuery Studio, Cloud SQL Studio (supporting Postgres, MySQL, and SQL Server), AlloyDB Studio, and Cloud Spanner Studio. Users can find text-to-SQL features within the SQL Editor, SQL Generation tool, and the "Help me code" functionality. Additionally, AlloyDB AI offers a direct natural language interface to the database, currently available as a public preview. These integrations leverage Gemini models accessible through Vertex AI, providing a foundation for advanced text-to-SQL functionalities. Current state-of-the-art LLMs like Gemini 2.5 possess reasoning skills that enable them to translate intricate natural language queries into functional SQL code, complete with joins, filters, and aggregations. However, challenges arise when applying this technology to real-world databases and user questions. To address these challenges, Google Cloud is developing methods to provide business-specific context, understand user intent, manage SQL dialect differences, and complement LLMs with additional techniques to offer accurate and certified answers. These methods include context building, table retrieval, LLM-as-a-judge techniques, and LLM prompting and post-processing, which will be explored further in future blog posts. Recommended read:
References :
@Salesforce
//
Salesforce is aggressively expanding its Agentforce platform to capitalize on the growing demand for AI-powered digital labor. A key component of this strategy is the acquisition of UK-based AI automation startup, Convergence.ai. This acquisition will bring Convergence's team and technology, including their expertise in AI agent design, autonomous task execution, and adaptive systems, into the Agentforce platform. The move is intended to accelerate the development of sophisticated AI agents capable of handling complex digital workflows within enterprise environments.
Convergence.ai, founded in 2024 by machine learning scientists Marvin Purtorab and Andy Toulis, has developed innovative AI agents, including "Proxy," designed to learn, evolve, and collaborate with humans. Their technology enables AI agents to navigate and execute tasks within dynamic digital systems, adapting to challenges such as pop-ups, system errors, and user interface changes. According to Salesforce, this acquisition will enable Agentforce to deliver AI that truly transforms how work gets done, allowing for AI assistants that can intelligently manage intricate tasks with human-like ingenuity. To further drive adoption of Agentforce, Salesforce has introduced new flexible pricing. The new ‘Flex Credits’ pricing model is designed to enable businesses to scale AI-powered digital labor to every employee, department, and process. This initiative aims to meet the rapidly accelerating demand for digital labor and positions Salesforce as a leader in the evolving landscape of AI-driven automation and customer interaction. Salesforce executives envision a future where Agentforce empowers customers with AI agents that not only follow instructions but also perceive, reason, and adapt to the complexities of modern digital workflows. Recommended read:
References :
Tom Krazit@Runtime
//
References:
Runtime
ServiceNow is spearheading the integration of AI agents within enterprise IT, outlining a plan to automate various IT functions and processes. CEO Bill McDermott emphasized that the "control tower" for autonomous agentic AI within corporations will originate from the IT department. This vision positions IT, and specifically the chief digital information officer, as central to managing and securing AI deployments across the enterprise. ServiceNow introduced AI Control Tower, a centralized platform to govern, manage, and secure AI agents, models, and workflows, irrespective of whether they are from ServiceNow or third-party vendors.
The company has also launched numerous new AI agents designed to integrate with its core IT management software, including IT service management, operations management, and asset management. These agents aim to alleviate the burden of routine tasks on IT departments, such as resolving support tickets and streamlining incident response. According to Pablo Stern, executive vice president and general manager of Technology Workflow Products, these AI agents will handle "the menial work," ultimately reducing the number of human interventions required for various IT processes. While ServiceNow is a prominent advocate for agentic AI, other companies are also exploring its potential. The rise of AI agents is also impacting the Go-to-Market (GTM) strategies of many companies. Scale Venture Partners has highlighted that AI agents are goal-oriented systems capable of observing, deciding, and acting within a defined environment. These agents are redefining roles in prospecting, forecasting, and customer success by making intelligent decisions and scaling proven strategies. Furthermore, agentic AI software development is also transforming low-code and no-code platforms, potentially impacting the future roles of software engineers as AI-assisted coding becomes more prevalent. Recommended read:
References :
@the-decoder.com
//
OpenAI is making strides in AI customization and application development with the release of Reinforcement Fine-Tuning (RFT) on its o4-mini reasoning model and the appointment of Fidji Simo as the CEO of Applications. The RFT release allows organizations to tailor their versions of the o4-mini model to specific tasks using custom objectives and reward functions, marking a significant advancement in model optimization. This approach utilizes reinforcement learning principles, where developers provide a task-specific grader that evaluates and scores model outputs based on custom criteria, enabling the model to optimize against a reward signal and align with desired behaviors.
Reinforcement Fine-Tuning is particularly valuable for complex or subjective tasks where ground truth is difficult to define. By using RFT on o4-mini, a compact reasoning model optimized for text and image inputs, developers can fine-tune for high-stakes, domain-specific reasoning tasks while maintaining computational efficiency. Early adopters have demonstrated the practical potential of RFT. This capability allows developers to tweak the model to better fit their needs using OpenAI's platform dashboard, deploy it through OpenAI's API, and connect it to internal systems. In a move to scale its AI products, OpenAI has appointed Fidji Simo, formerly CEO of Instacart, as the CEO of Applications. Simo will oversee the scaling of AI products, leveraging her extensive experience in consumer tech to drive revenue generation from OpenAI's research and development efforts. Previously serving on OpenAI's board of directors, Simo's background in leading development at Facebook suggests a focus on end-users rather than businesses, potentially paving the way for new subscription services and products aimed at a broader audience. OpenAI is also rolling out a new GitHub connector for ChatGPT's deep research agent, allowing users with Plus, Pro, or Team subscriptions to connect their repositories and ask questions about their code. Recommended read:
References :
@techradar.com
//
References:
AWS News Blog
, Data Phoenix
,
AI adoption is accelerating rapidly, with Amazon reporting that a UK business is adopting AI every 60 seconds. This surge in adoption is highlighted in a recent AWS report, which indicates a 33% increase in the past year, bringing the total of UK businesses utilizing AI to 52%. Startups appear to be leading the charge, with 59% adoption rate, and are also more likely to have comprehensive AI strategies in place compared to larger enterprises, 31% versus 15% respectively. Benefit realization is also on the rise, with 92% of AI-adopting businesses reporting an increase in revenue, a substantial jump from 64% in 2024.
Amazon is also introducing new tools to assist developers in building and scaling AI solutions. Amazon Q Developer is now available in preview on GitHub, enabling developers to assign tasks to an AI agent directly within GitHub issues. This agent can develop features, conduct code reviews, enhance security, and migrate Java code. The tool aims to accelerate code generation and streamline the development process, allowing developers to quickly implement AI-driven functionalities within their projects. Installation is simple, and developers can begin using the application without connecting to an AWS account. Adding to its suite of AI offerings, Amazon has launched Nova Premier, its most capable foundation model, now generally available on Amazon Bedrock. Nova Premier is designed to handle complex workflows requiring multiple tools and data sources. It boasts a one-million token context window, enabling it to process lengthy documents and large codebases. One notable feature of Nova Premier is its model distillation capabilities, allowing users to transfer its advanced features to smaller, faster models for production deployment. Amazon is investing in AI training, with a UK initiative to train 100,000 people in AI skills by the end of the decade, collaborating with universities such as Exeter and Manchester. Recommended read:
References :
@techradar.com
//
AI agents are rapidly transforming business operations, marketing strategies, and customer support systems. Recent findings indicate a potential 25% decrease in browser searches by the end of the year as AI algorithms increasingly handle shopping tasks. Small and mid-sized businesses (SMBs) are already experiencing the benefits of AI-powered voice agents, with a Vida survey reporting increased revenue, improved customer engagement, and enhanced industry leadership positioning among early adopters. However, the adoption of AI voice agents remains uneven, primarily due to concerns about customer preferences for human interaction and potential implementation complexities.
Sendbird Inc. has launched an "omnipresent" AI agent designed to proactively address customer service issues across multiple channels, including web, mobile, email, SMS, WhatsApp, and voice. This agent maintains conversation history, allowing seamless transitions between channels and resolving issues before customers even contact support. By anticipating problems like delivery delays, the AI agent can autonomously handle rescheduling or cancellations, escalating to human support only when necessary. According to IDC, this shift towards omnipresent and proactive support enhances customer experiences, reduces friction, and improves loyalty. As AI agents become more integrated into business processes, security concerns are also escalating. Marta Dern Simon, senior product marketing manager at Oasis Security, emphasizes the emerging challenges in managing and securing these agents, particularly regarding permissions and potential cyberattacks. With AI agents expanding the attack surface, organizations must carefully authenticate, manage, and monitor them, assigning appropriate privilege levels and limiting access to align with specific tasks. RSAC 2025 highlighted the growing demand for Chief Information Security Officers (CISOs) as businesses prioritize protecting AI/ML models and data pipelines. Recommended read:
References :
@blogs.microsoft.com
//
Microsoft is aggressively promoting agentic AI as a key driver for business transformation, emphasizing its potential to unlock greater value for customers. Agentic AI, with its autonomous capabilities, combined with copilots and human ambition, is believed to offer real AI differentiation. Microsoft's vision involves embedding AI directly into business processes, enabling organizations to achieve more by leveraging the power of intelligent agents acting on their behalf. Judson Althoff, Executive Vice President and Chief Commercial Officer at Microsoft, highlighted the rapid growth of agentic AI and its crucial role in accelerating AI transformation for businesses. The recent introduction of Microsoft 365 Copilot further underscores this commitment to making AI accessible and beneficial to all.
Recent updates include the release of a comprehensive guide to failure modes in agentic AI systems by Microsoft's AI Red Team (AIRT). This guide aims to help practitioners design and maintain resilient agentic systems by addressing potential security and safety challenges. The guide categorizes failure modes across two dimensions: security and safety, each comprising both novel and existing types. Novel security failures include agent compromise, agent injection, and agent flow manipulation, while novel safety failures cover intra-agent Responsible AI concerns and biases in resource allocation. By providing a structured analysis of these failure modes, Microsoft seeks to foster the responsible development and deployment of agentic AI technologies. In addition to agentic AI, Microsoft is also urging the U.S. and its allies to double down on quantum computing investments to maintain technological leadership amid growing global competition. Microsoft President Brad Smith warned that the U.S. risks falling behind China in the quantum race unless it strengthens investment, workforce development, and supply chain security. Smith advocates for expanding federal research funding, boosting quantum talent development, and shoring up domestic quantum manufacturing. He emphasized that quantum computing is transitioning from theory to practice, with transformative potential for science, medicine, energy, and national security. Recommended read:
References :
Ryan Daws@AI News
//
Anthropic has unveiled groundbreaking insights into the 'AI biology' of their advanced language model, Claude. Through innovative methods, researchers have been able to peer into the complex inner workings of the AI, demystifying how it processes information and learns strategies. This research provides a detailed look at how Claude "thinks," revealing sophisticated behaviors previously unseen, and showing these models are more sophisticated than previously understood.
These new methods allowed scientists to discover that Claude plans ahead when writing poetry and sometimes lies, showing the AI is more complex than previously thought. The new interpretability techniques, which the company dubs “circuit tracing” and “attribution graphs,” allow researchers to map out the specific pathways of neuron-like features that activate when models perform tasks. This approach borrows concepts from neuroscience, viewing AI models as analogous to biological systems. This research, published in two papers, marks a significant advancement in AI interpretability, drawing inspiration from neuroscience techniques used to study biological brains. Joshua Batson, a researcher at Anthropic, highlighted the importance of understanding how these AI systems develop their capabilities, emphasizing that these techniques allow them to learn many things they “wouldn’t have guessed going in.” The findings have implications for ensuring the reliability, safety, and trustworthiness of increasingly powerful AI technologies. Recommended read:
References :
Michael Nuñez@AI News | VentureBeat
//
References:
AiThority
, AI News | VentureBeat
,
AI security startup Hakimo has secured $10.5 million in Series A funding to expand its autonomous security monitoring platform. The funding round was led by Vertex Ventures and Zigg Capital, with participation from RXR Arden Digital Ventures, Defy.vc, and Gokul Rajaram. This brings the company’s total funding to $20.5 million. Hakimo's platform addresses the challenges of rising crime rates, understaffed security teams, and overwhelming false alarms in traditional security systems.
The company’s flagship product, AI Operator, monitors existing security systems, detects threats in real-time, and executes response protocols with minimal human intervention. Hakimo's AI Operator utilizes computer vision and generative AI to detect any anomaly or threat that can be described in words. Companies using Hakimo can save approximately $125,000 per year compared to using traditional security guards. Recommended read:
References :
Ken Yeung@Ken Yeung
//
Microsoft is enhancing its Copilot Studio platform with new 'deep reasoning' capabilities, allowing AI agents to solve complex problems more effectively. This upgrade also includes 'agent flows' which blend AI's flexibility with structured business automation. The new Researcher and Analyst agents for Microsoft 365 Copilot represent a significant step forward in AI agent evolution, enabling them to handle sophisticated tasks requiring detailed analysis and methodical thinking.
Microsoft's Security Copilot service is also getting a boost with a set of AI agents designed to automate repetitive tasks, freeing up security professionals to focus on more critical threats. These AI agents are designed to assist with critical tasks such as phishing, data security, and identity management. These agents showcase the breadth of what can be created when combining enterprise business data, access to advanced reasoning models, and structured workflows. Recommended read:
References :
george.fitzmaurice@futurenet.com (George@Latest from ITPro
//
References:
www.itpro.com
, Databricks
The AI agent landscape is rapidly evolving, with major tech companies pushing 'do-it-yourself' agent platforms to drive AI adoption. Firms like Oracle, OpenAI, AWS, Salesforce, and Workday are releasing platforms that allow users to build custom agents, rather than offering pre-built solutions. This emphasis on customization stems from the understanding that AI agent use cases are often less deterministic and require tailoring to specific business contexts. Gartner analyst Pieter J. den Hamer highlights the need for customization, noting that end-users gain the most from agentic tools when they have full control over their functionality.
Dataiku offers a platform to build AI agents that optimize workflows, enhance productivity, and automate complex processes. They allow users to add tools that extend agent capabilities, allowing integration with external systems. China's Manus AI is emerging as a potential leader, moving beyond chatbots to autonomous agents capable of executing real-world tasks with minimal human oversight. Other offerings include Databricks Apps, which can be combined with React and Mosaic AI Agent Framework, to create enterprise chat solutions. Recommended read:
References :
Ellie Ramirez-Camara@Data Phoenix
//
Nvidia is making significant strides in the realm of AI agents, highlighted at this year's GTC 2025 conference. CEO Jensen Huang emphasized the transformative impact of agentic AI and reasoning models, predicting that these technologies will revolutionize industries and automate processes. To support this shift, Nvidia unveiled the Blackwell Ultra platform, designed to handle the demanding requirements of AI reasoning, agentic AI, and physical AI applications. The platform, which includes the GB300 NVL72 rack-scale solution and the HGX B300 NVL16 system, offers substantial performance improvements over previous generations, with the GB300 NVL72 delivering 1.5x more AI performance.
In addition to hardware advancements, Nvidia launched NVIDIA Dynamo, an open-source inference framework, to optimize reasoning AI services across thousands of GPUs. This framework is designed to maximize token revenue generation for AI factories deploying reasoning AI models by orchestrating and accelerating inference communication across GPU clusters. Major cloud providers and server manufacturers are expected to offer Blackwell Ultra-based products starting in the second half of 2025. These developments position Nvidia as a key player in the emerging landscape of AI agents and reasoning models, promising to drive significant advancements in AI capabilities and applications. Recommended read:
References :
george.fitzmaurice@futurenet.com (George@Latest from ITPro
//
References:
www.itpro.com
, Salesforce
Cisco has announced new agentic AI tools designed to improve both customer and employee experiences. The unveiling includes the Webex AI Agent, set to be generally available later this month, which aims to address customer service requests around the clock with human-like natural language. Additionally, the firm introduced 'Cisco AI Assistant for Webex Contact Center,' offering suggested responses and real-time conversation transcripts for human agents.
Other offerings include ‘Workflow Automation in Cisco AI Assistant for Webex’ integrating Cisco’s tools with enterprise apps like Salesforce, ServiceNow, and Jira. ‘AI Capabilities in Webex Control Hub’ functions as a platform for IT admins to manage AI ecosystems and view analytics on AI usage and employee adoption. Cisco CPO Jeetu Patel stated that agentic AI is reinventing how people and technology work together across both physical and digital realms. Recommended read:
References :
Charles Lamanna@Microsoft 365 Blog
//
Microsoft is enhancing Copilot Studio with new capabilities to build autonomous agents, set to be in public preview at Microsoft Ignite 2024. These agents are designed to understand the nature of users' work and act on their behalf, offering support across business roles, teams, and functions. The goal is to transform business operations by automating complex tasks and streamlining workflows.
These autonomous agents can be configured, secured, and tested, automating tasks across apps and data sources for entire teams. Organizations are already utilizing Copilot Studio to create agents for specific business workflows, such as Pets at Home, which developed an agent for its profit protection team that could potentially drive a seven-figure annual savings. Copilot Studio plays a crucial role in customizing Copilot and creating agents for an entire company, enhancing efficiency, customer experience, and driving growth. Recommended read:
References :
stclarke@Source
//
References:
Source Asia
, Microsoft Research
Microsoft is aggressively integrating AI across various platforms to enhance productivity and transform business processes. A key focus is on understanding human-AI interactions through a project called Semantic Telemetry at Microsoft Research. This project employs a novel data science approach using Large Language Models (LLMs) to analyze user behavior and classify these interactions, aiming to build and support increasingly high-value use cases for Copilot and other AI tools. This innovative approach reimagines traditional telemetry by using LLMs to generate meaningful categorical labels from chat log data, providing valuable insights into how users interact with AI systems.
The integration of Microsoft 365 Copilot is also enabling companies like Estée Lauder to reimagine trend forecasting and consumer marketing. By building a generative AI ecosystem with Copilot Studio, Azure OpenAI Service, and Azure AI Search, Estée Lauder is leveraging AI to gather data, identify trends, build marketing assets, and inform research, ultimately accelerating the process of bringing beauty products to market. This transformation is part of Estée Lauder's "Beauty Reimagined" initiative, aiming for leaner, faster, and more agile operations. Recommended read:
References :
|
BenchmarksBlogsResearch Tools |