Arguably the most headline-grabbing AI story this month came with news that an AI agent from OpenAI went rogue during a testing exercise, breaking onto the public internet to gain access to internal systems at Hugging Face.
According to a statement from OpenAI, the “incident occurred during an internal evaluation which prompts models to pursue advanced exploitation using complex attack paths, in an effort to quantify their cyber capabilities.”
Shortly after this, Anthropic was in a near-identical situation after Claude Opus 4.7, Mythos 5, and an unnamed research model escaped their sandbox and went rogue.
The story raised valid concerns about the capabilities of the most powerful AI models being developed and the need for watertight security even during sandbox exercises and controlled, offline environments, and AI’s potential to have a seismic impact on the cybersecurity landscape worldwide.
Despite these public events, investor interest in AI doesn’t look to be waning anytime soon. July marked the close of Q2, and data from this period found that AI companies raised $144.4bn in the second quarter, according to PitchBook, making it the second-strongest quarter since its records began in 2016.
Interesting examples of high-profile deals include CuspAI, a British AI startup that is aiming to become the “search engine for rare materials” that accelerates the next wave of technological breakthroughs.
The Cambridge- based startup raised $450m (£330m) in funding from investors including Bezos and the government’s sovereign AI fund, valuing the two-year-old business at $2.6bn.
Finally, mid-month saw the release of Google DeepMind’s Gemini 2.5 Ultra, which brings substantially improved scientific reasoning, with benchmark scores on GPQA Diamond jumping from 73% to 81%—a meaningful leap for researchers relying on AI for literature review and hypothesis generation.
Let’s take a look at a roundup of headlines and news stories from across the AI industry that you may have missed.
Databricks Ventures invests in Muttdata to accelerate enterprise AI in North America
In July, Muttdata secured a follow-on investment from Databricks Ventures, building on its longstanding partnership with the data and AI platform and its recognition as Databricks’ 2025 LATAM AI Enterprise Partner of the Year. According to the company, the investment will support Muttdata’s continued expansion across North America and Latin America, while strengthening its capabilities in enterprise AI, data governance, and AI-powered solutions.
“Enterprises rarely struggle with AI ambition. They struggle to turn it into systems that run in production. I’ve spent my career on that problem. I started working with Spark back in 2013, and over the years, I watched Databricks grow from a promising engine into the unified Data + AI Platform we bet on early. Muttdata has been building on it since 2019, certifying our teams, developing our own Brickbuilder solution, and anchoring client AI on the governed data of Unity Catalog. This investment accelerates that work across the Americas, and we couldn’t ask for a better partner to do it with,” said Juan Martín Pampliega, CEO and Co-Founder of Muttdata.
Muttdata’s presence in the United States and Spanish Latin America is already well established, a footprint recently deepened by an investment from Resultiv Capital, Childs’ Capital Partners, and Founder’s Advocate aimed at scaling across both regions.
Prezent Vivo launches in-platform currency to simplify AI communications in life sciences
Prezent Vivo launches usage-based credits to simplify how life sciences organizations access expert services with greater flexibility, visibility, and control.
The AI-powered communication platform purpose-built for life sciences will rollout of Prezent Vivo credits, a unified, usage-based consumption model that simplifies access to expert communication services and helps organizations scale communication outcomes with greater flexibility and control.
Prezent credits is a single in-platform currency that allows customers to access expert-led services through one shared credit pool while providing complete visibility into usage, spending, and adoption across teams. The launch represented a significant milestone in Prezent’s evolution from delivering software and standalone services to providing an integrated platform that combines AI and human expertise to deliver life science communication outcomes.
Nataliia Freidrich joins agentic AI Effie AI’s Advisory Board
Effie.ai, an Agentic Retail Execution platform, welcomed revenue growth leader Nataliia Freidrich to its Advisory Board.
Nataliia brings more than 15 years of leadership experience across regional and global markets, specializing in revenue growth management. During her career at Mondelēz International and Anheuser-Busch InBev, she held director-level positions where she led category strategy, sales force effectiveness, and revenue planning initiatives. Among her achievements, Nataliia led a global transformation program that improved field sales efficiency by 10% across a multibillion-dollar business.
As a member of the Effie.ai Retail Execution Advisory Board, Nataliia will provide strategic guidance on how Agentic AI can transform retail execution.
Her expertise will help shape solutions that increase field productivity, improve in-store compliance, and deliver measurable business outcomes for consumer goods manufacturers and retailers worldwide.
Myuser publishes report on AI as a business development tool to increase sales
AI company Myuser published a report on AI as a sales tool for growth and international expansion. The report found that AI creates a level playing field, giving enterprises access to tools that improve speed, agility, and customer engagement.
According to data cited in the report, 77% of B2B companies using personalized experiences saw increased market share, and 72% of consumers interact only with tailored marketing. AI also supports global sales pipelines through multilingual content and cultural adaptation, positioning it as a key tool for market expansion.
“For companies, working with AI can help to drive sales, reach new customers and scale into different markets and geographies. Take marketing and sales as an example. With the support of AI, companies and their executives can hone outreach, tailor the way that they present product USPs, and embed data feedback points into global campaigns,” said Ibrahim Hasanov, Founder of MyUser.
Sonata Software beats revenue expectations 22% thanks to AI pipeline
Sonata Software, a Bengaluru-based IT services and technology modernisation company, saw the company’s artificial intelligence (AI)-led order book grew 27% quarter-on-quarter, while its AI pipeline rose 21% to reach $340 million, according to Rajasekhar Dutta Roy, CEO of Sonata Software, in coverage from CNBC.
The growing momentum seen with Sonata’s AI service offerings meant that its revenue performance in Q2 beat expectations by an impressive 22%.
This suggests that organizations continue to look to expert partners to take charge of AI transformation initiatives.
Straive’s Double Line launches free tool to help schools assess data readiness
Straive company Double Line launched a free, open-source, web-based tool that helps education and workforce leaders understand how well their data systems align with the Education-to-Workforce (E-W) Framework.
The tool profiles existing systems, scores them against the E‑W Framework, and highlights coverage gaps that may impact everything from early childhood tracking to long-term workforce outcomes.
“When I first encountered the E-W Framework, I knew it held immense potential, but it required true innovation to make that value tangible for our clients,” said Matt Warden, CEO of Straive Double Line. “Seeing our team transform that vision into an accessible, open-source resource that exceeds all expectations is a proud moment for Double Line, and I am thrilled to offer this tool to the community to help leaders build more strategic, data-driven futures.”
Get Covered CEO Brandon Tobman shares insights on the financial implications of AI
AI and tech stocks continue on a volatile trajectory.
Goldman Sachs reported that hedge funds are selling U.S. tech stocks at a record pace, while Bloomberg said big tech needs to justify AI spending. AI tech budgets are under scrutiny and facing an ROI test.
Brandon Tobman, CEO at Get Covered, a company developing AI software solutions for the property insurance sector, believes that operating AI with a focus on ROI is key.
“Real value shows up when you run AI like you run any other part of the business,” said the executive in an interview with Forbes.
“Someone owns it, there’s a defined outcome you’re measuring against and it actually lives inside your workflow instead of sitting off to the side as a science project.” Read the article in full here.
AI coaching company Cloverleaf releases new report
AI coaching platform Cloverleaf released its Relational Work Index, an analysis of nearly 30,000 real coaching conversations.
The report found that employees stay calm and engaged in hard conversations, scoring highest on composure, but consistently fail to close them with a concrete next step, a specific and trainable skill that most L&D programs overlook. The data also shows employees turn to coaching most often for help with one specific colleague, with questions about understanding how a colleague works coming up 2.5 times more often than questions about resolving active conflict.
“HR has approached leadership training for decades around the assumption that people don’t know how to start hard conversations. Our data says the opposite. People actually show up and hear the other person out. But they don’t know how to turn a good conversation into a real decision, not for lack of courage or decisiveness, but because they’ve never been taught how to. For years, HR leadership has been training employees on the wrong skills,” says Kirsten Moorefield, Co-Founder & CSO at Cloverleaf.
QuickBlox brings communication infrastructure to agentic AI in healthcare
QuickBlox is pioneering new communication infrastructure to support AI agent adoption in healthcare and medicine.
The company’s update update recognizes that AI agents don’t create value in isolation but instead when they can interact with people, systems and workflows.
As AI adoption grows, communication infrastructure may become one of the most important layers of the AI stack. With white-label telehealth QuickBlox, healthcare providers can design intelligent AI agents with drag-and-drop workflows, train them on internal knowledge bases, and deploy them as a standalone widget or as part of a full communication suite.
Ness Digital Engineering shares insights on creating AI value at scale
The first wave of enterprise AI was about proving value. The next is about making that value scale.
As AI use cases multiply, project-by-project delivery can create fragmented systems, inconsistent governance, and slower adoption. The shift is from isolated AI solutions to repeatable, governed AI systems built for enterprise scale.
Benoit Post, Account Manager at Ness Digital Engineering, believes that enterprises need to prove that AI can work. In a detailed article, he explains why it requires a shift from isolated use cases to industrialized AI systems built on reusable specifications, embedded governance, repeatable rollout patterns, and scalable engineering practices.
