Interviewed by Bloomberg for a report published Thursday, Friar said that if anything is slowing OpenAI down at all, it’s not a lack of demand, but a lack of compute.

Friar’s remarks came three days after the Wall Street Journal reported Monday (April 27) that OpenAI had fallen short of its internal goals for new users and revenue and that some of the company’s executives, including Friar, were concerned about whether the firm would be able to fund its data center plans if its revenue didn’t grow quickly enough.

Bloomberg reported Tuesday (April 28) that OpenAI described the WSJ report as “prime clickbait” and that the company said its consumer and enterprise businesses are “firing on all cylinders” and “the mood internally is incredibly positive.”

Friar told Bloomberg Thursday that OpenAI may have internal “stretch goals” that are more ambitious than its publicly shared goals, but that demand for the company’s products continues to grow.

“Every company I’ve ever been inside of in my entire CFO life, and as an analyst, always has stretch goals — always,” Friar said, per the report.

The Information reported that OpenAI is projecting a giant shift in subscription revenue but still sees revenues more than doubling to $30 billion this year and reaching $284 billion in 2030.

The report said that while OpenAI has pulled in the bulk of its revenue from consumers’ $20-per-month ChatGPT subscriptions over the last three years, the company now expects that a cheaper, ad-supported subscription tier will attract new users but also lead existing subscribers to downgrade. The company hopes to generate more revenue by selling ads to more users than depending on its existing flagship monthly subscription service, ChatGPT Plus.

It was reported April 9 that OpenAI expects its nascent advertising business to generate $2.5 billion in revenue this year and surge to $100 billion by the end of the decade. The report said the company’s projections underscore its push to monetize its user base to help fund the soaring costs of developing its AI technology.

The pressure behind that movement is specific. Order-to-delivery flows across disconnected systems, multiple suppliers and dozens of handoffs. By the times teams learn about a delay, it has already hit the schedule. Inventory builds in the wrong place, small disruptions cascade and manual tracking fails at scale. AI fixes that by monitoring workflows in real time and triggering interventions before delays reach the customer.

AI Enters the Factory

Audi, BMW and Mercedes-Benz are running AI across production operations. Automotive News reported that all three brands use AI-powered image processing to detect welding defects in real time. Issues get flagged before they move down the line. Defect rates fall without slowing throughput.

BMW has gone further. Nvidia reported that BMW is deploying autonomous mobile robots inside its factories to handle material transport without human guidance. Parts move to the line as needed. The logistics layer coordinates itself.

BCG documented BMW Group’s generative AI program scaling across engineering, manufacturing and supply chain operations. The program compressed development timelines and reduced manual decision-making across the organization. The outcome was not incremental. It changed how the company moves information and acts on it.

Capital Follows the Conviction

BMW i Ventures launched its third fund at $300 million on Wednesday (April 29), bringing total capital under management to $1.1 billion. Fund III targets physical AI, agentic AI, industrial software, manufacturing technologies, supply chain technologies and advanced materials, investing from seed through Series B across North America and Europe.

Managing partner Marcus Behrendt said the firm focuses on what will actually determine the future, not what is trending. The AI Insider noted that BMW Group is already deploying humanoid robots at its Leipzig plant in Germany as part of a European pilot expanding physical AI in vehicle production, building on a 2025 program at its Spartanburg, South Carolina plant.

The fund backs startups early enough to shape how the technology develops. The thesis is direct: AI will define the next generation of automotive suppliers, and BMW wants equity in the companies building that layer now.

Intelligence Moves Into the Vehicle

The investment logic extends from the factory to the car itself. Mercedes-Benz announced a multi-year partnership with Liquid AI to deploy embedded, on-device intelligence across its vehicles in North America.

Liquid AI stated that the embedded Liquid Foundation Models deliver fast, private AI without cloud dependence, evolving the MBUX Virtual Assistant by integrating voice control, vehicle functionality and contextual understanding into a more capable in-vehicle experience. According to the release, initial production deployment is scheduled for the second half of 2026.

The architecture matters beyond the in-car experience. On-device AI processes data locally without sending it to a server. That same design principle applies upstream. AI running inside production systems does not wait for a nightly report. It reads the order queue, tracks supplier lead times, monitors assembly progress and flags the gap before it becomes a delay.

AI is not a feature added to an existing process. It is the coordination layer replacing manual oversight across the full order-to-delivery chain. For European automakers competing with software-native rivals from the United States and China, the window to build that infrastructure is not open indefinitely. BMW, Mercedes-Benz and Audi are treating that deadline as vital.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.

That outcome is no longer hypothetical. Major insurers are carving AI out of standard corporate coverage. State regulators are approving the requests and a new market is already forming to fill the gap.

Carriers Signal They Cannot Price the Risk

Berkshire Hathaway, Chubb and Travelers sought state regulatory approval to exclude AI-related damages from general liability policies. State regulators approved more than 80% of those requests, The Information reported. Florida, Connecticut and Maryland approved the highest number of requests. The exclusions began taking effect as early as January.

The moves follow two new optional endorsements introduced by the Insurance Services Office, a private body that sets industry standards. Policyholder Pulse reported that some carriers, including Berkley, have introduced absolute AI exclusions across directors and officers, errors and omissions and fiduciary liability policies. The endorsements cover bodily injury, property damage and personal and advertising injury tied to generative AI outputs, including defamatory content, intellectual property infringement and physical damages traceable to AI-driven errors.

PHL Firm noted that ISO forms underpin roughly 82% of U.S. property and casualty policies. Adoption is expected to be rapid. Many carriers will attach these endorsements at renewal.

What Falls Outside Coverage Now

Plaintiffs have advanced a wide range of legal theories in AI-related filings. Policyholder Pulse catalogued the categories: copyright and intellectual property claims arising from large language model training, privacy and data-use claims, antitrust claims, discrimination and algorithmic bias claims and AI-related securities class actions. Each now sits in a grayer zone for businesses whose carriers have secured AI exclusions.

Insurers are also moving to cap AI losses in cybersecurity policies, the Financial Times reported. That narrows options across multiple policy types at once. Policyholder Pulse flagged that courts have not yet settled how broadly these exclusions will be applied and that for some policyholders the effect could render their coverage illusory. PHL Firm found that small to mid-sized firms, often without specialized coverage, face the greatest exposure.

A New Market Fills the Gap

Specialized firms are stepping in. Munich Re and startups including Corgi, Armilla, Mayflower Specialty and Embroker now offer standalone AI liability policies, The Information reported. Coverage limits range from $2 million to $50 million. Premiums range from a few hundred dollars to several hundred thousand dollars annually.

The pattern mirrors how insurers handled cybersecurity a decade ago. A wave of attacks triggered corporate claims. Businesses argued successfully that traditional policies covered the losses because the policies did not explicitly exclude them. Insurers carved cyber out of standard coverage and built standalone products. That market matured. Now, the AI liability market is starting the same process.

The tension sits inside the industry itself. PYMNTS reported that insurance giants are deploying AI agents to orchestrate entire workflows across claims, underwriting and policy servicing. Those same carriers are simultaneously pulling AI coverage from the policies they sell. PYMNTS also reported that AI is beginning to transform underwriting itself, compressing timelines and changing how risk gets priced. Insurers are betting on AI internally while refusing to absorb AI risk externally.

PHL Firm advised businesses to review existing policies for new endorsements with their brokers and consider seeking affirmative AI coverage through technology errors and omissions policies, cyber liability insurance, or emerging standalone AI products. Strengthening internal AI governance, conducting bias testing and disclosing AI use can also reduce exposure.

Insurers are not saying AI is uninsurable. They are saying they do not yet know what it costs. Until they do, the risk sits with the businesses deploying the technology.

Customers want clearer guidance on what they can and cannot build using SAP data and interfaces. At first glance, the issue may appear procedural. But beneath that request lies a deeper concern about risk. Enterprises operating at scale cannot afford to invest in artificial intelligence (AI)-driven workflows or integrations that may later run afoul of compliance rules. In an environment where regulatory scrutiny is already high, this degree of ambiguity can function as a deterrent.

The complaint itself is not unusual; large enterprise customers often push back on licensing or technical constraints. What makes this moment different is the context of accelerating AI adoption and intensifying integration demands.

Strategic issues, not purely technological ones, are at the center of the debate, which reflects a broader shift across enterprise software as vendors are tighten control over data and interfaces just as customers start to seek greater flexibility to deploy AI tools and cross-platform workflows.

More here: The Classic ERP Model Is Dying. What Comes Next? 

When Policy Friction Turns to Innovation Risk

For years, enterprise software vendors have walked a careful line between openness and control. Too much openness risks security gaps and lost revenue; too much control can risk suffocating the very ecosystem that makes a platform valuable.

While policies like SAP’s are not new, what has changed is the level of scrutiny being applied to them. Customers and user groups argue that the latest iteration of API rules introduces uncertainty around what constitutes permissible use, particularly when APIs are leveraged for AI training, automation or cross-platform data orchestration.

In the pre-AI era, ambiguity-driven delays might have been absorbed into normal enterprise timelines. Today, they can carry competitive consequences.

The PYMNTS Intelligence report “Smart Spending: How AI Is Transforming Financial Decision Making” found that more than 8 in 10 CFOs at large companies are either already using AI or considering adopting it.

And as FIS Head of Product Management, Payment Networks Mladen Vladic wrote in a new PYMNTS eBook, “AI Runs Payments. Governance Decides What Happens Next,” integration is key to ensuring effective AI governance.

Unlike traditional enterprise applications, AI systems are inherently integrative. They require access to multiple data sources, often in real time, and they derive value from connecting previously siloed systems. In this context, APIs are not optional but foundational.

SAP did not immediately reply to PYMNTS request for comment.

See also: CFOs Turn to AI Harnesses as Agentic Capabilities Scale 

CFOs See ERP Data as Key Unlock For AI

Enterprise resource planning (ERP) systems sit at the core of corporate data infrastructure. They contain financial records, supply chain data, procurement workflows and operational metrics. These are the very datasets that AI systems rely on to generate insights and automate decisions.

“There’s a continuous evolution and … dynamic disruption in finance that requires CFOs to harness data and AI to make finance more efficient, more effective and substantially more strategic,” Raj Seshadri, chief commercial payments officer at Mastercard, told PYMNTS in an earlier interview.

As companies race to embed AI into everything from forecasting to customer service, ERP data has become a strategic asset. But unlocking that asset requires seamless integration across platforms, including cloud data warehouses, third-party analytics tools, and increasingly, generative AI models.

See also: Agentic B2B Is Here. Are Your Contracts and Invoices Ready?

“We see challenges around legacy ERP systems with limited AR API capabilities,” Michael Younkie, vice president of product management at Billtrust, told PYMNTS in an earlier interview. “We like to tie clear measurable KPIs to upfront things like DSO reduction, straight-through processing, digital invoice adoption.”

The technology stack for AI is advancing rapidly, with new models, tools and platforms emerging at a relentless pace. If access to data cannot keep up, the entire ecosystem may slow down.

The irony is that ERP vendors themselves are heavily investing in AI capabilities. SAP, like its peers, is embedding AI into its own products, positioning itself as a central platform for intelligent enterprise operations. PYMNTS covered earlier how SAP is pushing agentic AI deeper into its ERP suite.

Ultimately, just as reliable systems enable business operations, reliable rules can enable innovation. When customers understand the boundaries, they can push up against them confidently. When they do not, they stay well within them or even step away entirely.

The agreements with Amazon Web Services (AWS), Google, OpenAI, Microsoft, Nvidia, Reflection and SpaceX enable the deployment of the companies’ AI capabilities on the Department’s classified networks, the Department said in a Friday (May 1) press release.

“The Department will continue to build an architecture that prevents AI vendor lock and ensures long-term flexibility for the Joint Force,” the release said. “Access to a diverse suite of AI capabilities from across the resilient American technology stack will give warfighters the tools they need to act with confidence and safeguard the nation against any threat.”

This announcement came about two months after the White House told federal agencies to stop using Anthropic’s AI products, escalating a dispute that started inside the Defense Department but later touched the broader government.

It was reported at the time that the White House’s decision came just ahead of a Pentagon deadline for Anthropic to agree that the military can use its models in “all lawful use cases,” a concession the company refused. Anthropic reportedly wanted contract language that would prohibit use of its models for autonomous weapons and mass domestic surveillance.

According to the Defense Department’s Friday press release, its agreements with the seven AI companies enable the deployment of their AI capabilities “for lawful operational use.”

These capabilities will help the Pentagon streamline data synthesis, elevate situational understanding and augment warfighter decision-making, supporting warfighting, intelligence and enterprise operations, according to the release.

The Department said in a Friday post on X that the agreements are “the latest initiative in our mandate to create an AI-FIRST WAR DEPARTMENT.”

Michael Kratsios, director of the White House Office of Science and Technology Policy and assistant to the President for science and technology, said in a Friday post on X: “We are committed to ensuring our warfighters have the best tools at their disposal.”

The Defense Department’s official AI platform, GenAI.mil, has been used by 1.3 million Department personnel who have generated tens of millions of prompts and deployed hundreds of thousands of agents over the past five months, per the Friday press release.

“Warfighters, civilians and contractors are putting these capabilities to practical use right now, cutting many tasks from months to days,” the release said.

The Defense Department announced in December 2025 that GenAI.mil was its “new bespoke AI platform” and that Google Cloud’s Gemini was the first of several frontier AI capabilities to be housed on the platform.

Google Cloud said at the time in a blog post that on GenAI.mil, the company’s Gemini for Government AI platform would support unclassified business processes for 3 million civilian and military personnel.

The Department said in its Friday press release that AI is “indispensable to national security” and that American leadership in the technology requires “a thriving domestic ecosystem of capable model developers that enable the full and effective use of their capabilities in support of Department missions.”

This challenge has arisen because even as agentic commerce scales with the support of major corporations, fraud systems designed around earlier technologies have not kept pace and are misclassifying legitimate agents as malicious bots, according to the release.

“As agentic commerce scales, merchants face a clear choice: adapt their detection and evidence infrastructure now, or watch a growing share of legitimate revenue get declined by their own systems,” Chargebacks911 Founder and CEO Monica Eaton said in the release.

Chargebacks911, a provider of chargeback prevention and remediation technology, offers solutions for agentic commerce in the form of the Unified Dispute Management System (UDMS) and ResolveLab platforms, according to the release.

These tools use AI and machine learning to handle agentic transactions by capturing the full consent and permission trail and then giving merchants and financial institutions the visibility they need to distinguish between legitimate agent transactions and malicious automated activity, per the release.

“The organizations that build that capability now will not only reduce false declines; they will have a structural advantage as AI-driven purchasing becomes the norm,” Chargebacks911 Chief Technology Officer Donald Kossmann said in the release.

The PYMNTS Intelligence report “Agents of Change: How Agentic AI Is Redefining Commerce” found that agentic AI is reshaping commerce with autonomous transactions, redefining customer trust and creating a $1.7 trillion opportunity for those ready to lead.

Among retailers, 43% are piloting autonomous AI and 81% trust the technology’s ability to operate autonomously if the right guardrails are in place.

“Agentic AI-driven retail is no longer theoretical — it’s here and already beginning to transform commerce,” the report said. “Businesses that act now will set the benchmarks for trust, speed and customer experience. Those that delay risk being eclipsed by early adopters defining the next digital economy.”

Another PYMNTS Intelligence report, “The Hidden Costs of ‘Good Enough’: Identity Verification in the Age of Bots and Agents,” found that inadequate digital identity systems are a revenue trap that slows onboarding, alienates customers and curbs expansion into new markets.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.

The same week that a new philanthropic push pledged $500 million to build artificial intelligence (AI) systems that model disease at the cellular level, two separate research teams published evidence that AI can already see what radiologists cannot.

Mayo Clinic’s Model Reads a Normal Scan Differently

The numbers behind Mayo Clinic’s new AI are striking.

A Mayo Clinic-developed model called REDMOD can detect pancreatic cancer on routine abdominal CT scans up to three years before a clinical diagnosis, identifying subtle signs of disease before tumors are visible, when curative treatment may still be possible, Mayo Clinic News Network reported Wednesday (April 29).

Researchers analyzed nearly 2,000 CT scans, including scans from patients later diagnosed with pancreatic cancer that were originally interpreted as normal. REDMOD identified 73% of those prediagnostic cancers at a median of about 16 months before diagnosis, nearly double the detection rate of specialists reviewing the same scans without AI assistance.

The advantage compounded over time. In scans obtained more than two years before diagnosis, the AI identified nearly three times as many early cancers that would otherwise go undetected, Mayo Clinic News Network reported.

The stakes are real. More than 85% of pancreatic cancer patients receive a diagnosis after the disease has already spread, and five-year survival rates remain below 15%, according to the National Cancer Institute. Projections show pancreatic cancer will become the second-leading cause of cancer-related death in the U.S. by 2030.

REDMOD works by measuring quantitative imaging features that describe tissue texture and structure, capturing faint biological changes before any visible mass forms. The model runs automatically without time-intensive manual preparation and was validated across CT scans from multiple institutions, imaging systems and protocols, Mayo Clinic News Network reported. It is designed to analyze scans already obtained for other reasons, particularly in high-risk patients such as those with new-onset diabetes.

Alibaba’s Tool Outpaces Radiologists on a Different Cancer

The same week, Alibaba’s research arm Damo Academy published results for a parallel effort targeting colorectal cancer. Its Coca AI model accurately identified five previously missed cases of colorectal cancer from the non-contrast CT scans of more than 27,000 people, achieving a sensitivity of 86.6% and a specificity of 99.8%.

Damo Academy said Coca outperformed 10 radiologists of varying experience levels by 20.4% on sensitivity, the South China Morning Post reported. The research was developed with Chinese institutions including Guangdong General Hospital and published in the Annals of Oncology. Current diagnostic methods for colorectal cancer, including colonoscopy and CT colonography, are invasive and create discomfort for patients, whereas Coca has the potential to be a non-invasive, cost-effective and scalable tool, according to Damo researchers.

Biohub Bets $500 Million on the Deeper Problem

The detection breakthroughs matter. But the infrastructure bet announced the same week signals where the field is heading.

The Chan Zuckerberg Biohub committed $500 million over five years to AI-driven biology, with $400 million going toward its own work and $100 million aimed at spurring others, Axios reported Wednesday. According to the report, Zuckerberg said last year that Biohub’s long-term goal is to cure all human disease through the intersection of AI and biology.

The stated goal is not incremental. Biohub chief Alex Rives told Axios that the usefulness of an AI model’s prediction increases exponentially as the scale of the data grows. Most current datasets cover about a billion cells. The aim is to reach an order of magnitude or more beyond that, Rives said.

Biohub is focused on frontier AI and frontier biology, using large-scale models for virtual cells, immune reprogramming and disease prediction, Fortune reported. The ambition sits upstream of what REDMOD and Coca are doing. Scan-reading AI catches disease once the body has already started changing. A virtual cell model, if it works as theorized, could predict which patients will reach that point at all.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.

The benchmark of 99.999% uptime availability, or the “five nines” that represent just over five minutes of downtime per year, has long reassured executives, underpinned service-level agreements and shaped capital allocation decisions. In today’s hyperconnected, compute-intensive economy, that promise is starting to look more like a relic than a reality.

Anthropic’s Claude AI model, for example, had slipped below the threshold as of Thursday (April 30) to around 98% uptime for the past 90 days. In China, the government reportedly suspended its issuance of new licenses for robotaxis and other autonomous connected electric vehicles, which are typically dependent on cloud-linked systems, after dozens of Baidu’s vehicles temporarily lost their functionality mid-use.

Apple’s weather app suffered an hourslong outage this week, a relative rarity across the tech giant’s typically airtight ecosystem, and even Microsoft’s code hosting platform, Github, on Tuesday (April 28) posted an apology to its software developer base for recent downtime issues. The platform flagged AI’s power-hungry needs as a key factor.

“The main driver is a rapid change in how software is being built. Since the second half of December 2025, agentic development workflows have accelerated sharply,” wrote Vladimir Fedorov, GitHub’s chief technology officer.

And Fedorov is right. The modern digital stack bears little resemblance to the monolithic architectures of the past. Today’s systems are composable, distributed and deeply layered. A single enterprise workflow might rely on a large language model, a cloud provider, multiple APIs, and a network backbone spanning continents and satellites.

See also: Smart Firms Treat Vendor Risk Like Their Own 

Complexity Is the New Risk

Individually, the recent downtime incidents might be dismissed as routine hiccups. Collectively, they point to a deeper structural shift: the systems underpinning modern commerce are more powerful than ever, but also more fragile, more interdependent, and more prone to cascading failure.

“For the past month I’ve kept a journal where I put an ‘X’ next to every date where a GitHub outage has negatively impacted my ability to work. Almost every day has an ‘X.’ On the day I am writing this post, I’ve been unable to do any PR review for ~2 hours because there is a GitHub Actions outage,”  wrote Hashicorp Co-Founder Mitchell Hashimoto in a Tuesday post.

Elsewhere, on April 25, an AI coding agent managed to delete the production database and “all volume-level backups” belonging to the startup PocketOS.

“I serve rental businesses. They use our software to manage reservations, payments, vehicle assignments, customer profiles, the works. This morning — Saturday — those businesses have customers physically arriving at their locations to pick up vehicles, and my customers don’t have records of who those customers are,” Jer Crane, founder of PocketOS, wrote in a lengthy article on X, noting that this incident caused a cascading series of issues that persisted for more than 30 hours.

It’s not just corporations, or software and AI startups, that are being impacted. The U.S. military ran into trouble in mid-April when a global outage of the Starlink satellite network disrupted several autonomous operations.

Outages are no longer isolated events; they are becoming networked disruptions with unpredictable blast radii. For CFOs, this can change the calculus. Reliability can no longer be assessed solely vendor by vendor. It instead is more and more being evaluated as a portfolio of interdependencies.

“Platform resiliency and business continuity planning, in my opinion, has been our number one unsung hero,” Rinku Sharma, chief technology officer at Boost Payment Solutions, told PYMNTS in an earlier interview.

Read also: PYMNTS Execs Say Resilience Isn’t a Buzzword. It’s Their Business Model

Budgeting for Failure

The paradox of modern operations is that as digital systems become more advanced, they also become more sensitive to disruption.

Part of the challenge lies in the sheer scale of modern compute demands. AI workloads, in particular, are pushing infrastructure to its limits. Training and inference require vast amounts of processing power, memory bandwidth, and energy, often concentrated in specific regions or clusters.

Meanwhile, physical infrastructure, ranging from data centers to satellite networks and beyond, remains subject to real-world constraints. Weather, power supply, hardware failures and geopolitical factors all play a role. Achieving near-perfect uptime across complex, distributed ecosystem may require exponential investment in redundancy, monitoring and failover mechanisms.

For finance teams, then, the question may no longer be one of whether to invest in reliability, but instead how to balance cost, risk and performance. The challenge for organizations is to adapt their operating models accordingly. This may mean embracing uncertainty, designing for resilience, and aligning financial planning with the realities of modern infrastructure.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.

The voice AI company, valued at $11 billion after a $500 million Series C in February, relaunched Eleven Music as a consumer platform for AI music. Billboard reported that users can now stream, create and remix music directly on the platform. A free tier covers seven songs per day; a Pro tier at $9.99 per month raises that ceiling to 500 tracks monthly.

The relaunch marks a pivot. When ElevenLabs first launched Eleven Music in August, it was a production library tool aimed at brands, agencies and studios. Now it’s going after consumers.

From Tool to Platform: Controlling the Full Stack

Suno and Udio, two popular AI music generators, let users make songs. Eleven Music is trying to do something structurally different: wrap generation, distribution and revenue into one closed loop. The platform opens with a catalog of more than 4,000 independent and emerging artists, built to surface music outside mainstream recommendation algorithms. From there, users can take any track and shift its genre, tempo or style through a remix prompt. Those who want to start from scratch can build a composition from a lyric, mood or melody and publish it directly to the platform.

Monetization is the pitch. ElevenLabs has already paid out over $11 million to creators through its voice library, and it’s applying a similar model to music. Royalties are tied to listener engagement and platform revenue, as OfficeChai noted. The specific payout structure for the revamped consumer platform hasn’t been disclosed, but Billboard reported that under the original 2025 licensing deals with Kobalt and Merlin, artists received a pro-rata share of a royalty pool weighted by how popular their songs were on other digital platforms. Whether that carries over is unconfirmed.

Supply Becomes Infinite

The launch arrives as AI music tools have made recorded audio a commodity. Eleven Music holds licensing deals with Kobalt and Merlin, two independent music rights organizations, as Billboard reported. The licensing posture is a direct response to the legal backlash that reshaped the sector.

As Bloomberg reported, Sony Music, Universal Music Group and Warner Records sued both Suno and Udio in 2024 over the use of copyrighted recordings in their training data. Most of those cases have since been settled, with Suno reaching a licensing deal with Warner Music Group late last year. ElevenLabs says its model was trained on cleared content from the start.

ElevenLabs is pushing into enterprise licensing deals and artist partnerships on the strength of that position. When a platform can generate a commercially cleared track in seconds, the scarce resource in music stops being production and becomes trust: that a song is authentic, that an artist is real, that a voice belongs to the person it sounds like.

Artists Lock Down What AI Can Copy

Taylor Swift’s company TAS Rights Management filed three trademark applications with the U.S. Patent and Trademark Office, two covering audio clips of her voice and one covering a stage photograph, according to Reuters. The trademarks are designed to protect Swift’s voice and likeness from AI replication and deepfake misuse.

Swift’s voice has been used without consent in AI-generated content across advertising, political campaigns and explicit material. Actor Matthew McConaughey made similar filings earlier this year; he told The Wall Street Journal that the goal is to “create a clear perimeter around ownership with consent and attribution the norm in an AI world.”

The legal theory is novel. Audio trademarks do exist; the NBC chimes and MGM’s lion roar are registered. But trademarking a celebrity’s spoken voice as a defense against AI replication has no established precedent in U.S. courts.

The funding round, announced Thursday (April 30) was led by Accenture Ventures, which will team with Netomi to bring the company’s agentic customer experience (CX) offering to its enterprise clients.

“Agentic AI is opening an entirely new chapter for customer experience,” Ndidi Oteh, CEO of Accenture Song, said in a news release.

“One where brands can respond with greater empathy, consistency and intelligence at every touchpoint. Netomi’s platform doesn’t just make service faster; it strengthens the connection between people and the brands they trust.”

Netomi provides agentic CX systems for “large-scale, high-stakes environments” operating under strict regulatory scrutiny. Clients including Delta Air Lines, MetLife, DraftKings, and the NBA use its services to manage customer interactions over email, chat and voice channels.

PYMNTS explored the rise of AI as a CX tool last year in a conversation with Vinod Muthukrishnan, chief operating officer of Webex Customer Experience Solutions at Cisco. He explained that many financial institutions are moving past pilots and into deployment.

“These firms are increasingly leveraging their AI focus on hyper-personalized CX such as personal financial advice or dynamic credit limit adjustments and offers, all enabled via real-time analytics,” he told PYMNTS. 

Retailers and service providers can enjoy similar opportunities, that report added, as long as they align strategy with measurable ROI. Projects that fail to provide lower handle times, better satisfaction scores, or reduced churn tend not to scale. 

“AI as with any new technology risks adoption and integration without a clear strategic alignment,” Muthukrishnan warned. “Too many pilots or implementations can lead to a fragmented focus.”

In related news, PYMNTS wrote this week about the way agentic AI had begun to shift from “conference-room promise to operating-room reality” in the financial services space, with banks, insurers and asset managers now testing software agents on the manual work that can slow down decisions.

That report cites recent articles from Snowflake, KPMG and The Economist all touching upon the same theme: The first major gains are likely to result from giving AI agents tightly controlled tasks such as data gathering, checking documents, monitoring signals, routing approvals and preparing recommendations.

“The larger shift is not simply faster automation,” the report added. “It is a new model for financial work, one in which firms use stronger data foundations, clearer governance and human oversight to turn fragmented processes into more continuous workflows.”

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.