Agentic AI News

Delinea launches runtime authorisation for AI agents

Can Delinea stop AI agents before they touch databases, clouds, or SSH hosts?

Delinea has introduced runtime authorisation for AI agents, applying policy controls before actions are executed. The system extends its identity security platform to govern what agents do after they gain access to tools like databases, cloud consoles, Kubernetes clusters, and SSH hosts. The timing reflects a growing worry that access control alone is not enough once agents start acting on their own. For those tracking agentic AI products, this shows the market is moving deeper into action-level governance. The missing detail is how Delinea decides which agent steps get blocked in the moment.

New Research Finds Enterprise AI Failures Are Shifting Beyond Hallucinations as Companies Move from Chatbots to Agents

Why did hallucinations make up under 10% of ChatSee's 10,000 AI failure events, while escalation breaks kept growing?

ChatSee.ai says its study of more than 10,000 enterprise AI failure events found hallucinations caused less than 10% of cases. The biggest problems were resolution and escalation breakdowns, which made up 31.1% of failures, while execution and action failures rose 62% from the Q2 2024 baseline. For those tracking AI customer support agents, this signals the main risk is shifting from bad answers to broken handoffs and incomplete actions. The sharpest finding is that the systems may sound smarter, but they still fail when they have to finish the job.

JetHost Launches AI Connector, Bringing AI-Native Hosting Management to ChatGPT & MCP-Compatible AI Assistants

Can JetHost AI Connector let ChatGPT and Claude manage hosting safely through MCP today?

JetHost has launched JetHost AI Connector, a secure MCP integration for managing hosting accounts through AI assistants. The tool connects with ChatGPT, Claude, Cursor, Lovable, and other MCP-compatible systems. For those tracking agentic AI product launches, this signals assistants are moving from answering questions to controlling real infrastructure. The big question is whether natural-language access to servers and hosting stays secure under everyday use.

Stairwell Launches Backstory, the First Agentic Investigation Platform for Malware Blast Radius

Stairwell's Backstory traces 2.4 malware variants per sample, and says it can map an incident's blast radius in seconds.

Stairwell announced Backstory on July 29, 2026 as an agentic investigation platform for malware response. The launch follows analysis showing 2.4 related variants for every sample named in public threat reports. For those tracking agentic AI, this signals security tools shifting from alerts to autonomous investigation. The key promise is how fast Backstory can identify affected systems and connect the dots across an attack.

EliseAI Unveils New State of AI in Multifamily Report; 85% See Reduced Operating Expenses

EliseAI says 85 percent see lower operating costs, even as multifamily firms split on real AI adoption.

EliseAI released its second annual State of AI in Multifamily report on July 29, 2026, based on surveys of 350 U.S. decision-makers and 500 residents. The company says the data shows a sharp divide between surface-level adopters and firms rebuilding operations around AI. For those tracking AI agents in business operations, this points to a market where ROI is starting to separate serious buyers from casual users. The most useful detail is how much of the reported savings comes from operational changes rather than software alone.

BrowserStack Launches Test Companion, Agentic AI That Brings Complete Test Automation Into the IDE

BrowserStack's Test Companion brings AI test automation into the IDE, where release speed has been stuck behind coding speed.

BrowserStack launched Test Companion on July 29, 2026 as agentic AI for test automation inside the IDE. The tool is built to speed up test authoring, execution, debugging, and maintenance across web and mobile apps. For those tracking agentic AI, this signals testing becoming a bigger target for autonomous workflows. The detail to watch is whether moving testing into the IDE finally closes the gap between code and release.

Vendasta Unveils Autonomous AI Employees to Scale Organic Search and Social Marketing for Small Businesses

Vendasta logged 500-plus deployments in 24 hours as its AI Social Media Manager and AI Blogger went live.

Vendasta announced general availability of two autonomous AI Employees, the AI Social Media Manager and AI Blogger, on July 29, 2026. The launch targets SMBs that need a constant search and social presence as generative AI changes how people find businesses. For those tracking AI for sales and marketing, this shows tools that do the work, not just suggest it. The surprise is how fast customers adopted it, with 500-plus production deployments in the first day.

Fenergo launches Fen-AI to put compliance on autopilot

Fenergo's Fen-AI puts compliance on autopilot for banks, but can governed agents pass regulators' audit tests?

Fenergo has launched Fen-AI, an agentic AI orchestration platform for financial institutions handling KYC, AML, and CLM work. The platform is built to automate tasks across the client lifecycle while keeping humans in the loop, uniform policy controls, and a full audit trail. For those tracking agentic AI, this signals a push to move beyond chatbots into regulated workflows where action, oversight, and compliance must all work together. The big question is whether banks will trust KYRA, Fenergo's governed agentic workforce, to handle sensitive onboarding work at scale.

Surviving the paper deluge: a one-year study in learning from demonstration - Robohub

Can learning from demonstration survive the paper deluge, or will AI make robotics research harder to track?

Robohub highlights a one-year study on learning from demonstration as robotics research keeps expanding. The piece asks whether AI can help researchers keep up with the flood of papers or add to the overload. For those tracking agentic AI, this signals a core challenge in teaching systems to learn tasks from examples. The most useful takeaway may be how the field itself is struggling to organize its own knowledge.

Businessolver Introduces AI-powered Client Onboarding and Universal Data Gateway, Reducing Implementation Effort by 40%

Businessolver says AI onboarding cuts implementation effort 40% as it rolls out universal data gateway today.

Businessolver announced new AI-powered onboarding tools and a universal data gateway on July 28, 2026. The company says the changes reduce implementation effort by 40% and make benefits data easier to move across systems. For those tracking AI agents in business operations, this signals another attempt to turn automation claims into measurable workflow savings. The sharper test is whether the improvement holds up outside a promotional launch.

Cyabra Launches AI Agent That Automatically Investigates Coordinated Online Activity and Delivers a Verdict

Cyabra's new AI agent scans coordinated activity at scale, then returns one verdict with evidence behind it.

Cyabra announced Coordinated Activity, an AI agent that automatically investigates coordinated online behavior and delivers a single evidence-backed verdict. The launch was reported on July 28, 2026 and is aimed at governments and enterprises that need faster digital trust checks. For those tracking agentic AI, this shows agents moving into judgment-heavy workflows, not just task automation. The interesting part is how the system copies an analyst's method and then packages the result into one call.

Torq Unveils the SOC Brain: The AI SOC That Learns, Not Just Remembers

Torq says its SOC Brain learns from every analyst decision and historical incident inside its AI SOC platform.

Torq launched SOC Brain, a new layer for its AI SOC Platform that learns from historical investigations, analyst decisions, and organization-specific security operations. The company says the system builds a self-learning security operations center from day zero rather than relying on static rules alone. For those tracking agentic AI, this is a clear example of memory and precedent shaping autonomous decisions. The key detail is how the platform turns old investigations into operational judgment for new cases.

Soracom Introduces Air RTC Gateway, Routing IoT Device Voice by SIM Identity to Contact Centers, VoIP Providers, and AI Agents

Soracom's Air RTC Gateway routes device calls by SIM identity, cutting phone-number management for AI triage.

Soracom launched Air RTC Gateway to let IoT devices route voice calls by SIM identity to contact centers, VoIP providers, and AI agents. The product targets devices like parking terminals, gates, elevators, and emergency buttons that need live voice help. For those tracking AI agents in business operations, this shows how voice automation is moving into field devices and support intake. The key question is whether SIM-based routing can replace the costly phone-number systems that slow these deployments.

Microsoft Unveils Lower-Cost AI Model for Cybersecurity

Microsoft's MAI-Cyber-1-Flash joins GPT-5.4 in Project Perception, beating rivals on CyberGym while hunting code flaws.

Microsoft introduced MAI-Cyber-1-Flash, its first AI model built to find security flaws in source code, and paired it with GPT-5.4 inside Project Perception. The company said the agent system beat competing security models on the CyberGym benchmark. Cybersecurity teams are under pressure to catch bugs faster as codebases grow and attacks get more automated. For those tracking AI developer tools, this signals a push from code generation into code defense, with the benchmark win hiding one key detail about how the system works.

Mond Inc. Introduces All-in-1 Marketing Automation Platform AMONDLAB... Navigating from Planning to Uploading with one URL

Can Mond Inc.'s AMONDLAB really scrape one URL and run the whole marketing workflow by itself?

Mond Inc. has introduced AMONDLAB, an AI marketing agent service for small business owners and solopreneurs. The system automates steps from brand URL scraping and creative design to scheduling and publishing. For those tracking AI workflow automation, this is a clear example of a multi-step agent aimed at everyday business operations. The hook is whether its single-URL promise holds up once campaigns need revisions, approvals, or exceptions.

BridgeApp Launches AI Orchestration Layer That Automates the Full Software Development Cycle

BridgeApp launched July 27, 2026, promising to turn software tasks into pull requests without manual handoffs.

BridgeApp introduced an orchestration layer on July 27, 2026 that connects people, AI agents, tasks, and context in one workspace. The system is built to move software work from a to do item to a finished pull request with no tool switching. For those tracking agentic AI product launches, this signals a push toward tools that can own more of the development cycle. The catch is whether BridgeApp can really learn a codebase well enough to ship production-ready changes safely.

Gravity Launches AI Agent for Sustainability Teams

Gravity's new AI agent cuts data migration and disclosure drafting time by at least 70% for early customers.

Gravity has launched an AI agent for sustainability teams that handles carbon accounting, energy management, and disclosure work. Early customers say it cut time spent on data migration, disclosure drafting, and energy analysis by at least 70%. For those tracking agentic AI, this is a practical example of workflow automation with clear business impact. The hook is the mix of 85-plus skills and governance rules like human approval and audit trails.

RHA OneAI Announces Four Strategic Enterprise Client Wins in First Fully Operational Quarter

RHA Technologies said July 27 that RHA OneAI signed four strategic enterprise clients in its first fully operational quarter. The platform is pitched as a secure workspace that connects company knowledge, business apps, and multiple AI models. For those tracking AI workflow automation, this signals early demand for tools that can cut manual work across departments. The bigger question is which business processes those customers will actually let RHA OneAI run.

Metallic Minerals' La Plata Project Selected for U.S. DOE Genesis Mission AI Research Initiative Award Funding

Metallic Minerals joins DOE's Genesis Mission as AGAPEX targets faster mineral finds with agentic GeoAI funding in 2026.

Metallic Minerals said its La Plata Project was selected for U.S. DOE Genesis Mission funding through the AGAPEX research initiative on July 27, 2026. The project, led by the Colorado School of Mines, aims to build agentic GeoAI tools for critical minerals exploration. For those tracking agentic AI, this signals autonomous systems moving from chat and code into scientific search and field discovery. The key question is how much of the mineral hunt this research can automate.

Ode with Anthropic and LogicGate Announce Partnership to Scale Expert-Led Client Outcomes through Accelerated Agentic Capabilities

Anthropic and LogicGate are pushing agentic tools into GRC, where faster updates can change enterprise risk work.

Anthropic, Ode, and LogicGate announced an initiative to accelerate new agentic capabilities for LogicGate customers. The companies say the work will help enterprises build, implement, and update GRC programs faster. For those tracking agentic AI, this points to a broader push to embed autonomous tools inside compliance software. The key detail is how much expert-led workflow can be automated without losing control.

Unstop Launches Seven AI Agents to Automate End-to-End Enterprise Hiring

Can Unstop's seven AI agents really run hiring end to end, from sourcing to secure proctoring, in one system?

Unstop announced seven purpose-built AI agents for enterprise hiring on July 27. The tools cover sourcing, resume screening, calling, assessments, interviews, scheduling, and secure proctoring. For those tracking agentic AI, this is a clear example of end-to-end workflow automation in a high-volume business process. The most interesting part is how the system splits one hiring job into seven specialized agents.

Synopsys, Inc. and NVIDIA Announce Advancements to Agentic Ai for Engineering in Collaboration

Can Synopsys, Inc. and NVIDIA really run fully autonomous chip design agents with Nemotron and OpenShell?

Synopsys, Inc. announced new agentic AI advances for engineering with NVIDIA, including fully autonomous long-running capabilities for chip and electronics design. The system uses NVIDIA Nemotron on accelerated computing and the NVIDIA OpenShell runtime, built on Synopsys AgentEngineer technology. For those tracking agentic AI, this signals vendors are pushing deeper into hands-off engineering workflows. The catch is that autonomy in chip design is only useful if the agent can stay reliable across long, complex jobs.

KwaiKAT Team Releases KAT-Coder-V2.5: An Agentic Coding Model Trained on 100,000+ Verifiable Repository Environments

Kuaishou

Access Bank unveils app for SMEs in Nigeria

Access Bank launched an app for SMEs in Nigeria to adopt AI, boost productivity, and expand operations.

Access Bank Plc unveiled a new app for small and medium businesses at its maiden MSME Conference in Lagos. The bank says the tool is meant to help Nigerian firms adopt AI, improve productivity, and expand operations. For those tracking AI agents in business operations, this is a sign that banks are packaging AI adoption for smaller firms, not just large enterprises. The open question is how much real automation the app delivers beyond general AI guidance.

Bluesky’s AI assistant Attie expands into an open social research tool

Bluesky's Attie can now scan news and conversations across apps, turning social chatter into a searchable research tool.

Bluesky says users can now ask Attie questions about news, trends, and conversations on Bluesky and other apps on the AT Protocol. The expansion pushes the assistant from simple chat into cross-app research, where it has to gather and synthesize information on demand. For those tracking agentic AI, this signals a move toward assistants that can explore live social data instead of waiting for prompts. The key detail is whether Attie can pull useful answers from a messy stream of posts without missing context.

Yoroll Launches AI Odyssey, a Playable AI Film Inspired by Homer's Epic

Yoroll's AI Odyssey packs 140 AI videos and 11 endings, built solo in two days for about $250, with one twist.

Yoroll launched AI Odyssey today, a playable AI film built on its AI-native interactive video platform. Creator Amy Sun made it in two days on a budget of about $250, and the project uses more than 140 AI-generated cinematic videos. For those tracking AI video tools, this signals how interactive formats are moving from demo to real production workflows. The big question is which part of the experience keeps viewers coming back through 11 endings and branching choices.

From Stock Screeners to Smarter Research: How AI Is Changing Retail Investing

Can StocksToFind help retail investors sort thousands of filings, earnings reports, and news items into actionable research?

Retail investors now have access to filings, earnings reports, analyst estimates, insider trades, market news, charts, valuation metrics, and economic data across thousands of sites. The challenge has shifted from finding information to deciding what matters and connecting the dots faster. For those tracking AI Data Analytics Agents, this signals rising demand for tools that turn messy market data into a repeatable research workflow. The key question is whether the tools can do more than summarize and actually narrow the field to real opportunities.

[AINews] Black Forest Labs FLUX 3 - Multimodal Flow Models that beat Seedance 2.0, Gemini Omni and Grok Imagine, and FLUX-mimic video-action robotics model

Can Black Forest Labs' FLUX 3 really beat Seedance 2.0, Gemini Omni, and Grok Imagine on multimodal output?

Black Forest Labs says FLUX 3 outperforms Seedance 2.0, Gemini Omni, and Grok Imagine in multimodal flow models. The claim points to a sharper race in AI media tools that can move from images to video and robotics-style action. For those tracking AI video and media generation, this signals a model push that could reshape production workflows if the benchmarks hold up. The big question is which tests and tasks FLUX 3 won, and whether the gains survive real-world use.

JAMS Launches AI for Enterprise Job Scheduling: JAX and JAMS MCP

JAMS Software's July 24 launch lets enterprises query job runs without exposing onshore data, but one connector changes the risk.

JAMS Software said its JAX agent and JAMS MCP connector are now generally available for enterprise job scheduling. The tools let IT teams query, diagnose, and manage automation in plain language while keeping operational data inside their own network. For those tracking agentic AI, this is a clear push toward practical workflow control instead of chat-only assistants. The detail worth clicking for is how the open-standard connector and deployment model are meant to keep sensitive data onshore.

ExtraHop launches Agentic SOC Alliance with 15 members

ExtraHop and 15 partners launched an Agentic SOC Alliance to standardize autonomous security operations before buyers pick winners.

ExtraHop has launched the Agentic SOC Alliance with 15 founding members, including CrowdStrike, Dropzone AI and Torq. The group says it wants a shared operating model for autonomous security operations. Security teams are under pressure to cut alert backlogs and move faster on triage and response. For those tracking agentic AI, this signals a push to turn security automation into a common standard instead of isolated tools.

OutRival Launches Travel Division and Appoints Jeff Herzfeld

OutRival hired Jeff Herzfeld to run Travel as it pushes voice agents into hotels, cruises, and resorts.

OutRival has launched a dedicated Travel division and named Jeff Herzfeld as head of Travel. The company says the new unit will serve cruise lines, hotels, resorts, tour operators and travel agencies. Travel brands are looking for ways to handle more customer calls without adding staff. For those tracking agentic AI, this is a sign that voice agents are moving deeper into service and sales workflows.

Genesis Mission launches 278 AI-powered research projects across the US

The DOE picked 278 Genesis Mission projects across all 50 states, but the real race is what they’ll automate.

The U.S. Department of Energy has selected 278 research projects under its Genesis Mission, spanning all 50 states. The agency says the response was its largest ever for a funding call. The effort is aimed at AI-enabled scientific workflows to speed discovery in energy and science. For those tracking agentic AI, this shows where governments want systems that can help run research, not just answer questions.

TraceLink Unveils a Breakthrough Release of the OPUS AI Platform

TraceLink's OPUS AI Platform could unlock $350 billion in productivity gains, if companies actually trust no-code agents.

TraceLink unveiled a major release of its OPUS AI Platform for the life sciences and healthcare supply chain. The update lets organizations build custom OPUS Agents with no-code tools for business transactions and end-to-end network processes. For those tracking agentic AI, this is a direct bet that workflow automation will spread fastest where accuracy and traceability matter most. The open question is whether supply-chain operators will let software agents handle enough of the process to matter.

Dogpile launches web search API for AI agents By Investing.com

System1's Dogpile Fetch launches in under 2 minutes, giving AI agents web search, MCP access, and Deep mode.

System1 announced Dogpile Fetch, a web search API and Model Context Protocol server built for AI agents. It installs in Claude Code, Cursor, or other MCP-compatible clients in under two minutes and returns ranked results, knowledge graph data, and People Also Ask questions. For those tracking agentic AI products, this is a fresh piece of infrastructure that helps agents search and act inside workflows. The surprise is how fast it plugs in, and how much search behavior it packs into one tool.

dxFeed Launches dxScript: An AI-Native Scripting Language for Technical Analysis

dxFeed launched dxScript, a JavaScript-based scripting language built for technical analysis and tied to its market data feeds. The company says the platform includes a native integration layer for AI agents and large language models. For those tracking agentic AI, this matters because it gives bots a way to compute indicators and work directly with real-time and historical financial data. The biggest question is how that agent integration works inside the new language.

Lumonic Launches MCP Server, Bringing Audit-Ready Portfolio Data to Leading AI Assistants

Can Lumonic's MCP server let Claude and ChatGPT query private equity data without breaking audit rules?

Lumonic launched general availability of its MCP server for all customers. The server connects approved assistants like Claude and ChatGPT to Lumonic's portfolio monitoring workspace so users can query data directly. For those tracking agentic AI, this is another sign that assistants are becoming tools for real business work, not just chat. The interesting part is how the setup handles sensitive portfolio questions like missing financials and audit readiness.

Thryv Announces Integration with Breesy AI to Boost Growth for Franchise-Based Service Businesses

Thryv and Breesy AI say one integration can expose missed calls, lost revenue, and overloaded teams before jobs slip away.

Thryv announced an integration with Breesy AI aimed at franchise-based service businesses. The tools are built to show which leads convert, where revenue is lost, and whether teams can handle more call volume. The move fits a broader rush to use AI in customer intake and growth operations. For those tracking AI agents in business operations, this signals more vendors are selling decision support around frontline workflows, even if the automation is still limited.

Acrab Unveils GLIX 1 SoC and Agent Box, Bringing State-of-the-Art AI to the Edge

Acrab's GLIX 1 chip targets 100 billion-parameter models, but the company says the real prize is Agent Box.

Acrab unveiled GLIX 1, its first edge AI system-on-chip, alongside Agent Box, a personal edge AI system powered by the company's software stack. The launch centers on private, responsive AI agents running close to the user instead of in the cloud. For those tracking agentic AI, this signals another push to move capable assistants onto local hardware where latency and privacy matter more. The standout detail is that Acrab says the chip can support models in the 100 billion parameter class.

LALAL.AI Launches Lynx: Its First AI Model Built Exclusively for Voice Isolation & Noise Removal

LALAL.AI's first Lynx model can rescue location audio editors would have cut, but what noise does it erase?

LALAL.AI launched Lynx on July 23, 2026 as its first model built only for voice isolation and noise removal. The model targets background music, street noise, and other acoustic interference in recorded audio. For those tracking AI audio tools, this signals another push to make rough on-location sound usable inside professional video workflows. The key question is how much cleanup Lynx can deliver before editors decide to keep the take.

redSling Unveils Zenith, a Sovereign Agentic AI-Assisted Development Platform for the Enterprise

redSling launches Zenith for enterprises that want agentic AI development without giving up control of data or deployment.

redSling has launched Zenith, an agentic AI-assisted no-code development platform for enterprises. The company says it lets organizations build intelligent applications while keeping control of software, data, AI models, and deployment environments. For those tracking agentic AI products, this is a vendor pitch aimed at enterprise teams that want AI to build without losing sovereignty. The key tension is how much autonomy buyers will accept when the stack stays fully in-house.

Qoder Launches Qoder Security, Putting Three Layers of Security into the AI Coding Session

What are Qoder's three security layers catching inside the coding session if they boost detection 60%?

Qoder launched Qoder Security on July 23, 2026, adding built-in security review to its AI coding platform. The company says the feature moves code security from after-the-fact scanning to protection during the coding session. For those tracking agentic AI products, this shows vendors are pushing safety deeper into the action itself, not just the output. The catch is how much better it performs than older tools, with Qoder claiming about 80% fewer false positives.

Research-Grade EdgeBench Analysis: AI Agent Benchmarking, Leaderboard Analytics, Scaling Laws, and Evaluation Metrics

The piece breaks down EdgeBench as a benchmark for evaluating advanced AI agents across task types, runtimes, and interaction-time budgets. It pulls the dataset from Hugging Face and analyzes the taxonomy, execution settings, judging logic, and scoring metadata. The timing matters because agent builders are under pressure to prove reliability, not just demo flashy outputs. For those tracking the agent platform ecosystem, this signals that benchmarking and observability are becoming product features, and the leaderboard methods may matter as much as the models.

YubiKey 5.8 Ships Hardware-Backed Authorization for AI Agent Workflows

On July 21, Yubico shipped YubiKey 5.8 with CTAP 2.3 signing that can approve AI agent actions, not just logins.

Yubico released YubiKey 5.8 on July 21, adding hardware-backed signing for specific actions in AI agent workflows. The update uses CTAP 2.3 and WebAuthn so a human can cryptographically approve a database change or other agent task in the moment. For those tracking AI agent infrastructure, this points to a growing need for audit trails and human gates around autonomous systems. The most important shift is that identity alone is no longer enough to control what an agent can do.

AI Adoption Is Outpacing Leadership Readiness as New Research Finds Only 3% of Organizations Say Their Leaders Are Fully Prepared to Lead AI-Enabled Teams

How will ManpowerGroup Talent Solutions close the gap when only 3% of leaders are ready to run AI teams?

ManpowerGroup Talent Solutions says new research finds only 3% of organizations believe their leaders are fully prepared to lead AI-enabled teams. The finding comes as companies pour more money into AI and expect faster results from workers and managers. For those tracking AI data analytics agents and enterprise adoption, this signals that leadership gaps may slow real deployment more than the software itself. The sharpest warning in the report is not about model quality, but about trust, confidence, and who gets left behind.

Ushur launches the Ushur Agentic Platform: AI agents that finish the job

While enterprises spent years simplifying intake, Ushur’s July 22 platform targets the step that determines whether agents get trusted.

Ushur introduced the Ushur Agentic Platform on July 22, 2026. The platform is built to run AI agents that complete customer journeys end to end, not just start them. For those tracking agentic AI, this is a direct bet on operational automation with governance built in. The hook is the claim that enterprises can hand off the whole journey, not just fragments of it.

Zoom Revenue Accelerator unveils AI that drives revenue action, not just insights

Zoom Revenue Accelerator

Reserv launches usage-based AI claims licensing platform | The Insurer

Reserv spun out AiDE to license its AI claims stack, but insurers may want the pricing details.

Reserv has launched Apeiros Insurance Data Exchange, or AiDE, to license the AI claims technology behind its own adjusting business. The package includes models, infrastructure, data storage, and orchestration tools for insurers and brokers. That matters because vendors are now selling agent-like workflows as standalone products, not just internal systems. For those tracking agentic AI, the unanswered question is how this stack will be priced and adopted outside Reserv's core business.

Kim.cc Launches Sentinel Beta, Combining AI Insights with Human Oversight for E-Commerce Support

Kim.cc

Quantios Unveils Agentic AI Roadmap for Trust and Corporate Services

Quantios is rolling out agents across 600 TCSP firms, starting with a Timesheet Agent and a wider roadmap.

Quantios has published a platform-wide AI roadmap for trust and corporate service providers and launched a new Timesheet Agent. The company says earlier agents are already live across more than 20 firms in its early adopter program. The move comes as enterprise buyers want workflow tools that fit into regulated day-to-day work. For those tracking agentic AI, this is a clear example of a vertical vendor turning roadmap promises into active deployments.

Azumo Launches Valkyrie, an Open-Weight Coding Agent Built for Predictable AI Development Costs

Azumo opens Valkyrie to early users with flat-rate pricing, promising coding agents without runaway development bills.

Azumo has launched Valkyrie, an open-weight coding agent, and is starting with a friends and family test period before public release. The pitch is simple: give engineering teams AI help with a flat, predictable cost. That matters now as companies look for agentic tools that can write code without blowing up budgets. The early access rollout leaves one big question about how much control and capability the service will really deliver.

Michaels® Unveils 'Ask Mike' AI Assistant, Built With Google Cloud's Gemini Enterprise for Customer Experience

Michaels launches Ask Mike with Gemini Enterprise on July 21, and it guides shoppers from inspiration to purchase.

Michaels launched Ask Mike, a new AI shopping assistant built with Google Cloud’s Gemini Enterprise for Customer Experience, and it is live on Michaels.com. The tool is meant to improve product discovery and make shopping easier from browsing to checkout. For those tracking agentic AI, this shows retailers using assistants that can move customers through real purchase flows. The key question is how much of the shopping journey Ask Mike can handle on its own.

Mithra AI launches to verify data before AI models produce answers

Can Mithra AI stop bad data before models answer? The product screens content against a cryptographic source of truth.

Mithra Technologies, a subsidiary of ShelterZoom, launched Mithra AI to vet data before enterprise models generate responses. The product works one layer below the large language model and checks authenticity and user clearance as data enters the system. For those tracking agentic AI, this matters because agents are only as trustworthy as the data they read. The key detail is how the Single Source of Truth changes what agents are allowed to see.

PropulsionAI Launches Fay as Enterprise AI Agents Move Beyond Chatbots to Real Business Operations

PropulsionAI launches Fay on July 21, betting non-technical teams can automate complex operations without code.

PropulsionAI launched Fay, an enterprise AI agent and workflow platform built to help non-technical teams automate complex business operations. The company says the product is aimed at moving beyond chatbots and into real work across the business. For those tracking agentic AI, this signals growing demand for tools that can run workflows, not just answer questions. The interesting part is how far Fay goes in replacing manual steps across day-to-day operations.

S&P Global Launches Adaptive Retrieval, Giving Customers a New Way to Access Data Across AI and Agentic Workflows

S&P Global says Adaptive Retrieval serves autonomous multi-agent systems, while Deterministic mode still handles controlled pipelines.

S&P Global launched Adaptive Retrieval and says it now offers two ways to retrieve data inside its AI Data Portal. The company says the new setup supports both tightly controlled pipelines and autonomous, multi-agent workflows. For those tracking agentic AI, this shows data vendors building for both safe automation and more flexible agent behavior. The part worth watching is how customers choose between deterministic and adaptive access.

Crunchbase Launches MCP to Bring Private Market Intelligence Into AI Workflows

Can Crunchbase's new MCP give AI agents the private market data behind $510 billion in startup funding this year?

Crunchbase announced a new Model Context Protocol on July 21 that brings private market intelligence directly into LLM answers. The company says teams can now access its proprietary data and predictions inside AI tools and agents to move faster on decisions. For those tracking agentic AI, this is a practical example of data flowing straight into workflows instead of living in a separate dashboard. The key detail is how much of Crunchbase’s private market data agents will be allowed to use.

Saleoid Launches AI CRM for Small Businesses Starting at $5/Month

Saleoid launches a $5 AI CRM today, but the cheapest plan needs two years upfront.

Saleoid publicly launched its AI CRM for small businesses today, with pricing starting at $5 a month on two-year billing. The platform helps startups, agencies, consultants, and growing businesses capture leads, automate follow-ups, schedule appointments, and manage customer relationships. For those tracking AI sales tools, this signals another push to bundle outreach and CRM work into one low-cost product. The catch is that the headline price only applies if buyers commit to a long billing cycle.

Business News | PropulsionAI Launches Fay as Enterprise AI Agents Move Beyond Chatbots to Real Business Operations

PropulsionAI launched Fay on July 21, promising enterprise agents that let non-technical teams automate complex business operations.

PropulsionAI, founded by Harsh Nigam and Rohit Hazra, launched Fay as an enterprise AI agent and workflow platform. The product targets non-technical teams that want to automate business operations without building custom software. For those tracking agentic AI products, this is another sign the market is moving from chatbots to task-running systems. The unresolved question is how much real work Fay can handle before teams need human intervention.

Q-nomy Introduces AgentFlow to Orchestrate AI and Human Customer Journeys

Q-nomy announced AgentFlow, a new Q-Flow module, from Toronto on Tuesday. The product links AI-powered interactions with appointments, customer flow, business processes, and human-assisted service. For those tracking AI agent platforms, this signals a push to make agents work inside real enterprise workflows, not just chat on top of them. The catch is how much control Q-nomy keeps in the hands of rules, systems, and people behind each journey.

Veea Launches Free VigiLynx™ Service for Telcel Internet Gestionado -- Plug-and-Play Automation Platform Bringing Cybersecurity, Connectivity, Monitoring, and Edge Intelligence with Agentic AI in Mexico and Globally

Veea's free VigiLynx launch bundles four systems into one service for Telcel customers across Mexico and beyond.

Veea announced VigiLynx, a free commercial service now available through distributors for Telcel Internet Gestionado customers. The platform combines cybersecurity, connectivity, monitoring, and edge intelligence in one managed system. For those tracking agentic AI tools, this signals a push toward AI that handles real-world operations instead of chat alone. The launch stands out because Veea is packaging multiple products into a plug-and-play system with agentic AI at the center.

Chinese AI agent outperforms Anthropic's Claude Code in autonomous research

Qiushi Engine beat Anthropic’s Claude Code on ResearchClawBench, but the benchmark still asks which agent can really research.

As of Tuesday, Zhejiang University-led Qiushi Engine held the top spot on ResearchClawBench, with Open Science Desktop second and Claude Code third. The benchmark tests whether AI agents can independently carry out research and match human-written reference papers. It was created by a team led by the Shanghai Artificial Intelligence Laboratory. For those tracking agent platforms, this signals research agents are now competing on real task completion, not just chat quality.