Intuit’s CEO Says These AI Tools Will Streamline Your Accounting Department

As temperatures in New York City hit record highs this past Tuesday, an event space in Manhattan offered some much-needed respite from the heat—as well as a glimpse at how AI may soon change the face of small business bookkeeping. Intuit, the tech giant behind the QuickBooks accounting platform, previewed technology that it’s now making public: a suite of AI agents meant to handle some of the behind-the-scenes accounting and bookkeeping work. Each of four AI tools on display promised to automate a different aspect of business management on QuickBooks—payments, accounting, finance and customer acquisition—with video walk-throughs showing the software doing everything from reconciling books to suggesting what late fee a business owner should implement to ranking inbound sales leads as either “cold,” “warm,” or “hot.” Payroll, marketing and project management bots are also on their way, according to the company. It’s always a good idea to take these pre-screened tech demos with a grain of salt, especially when it comes to AI, a technology that can act in surprising and unpredictable ways. Still, Intuit seems all-in on the machine learning revolution, and as it rolls out these agents (a process that, for American customers, starts July 1), it’s worth understanding the company’s vision for the future of accounting. Inc. spoke with Intuit CEO Sasan Goodarzi about why his company is embracing AI and how he plans to maintain customer privacy and security in the process. What are you guys launching, and how does it fit into your vision of the role AI will play in bookkeeping? The biggest thing we’ve learned from our customers is that they’re using way too many apps, spending way too much money and don’t really know what’s going on in their business. And when you think about a business, whether it’s a solopreneur or a several hundred million dollar business, time is everything: so what we are launching is a virtual team of AI agents and AI-enabled human experts that can, in essence, do all the work for our customers while still leaving them in control. We have 12,000 experts that sit on our data and AI platform and are not only making our AI agents better every day, but are actually working with the AI agents to deliver a customer experience. Is this the first time you’ve integrated agentic AI into one of your products? This is years in the making in terms of all of the data investments that we’ve made—because AI is useless without data—and all of the data services we’ve built. So the essence of what we’ve launched has been in the works for years. About a year and a half ago, we launched multiple agentic experiences. But this is the first time we’ve launched a broad-based set of agents that can manage everything from leads to cash. So if I’m using QuickBooks, what does my interfacing with the AI look like? In very practical terms, we’ll go into their Gmail with their permission—if that’s the source of what they use to communicate—or SMS, and be able to help them manage which leads are hot versus cold. Then they can click on the hot leads and we’ll help them interact with that customer, suggest what they should communicate. We’ll create an estimate for them, and an invoice. So we feed them the most important actions and choices they should make so that it’s forefront—like, following up on invoices, or ‘It’s time to take out a line of credit.’ But they’re always in control. They always see why we’re recommending something, and if they choose to change it, they can. We’re in a CPA shortage right now. Do you see this as a workaround to having less than enough human accounting labor? In general, with anything that’s very driven by intensive labor, a lot of it will get automated by AI—but I also believe that it will fuel the reallocation of people’s time to things that are more productive. 75 percent of accountants are retiring, and the inflow of accountants has really dried up, so we see this as an opportunity not only to fuel the success of our accounting partners but to automate a lot of the manual work. People are obviously dealing with a lot of sensitive financial information and data in this system. What is your approach to data privacy and security? Trust is everything, and so we declared more than 20 years ago a set of data principles that we’ve adhered to ever since. One is that it’s the customer’s data, not ours. Two, we would never sell the data and we would only use the data for the benefit of our customers, which is why the way we have built our platform—both our data models and our AI models—only our Intuit financial large language models get trained by our customers’ data. The data never leaves our four walls. The data is never used to train LLMs, as an example, that are outside our four walls. We also have governance; we have a lot of human spot-checks in terms of how the models are operating to make sure there’s no bias in the models. We have AI models that inspect AI models as well as human experts that inspect the models. Accuracy is so key in something like accounting, and a lot of people are still concerned about hallucinations with AI. What can you say about how you ensure accuracy with high-stakes bookkeeping and financial decisions? Everything comes down to accuracy. That’s why what we’ve launched today is such a big deal, because it’s about financial forecasts, accounting, taxes. We’ve invested heavily in our own models. Anytime we detect an anomaly where we believe it’s not accurate, we automatically bring in a human expert to confirm. A lot of companies are still figuring out how to integrate AI into their workflows. What are some best practices you would encourage small business owners to adopt when implementing AI? In order for you to grow, you need to digitize your entire business so that everything from lead to cash is digitized. Then, when we show them what our platform can do and the capabilities which are all data and AI-driven, that’s where there’s a light switch, which is: ‘Rather than using all these different apps that don’t talk to each other, I can run my business in one place.’ So for a business that we serve, it’s all about having one platform that automates all their workflows, versus the customer automating their own workflows. Are there any mistakes you see small businesses make when they’re incorporating AI? Customers are over-digitized today. Four years ago, when I talked to customers, whether they were businesses or accountants, it was: ‘How do they move from Excel, Google Sheets, shoeboxes with receipts to using a platform to run their business.’ Today, they’re using up to 10 apps to run their business—and the larger they are, the more apps they use. They use one app to manage their pipeline; they use one app for estimating; one app for invoicing; one app for accounting. All their data is trapped in a bunch of different apps. BY BRIAN CONTRERAS @_B_CONTRERAS_

I Asked ChatGPT a Simple Question. It Literally May Have Saved My Life

As a general rule, one thing you should definitely not do is start typing random symptoms you might be experiencing into Google Search, or, say, an AI chat box. If you’re concerned that something might be wrong, you should probably just go see your doctor. I use ChatGPT a dozen or so times a day for research and getting answers to general information questions. I ask it about everything from how to fix a chainsaw to explaining company earnings reports. That’s probably why I did the thing you aren’t supposed to do, even though I know better. It started a few months ago with something I couldn’t quite put my finger on. I’d get short of breath when doing ordinary things like walking up a flight of stairs or mowing the lawn, or carrying a bag chair to the sideline of a soccer game. In May, I was exhausted after walking with my roller carry-on to the gate of a redeye flight. All of those are things I’ve done hundreds of times, but recently it seemed to take a lot more effort. For months, I chalked it up to getting older, needing more exercise, or maybe just a busy season catching up with me. However, the symptoms didn’t go away—they got worse. I began coughing more often. Breathing deeply made me wince. And worst of all, I was tired all the time. Still, I would have put the symptoms into the category of annoying, not alarming. I have long suffered from seasonal allergies, so the trouble taking deep breaths and coughing was easy to play off as just a bad year for tree pollen. I didn’t think it was an emergency. I wasn’t sick in the usual sense—no fever, no sore throat, no runny nose. Just a slow, steady decline. I’d heard stories about long COVID, maybe that was it? Or maybe it was just stress. I kept putting it off. Until one night when I randomly did the thing I do dozens of times a day: I asked ChatGPT. “I’ve been having some strange symptoms over the past few months. What is most likely wrong?” Then I listed what had been bothering me: an increased shortness of breath when doing routine physical activity like walking up stairs or mowing the lawn a feeling of congestion in my lungs, resulting in coughing that sometimes brings up material… sometimes that includes a slight reddish tint extreme fatigue and increased blood pressure. The AI didn’t hedge. The very first possibility it gave me wasn’t a cold. It wasn’t stress. It wasn’t even anxiety. The top result: “The combination of progressive shortness of breath, cough with possible blood (hemoptysis), and fatigue, especially when paired with elevated blood pressure, could point toward a cardiopulmonary issue, with congestive heart failure (CHF) or pulmonary hypertension among the leading concerns.” Then came the most alarming part. “These symptoms warrant medical evaluation without delay,” it wrote. “Please see a cardiologist or primary care doctor right away — or go to urgent care or an ER if the symptoms are getting worse.” I froze. Heart failure? That felt dramatic. Like something that happens to much older people, or people with a known heart condition. I’m relatively young. I’m not an athlete, but I don’t smoke or consume more than a few glasses of alcohol a year. Still, the answer rattled me enough that I called my doctor the next morning. After a quick physical exam, they ordered a battery of tests. First, a CT of my lungs to rule out blood clots, then bloodwork, and finally an echocardiogram. It’s never a good sign when you’re having an ultrasound of your heart and the sonographer stops, gets up from their chair, and says, “I’ll be right back to finish the rest of these images,” before leaving the room. The only reason that happens is because they saw something very bad, and they’re going to get a doctor. Sure enough, a moment later, he came back in the room with a cardiac fellow. “Do you know why your primary doctor ordered this test?” he asked. “Well, he was worried about blood clots in my lungs, but we ruled that out with the CT,” I replied. “He ordered this test to check for congestive heart failure.” “Yeah, you have heart failure,” the doctor said. Specifically, the echo showed that my ejection fraction, a measure of how much blood your heart empties every time it pumps, was around 25. That’s a little less than half of what is normal. My BNP—a blood marker that rises when the heart is under stress—was dramatically elevated. My heart was struggling to keep up, and my body was sending all the signals. I just didn’t know how to read them. What’s scary is that I almost didn’t do anything. I almost didn’t ask the question. I almost didn’t listen. I assumed I was just out of shape and had a bad case of allergies. Without that initial answer pushing me to take it seriously, there’s a good chance I’d still be trying to power through it—getting worse without knowing why—all the while putting myself at risk of not knowing why. The truth is, heart failure can be subtle in its early stages. The signs aren’t always dramatic. You might not clutch your chest or collapse. You might just feel a little more tired. A little more winded. A little off. But when your heart is struggling to circulate blood effectively, your entire body feels it—and if you ignore it long enough, the consequences can be irreversible. Google, by the way, suggested I might have bronchitis or tuberculosis. Neither of those is something you should ignore, but I was pretty sure I didn’t have TB, and bronchitis doesn’t usually last months, and often comes with a fever. ChatGPT, on the other hand, noticed a pattern. It gave me language to describe what was happening. It raised a red flag. It helped me take myself seriously. And, it nudged me to take the whole thing seriously and get checked out. That’s not just impressive. In my case, it may have been lifesaving. It’s easy to treat AI chatbots like a novelty or a parlor trick. But the thing they are really good at is taking complex information and synthesizing in ways that you wouldn’t be able to do on your own. Sometimes, asking the right question at the right time—no matter who, or what, you ask—can change everything. Apple’s CEO, Tim Cook, has frequently said that it thinks the company’s greatest contribution could be in health care. I think that may be true. I am amazed by the ability of technology to reveal information about ourselves that we could never know or understand on our own. We are incredibly fortunate to live in a time where we have such ready access to heart monitoring and other data sensors built into our smartwatches. I share this story to encourage you. If your body is telling you something, listen. If something feels off, don’t wait until it becomes a crisis. Technology is not a substitute for medical care, but it can be a powerful tool for insight, direction, and action—especially when you’re stuck in that limbo between “I’m probably fine” and “I might need help.” Thankfully, modern medicine is incredible, and this is something doctors are well-equipped to treat. I’ve been given more meds than I ever thought I’d be taking at 45 years old, and we’re working on ruling out a few other possible causes of my heart failure (once the insurance company comes around about what’s actually “medically necessary”—but that’s a different story). This is something I’ll be managing for the rest of my life, but thanks to a simple question I asked ChatGPT, I’m hoping it will be a long one. I reached out to OpenAI for comment on this story, but the company did not immediately respond to my request. EXPERT OPINION BY JASON ATEN, TECH COLUMNIST @JASONATEN

Your Employees Hate These Tasks at Work. They Say AI Can Help

New research commissioned by AI writing tool Grammarly and conducted by Talker Research found nearly half of the workers who responded hate the repetitive office tasks that make up the daily grind. The 44 percent total is no surprise, and you’ve probably had similar thoughts when you have to fill in a travel budget request form for Steve in Accounts—yet again. But it’s the AI era, and workers are increasingly aware that there are tools that can help wipe out this recurring drudgery—and 62 percent of the survey respondents said there are plenty of tasks they’d like to speed up with AI. The Grammarly study, which involved 2,000 knowledge workers, defined as people who work with computers in some way in the office, showed that certain tasks that have to do interrupt their momentum, getting in the way of productive work. The New York Post reported that respondents they have to undertake 53 of these tasks a week, on average, and they cost the typical worker some 3.5 hours of meaningful work. And while over four in 10 people hate carrying out repetitive tasks, there’s a generational spread in the data: 57 percent of Gen Z workers don’t enjoy mundane tasks, only 42 percent of Gen X feel the same. Anecdotally, this makes sense—Gen Z is used to living online, dealing with the fast-paced, constantly changing digital realm of social media, memes, and a world that never seems to stand still. That’s part of why many reports say that they feel very differently about the typical office job and workplace norms. The survey also asked what kind of tasks workers would like to use AI for, and how they’d like the experience to be. Fully 35 percent of people wanted AI to help with the tiresome task of drafting an email, and 34 percent said they’d want an AI to help with repetitive tasks like sorting spreadsheet data. Another 33 percent wanted the AI to draft meeting notes so they wouldn’t have to, and 31 percent wanted AI to carry out workflows automatically. The speed at which AI tools are evolving is a big factor in respondents’ interest in convenience. The data also showed 49 percent of workers would favor a tool that’s easy to use, and 35 percent said they wanted one that’s easy to prompt. Prompting an AI is a skill that can be developed, and it can greatly impact a generative tool like a chatbot, leading to a tighter, more useful output if you properly refine your questions . Since only 38 percent of the respondents’ companies have a preexisting policy on AI use, this makes sense—many companies are lagging on training staff and openly allowing them to use AI to save time. Over 50 percent of respondents said they wished their company was more open to AI use, and Gen Z led this sentiment, with 67 percent feeling this way, compared to just 45 percent of Gen X. Why should you care about these statistics? They do, after all, simply back up plenty of reports about the rising popularity of AI, and the fact that younger workers are more keen to adopt the technology to relieve them of humdrum office tasks than older workers are—even though, on the whole, roughly half of all workers are keen to use AI to do mundane jobs. The data even back up recent advice from the likes of entrepreneur Mark Cuban, who pointed out that AI is already great at taking on repetitive tasks. You should care precisely because the survey shows you how your staff could be using AI to free up their day to tackle more productive work. And because if your company is pro-AI, and you really want to reap the full benefits, it would be best for your business if every age cohort in your workforce was as keen to use AI as every other. Grammarly’s data show that maybe you should put an official AI use policy in place. Then you should train users of every age on how to make the most of what AI can offer, instead of relying on your digitally savvy younger workers to carry out your AI wishes. BY KIT EATON @KITEATON

Tensions Are Flaring Between Microsoft and OpenAI

Cracks are reportedly forming in the previously-solid relationship between OpenAI and Microsoft as the two companies renegotiate the terms of their partnership. According to new reporting from The Information and The Wall Street Journal, the two companies have been negotiating for more than eight months regarding issues of ownership, profit-sharing, and exclusivity. Here’s the context: OpenAI is attempting to convert from a non-profit into a for-profit public benefit corporation (also known as a B Corp). According to The Information, OpenAI wants Microsoft, by far its largest investor, to only have a 33 percent ownership stake in the company. The same report alleges that OpenAI also wants to alter terms that give Microsoft exclusive rights to resell OpenAI’s API in the cloud, and prevent Microsoft from getting access to AI code editor Windsurf, which OpenAI is currently in the process of acquiring. In 2019, years before the debut of ChatGPT, Microsoft invested $1 billion in OpenAI as part of a deal that gave the Bill Gates-founded company exclusive rights to resell OpenAI’s API in its Azure cloud computing platform. In total, Microsoft has invested over $13 billion in OpenAI. According to a January 2025 blog post, OpenAI and Microsoft’s contract will expire in 2030. Other cloud computing providers, like Google and Amazon, would likely jump at the chance to resell OpenAI’s popular models. The year 2030 is also when Microsoft will lose exclusive access to OpenAI’s IP, and according to The Information, Microsoft is looking to extend that period of exclusivity. Microsoft uses OpenAI’s tech to power many of its new and AI-upgraded products, and OpenAI’s models serve as linchpin for Copilot, Microsoft’s lineup of AI-powered personal and work assistants. In addition, OpenAI is reportedly looking to change the terms of its revenue-sharing deal with Microsoft. Under the current deal, OpenAI will share 20 percent of its revenue with Microsoft until 2030. Now, OpenAI is hoping to reduce that percentage. As for Windsurf, a platform that uses AI to assist software engineers, the Information reports that leaders at OpenAI are concerned that their relationship with Microsoft could complicate a planned $3 billion acquisition of the company. That’s because Microsoft also owns GitHub Copilot, a major competitor to Windsurf. According to the Journal, tensions between Microsoft and OpenAI have deepend to the point that OpenAI executives have discussed going to antitrust regulators with accusations that Microsoft has been engaging in anticompetitive behavior. However, OpenAI and Microsoft told The Information in a joint statement that “we have a long-term, productive partnership that has delivered amazing AI tools for everyone. Talks are ongoing and we are optimistic we will continue to build together for years to come.” BY BEN SHERRY @BENLUCASSHERRY

5 Ways Entrepreneurs Are Rethinking SEO Amid the Rise of GenAI

Cody Barbo has a folder on his phone filled with all the AI assistants, and at least once every month, the Trust and Will co-founder and CEO asks ChatGPT, Claude, Gemini, Meta AI, Grok, and every other model the same question: What is the easiest, cheapest, and best place to set up my will online? The entrepreneur wants to see how his San Diego startup ranks against the rest of the estate planning industry when it comes to the answers that large language models are spitting out to hundreds of millions of users. “We’re competitive, so we’re like, is it us? Is it the OG, LegalZoom? Or is it a new player?” says Barbo, whose company has landed on the Inc. 5000 for the past two years after posting an average three-year growth rate of 1,127 percent. “Our search volume is getting eaten up by these tools.” There’s no standard term yet for what Barbo is doing, but founders have started calling this practice “search everywhere optimization,” “generative engine optimization,” and “answer engine optimization.” While the acronyms vary, founders agree that finding the SEO equivalent for AI is the future of marketing, as more consumers take their queries to AI agents instead of search engines. Even though the space is rapidly evolving in real time, founders are uncovering which strategies are most effective at surfacing company mentions in generative AI responses. Here’s what they’ve learned. Invest in long-form content Founders say one approach that has really worked so far is investing in long-form editorial content. That’s why Trust and Will has put so much emphasis on educational content on its website, publishing regular articles by industry experts and conducting a 10,000-person study in January. Brand-tracking software company Tracksuit has taken a similar approach, publishing articles about marketing, case studies, and original research. “Right now, it seems to be long-form content that works,” says Tracksuit co-founder and CEO Connor Archbold. “White papers and actual research, I think, will become important.” Become the trusted answer Length is not the only factor to consider when creating editorial content for AI to crawl and cite. Founders recommend leaning into a niche and becoming the go-to source for industry leaders’ frequently asked questions. “In the era of AI-powered search, visibility means becoming the trusted answer,” says Cassi Janakos Chavez, co-founder and COO of corporate lactation services company and two-time Inc. 5000 honoree Healthy Horizons. “We focus on maintaining our website’s authority with in-depth resources that show we are the leading expert on workplace lactation.” Prioritize founder-first storytelling When it comes to optimizing AI search, Tyler Eide, founder and creative director of the Seattle-based design agency Parker Studio, has been advising his clients to lean into founder-first storytelling. “There’s only so much you can control… so the best thing to do is to control what you can,” says Eide, whose company has worked with Google, Lululemon, and Zappos. That means having his founder clients ask themselves: “How can I get out into as many places as possible with the story that I want people to have?” Expand your entire digital footprint If links are the currency of search rankings, brand mentions are the most important factor for determining visibility in AI responses, says Andy Crestodina. The co-founder and chief marketing officer of Orbit Media Studios, a Chicago-based digital agency focused on web development and website optimization, says, “Have the biggest digital footprint. Make sure your brand is everywhere.” Go on podcasts, participate in webinars, post on LinkedIn, write for every possible website, issue press releases, and repeat your elevator pitch in any video that may get transcribed. Make sure your company is listed on every review site, trade association website, and conceivable directory. Crestodina calls this basic digital public relations. “Fill the web with your company name and surround it with relevant industry terms. Your new goal is to be in as much AI training data as possible,” says Crestodina, whose company has landed on the Inc. 5000 three times. “Be all over the place.” Keep testing Like Barbo, keep experimenting to see if and when your company name comes up after prompting AI agents with questions. “Best practices are good hypotheses. Everything I just suggested should be tested,” says Crestodina. “Use data to confirm or reject that hypothesis, and then iterate, rinse, repeat. It takes a lot of humility to be good at this.” BY ALI DONALDSON @ALICDONALDSON

Google’s New AI Feature Turns Search Results Into Podcasts

Auditory learners rejoice: Google has introduced an experimental new feature that turns some search results into AI-generated podcasts. The feature is called Audio Overview, and it can only be accessed by opting in on Google’s Search Labs page. According to Google, Audio Overview allows users to “listen to a concise conversation generated with AI, providing a preview of information from the top search results in response to your query, with links out to the web to explore more.” Audio Overview is powered by Gemini, Google’s family of AI models, and functions similarly to NotebookLM, Google’s popular app that enables users to upload documents in order to generate a podcast conversation between two AI-powered voices. These artificial voices summarize and discuss the uploaded material, and can be directed to structure their conversation in specific formats, such as an audio FAQ or as a study guide. Google advertises NotebookLM as “your personalized AI research assistant.” With Audio Overview, users don’t need to upload documents to NotebookLM in order to generate a podcast. When you type a question into the Google search bar, a widget may appear in the results asking if you’d like to generate an Audio Overview. According to Google, the Audio Overview feature will activate when Google’s systems “determine it might be useful.” When asked “how are small businesses using AI,” the Audio Overview generated a five-minute podcast in which the two AI voices discussed AI’s ability to help people generate content, analyze market trends, and create targeted advertising campaigns. Audio Overview also lists the websites it used to source its answer, so people can double-check its accuracy. Early users are encouraged to give feedback on the quality of the Audio Overviews as Google fine-tunes the feature. For entrepreneurs who do their best learning by listening to audiobooks and podcasts, or who spend a lot of time traveling, Audio Overview could be a useful tool for researching and studying a subject while multitasking. To activate Audio Overview, users will need to navigate to Google’s Search Labs page, where the company keeps its experimental new search features, and turn the Audio Overview experience on. Google says the feature is currently only available with English-speaking voices and only accessible in the United States. BY BEN SHERRY @BENLUCASSHERRY

New Research Finds That AI Is Creating More Jobs and Higher Pay

Opinions on how artificial intelligence will affect employment differ considerably—often radically. The most recent demonstration of the divide arose last week, when Anthropic CEO Dario Amodei warned that AI apps like the kind his company is developing risk pushing joblessness up to 20 percent. That led serial entrepreneur Mark Cuban to counter with the prediction it’s more likely the tech will enable full employment instead. Now, a new study by consultancy PwC comes down somewhere in between those views, albeit far closer to the optimistic Cuban perspective. In its 2025 AI Jobs Barometer report, PwC said forecasts by skeptics that AI will unleash employment doom have failed to materialize in industries already embracing the tech most. It also noted the initial consequences of AI adoption included higher job creation, increased pay for those new positions, and reduced inequality by allowing people without university degrees to combine their skills with many AI-executed tasks formerly reserved for knowledge workers. “In contrast to worries that AI could cause sharp reductions in the number of jobs available, this year’s findings show jobs are growing in virtually every type of AI-exposed occupation, including highly automatable ones,” said PwC global chief AI officer Joe Atkinson in the report. “AI is amplifying and democratizing expertise, enabling employees to multiply their impact, and focus on higher-level responsibilities.” Those and other findings in the study may serve to undermine some of the more dire predictions made by AI critics. To create the study, PwC analyzed 800 million job posting and combed through thousands of financial reports of businesses in a variety of sectors. In addition to finding companies that had integrated the tech most reported increased job creation, the consultancy found five other ways AI adoption creates positive effects. For example, wages rose “twice as quickly in those industries most exposed to AI compared to those least exposed,” it said—even for people working in highly automatable roles. One reason for that increased pay was surging productivity, with the companies having embraced AI the most quadrupling their output compared with slower adopting businesses. With AI increasingly automating research, writing, accounting, and a lot of other duties previously handled by knowledge workers, employees without college degrees were able to add those tasks to the skills they were initially hired for by using AI. By broadening the range and deepening the value of their work for employers, those people benefitted from the average 56 percent wage premium that businesses have accorded people with AI capabilities, PwC said. In other words, rather than pushing employees out of jobs by automating them, AI has thus far permitted workers to enhance their value to businesses by doing a wider number of tasks, boosting their productivity, and increasing their pay as they did. But to make that shift, the PwC report said, people—and their employers—had to be both willing and quick to adjust to the demands of the quickly developing tech. “AI’s rapid advance is not just reshaping industries, but fundamentally altering the workforce and the skills required,” said PwC global workforce leader Pete Brown. “This is not a situation that employers can easily buy their way out of. Even if they can pay the premium required to attract talent with AI skills, those skills can quickly become out of date without investment in the systems to help the workforce learn.” Of course, skeptics may argue that in examining initial consequences of the fairly recent and still developing tech, PwC may have merely focused on an initially positive phase the could be followed by mass job cuts as companies master AI and automate as many roles as they can. But PwC global chief commercial officer Carol Stubbing argues that history suggests otherwise. “We know that every time we have an industrial revolution, there are more jobs created than lost,” Stubbing said, with the caveat that employees will need to continually revolutionize their skills to harness the powers of AI. “So the challenge, we believe, is not that there won’t be jobs. It’s that workers need to be prepared to take them.” BY BRUCE CRUMLEY @BRUCEC_INC

At WWDC, Apple has to address its two biggest pain points: AI and developer relations

Apple’s annual Worldwide Developer’s Conference (WWDC) has always been more than just a showcase of software updates—it’s a statement of intent. It reveals to the developer community—as well as to the world at large—what Apple thinks is most important about each of its platforms. But as WWDC 2025 approaches, the stakes feel dramatically higher than usual. With tech giants like Google, Microsoft, and OpenAI sprinting ahead in artificial intelligence, the pressure is on Apple not only to catch up—but to prove it still belongs in the conversation. What we can expect This year’s WWDC is shaping up less like a grand unveiling and more like a make-or-break update. Apple already introduced Apple Intelligence at last year’s WWDC, promising privacy-first generative features and a smarter Siri. But in the year since, very little has materialized. The features arrived late, rolled out slowly, and failed to generate much excitement. Many of them—notification summaries, writing tools, and image playgrounds—have been problematic and remain a disappointment compared with the competition. There are suggestions that we may see updates that make Apple Intelligence slightly more accessible to third-party developers—such as new APIs for summarization, task automation, or suggested replies. And Apple could announce that it’s bringing Google Gemini on board as an option for users in the same way it announced ChatGPT integration last year. But none of this is expected to move the needle dramatically. What Apple is likely to do is what it always does: emphasize its privacy advantages, show off beautifully controlled demos, and wrap incremental upgrades in the language of elegant design and trustworthiness. What we won’t see What’s unlikely to show up is exactly what Apple arguably needs most: a clear leap forward in generative AI. Despite launching Apple Intelligence last year, Apple still doesn’t have a competitive large language model of its own, and there’s no indication it plans to introduce one. Most of the AI work is still being powered by OpenAI behind the scenes, with Apple acting more as a front-end for someone else’s technology than a platform leader. Also missing: any meaningful upgrade to Siri. For all the talk of AI-enhanced assistants, Siri remains inconsistent, brittle, and far behind the real-time reasoning and multimodal capabilities demonstrated by OpenAI’s GPT-4o or Google’s Gemini Live. There’s no expectation Apple will unveil a Siri that can navigate your device, interact fluidly with apps, or hold a context-rich, back-and-forth conversation. What’s also off the table is any kind of public roadmap toward Apple building its own foundational models. Unlike Meta, Google, or even smaller players like Anthropic, Apple has remained silent on training its own GPT-class LLMs or competing in the model layer. There’s also no public cloud-based developer model in sight, leaving Apple notably absent from the infrastructure layer of AI. So while Apple may refresh its AI pitch, expand access to existing features, and try to recapture some excitement, it will still be operating from behind—repackaging existing partnerships and product polish as innovation. The lack of a bold leap forward could once again reinforce the perception that Apple is falling further behind in a race that’s rapidly redefining the future of computing. Absence of a strategy The absence of a clear and ambitious AI strategy would be more than a missed opportunity—it could be a strategic blunder. In the past year, Microsoft embedded AI into Windows, Office, and Azure. Google placed AI front and center in Android and Search. OpenAI launched GPT-4o, a multimodal assistant capable of real-time voice conversation, coding, and document analysis. Nvidia became the world’s most valuable chipmaker, and Meta open-sourced increasingly competitive models. Apple, meanwhile, has remained conspicuously quiet. That silence, once mistaken for secrecy or caution, now risks being interpreted as stagnation. If Apple shows up at WWDC and doesn’t address its two biggest pain points—AI and developer relations—it will come across as out of step with the rest of the tech industry, and out of touch with one of its most important audiences. Worse, Apple’s hesitation could threaten its control over the user experience. If the most advanced AI models live inside ChatGPT, Gemini, or Copilot, then users will increasingly rely on those tools across their devices—even on iPhones. The irony? Apple, the company that revolutionized how we interact with devices, could become a middleman in someone else’s AI ecosystem. The moment of truth To be fair, Apple’s slow approach is not without logic. It prizes user privacy, and generative AI presents significant privacy risks. Running models on-device rather than in the cloud is a key part of Apple’s identity, and that kind of development takes time. Still, the market is shifting, and consumer expectations are changing. AI isn’t just a backend feature anymore—it’s becoming the front end of computing. If Apple doesn’t show a compelling AI vision at WWDC, it could find itself boxed out of that future, even if its devices remain dominant in the present. This WWDC won’t just be a software update; it’s a referendum on whether Apple can still lead in defining the next era of computing. The company that once made “insanely great” products must now convince the world it can make insanely smart ones, too. Tim Cook and Apple’s executive team are walking a tightrope. If they were to deliver a truly compelling AI story—especially one that respects user privacy and works seamlessly across devices—they could reset the narrative and take a leadership role on their own terms. All signs, however, point to Apple just pressing through without acknowledging what everyone else can already see: the company is far behind and has nothing to show for its efforts to catch up. EXPERT OPINION BY JASON ATEN, TECH COLUMNIST @JASONATEN

Meta Eyes $10 Billion AI Investment

Meta Platforms is reportedly considering a more than $10 billion investment in artificial intelligence startup Scale AI. The deal is still in talks and could potentially change, Bloomberg News reported on Sunday, citing “people familiar with the matter.” Scale AI declined to comment and Meta did not return Reuters’ request for comment, the outlet reported. Scale AI is a data labeling startup backed by Nvidia, Amazon, and Meta. It also helps researchers exchange AI-related information, with users in more than 9,000 municipalities. It was founded in 2016 and was last valued at nearly $14 billion.

Mark Cuban Just Made a Bold Prediction About the Future of AI and Human Interaction

Mark Cuban, billionaire entrepreneur, basketball team owner, pharmaceutical-industry-upsetter, and former regular on the iconic Shark Tank reality TV investment show, knows a thing or two about innovation. And yesterday, he took to Bluesky to make this bold prediction about AI: Within the next 3 years, there will be so much AI, in particular AI video, people won’t know if what they see or hear is real. Which will lead to an explosion of f2f engagement, events and jobs. Those that were in the office will be in the field. Call it the Milli Vanilli effect It’s worth taking notice. Cuban’s effectively saying we’ve reached the end of the complex “will AI steal my job?” debate. And the answer is yes, more or less. At the very least, Cuban painted a bold picture about how AI is going to transform society in the next 36 months or so, and that picture includes a dramatic change in where new jobs will be created. His argument is simple. There is going to be so much AI and it will advance so much, he thinks, that we just won’t be able to resist it. The content that AI systems will generate will also be so convincing, or so useful, that people may more or less give up trying to tell if something is real or not. So society will fall back to relying on genuine face-to-face (f2f, as Cuban abbreviates it) human connections for fun and work—something that AI just can’t replace. Milli Vanilli, in case you don’t know, was a successful German R&B duo from the 1980s and ‘90s, selling some 30 million singles internationally. But their fans were shocked when the band’s manager revealed the two singers, Fab Morvan and Rob Pilatus, were just lip synching and didn’t actually perform any of the vocals on their hugely popular tracks. Their performance had been so convincing that they won a Grammy Award for Best New Artist in February 1990. (It was subsequently revoked.) By invoking this musical scandal, Cuban suggests he thinks AI is going to do the same thing—act convincingly like the real thing, while actually being artificial—but pretty much everywhere. This is going to upset many traditional models for how we live our lives, including how we work. His “those that were in the office will be in the field” line is particularly eye-opening. This seems to be a prediction that the traditional office job will be more or less dead. That’s one in the eye for all the micromanaging demonstrated by company leaders enforcing strict “return to office” rules right now. Instead, Cuban imagines that AI will be so convincing that people won’t rely on phone calls, texts, emails or even video calls to complete their jobs—and they’ll have to be out and about, traveling, actually meeting clients, customers, contractors, and co-workers in person. In Cuban’s mind, this transformation will create a whole host of new jobs. Picture it as a 21st-century equivalent of the arrival of the desktop PC in the workplace—a revolutionary device that abolished the old-fashioned typing pool, and later even many secretarial posts. But it also created a whole new slew of jobs, from programmers to Excel experts to the engineers in company IT departments. Some commenters replying to Cuban’s post agree. “You’re absolutely right. The digital will become unreliable. We’ll have to get back together in person,” wrote one. Another noted that they already “tell up and coming musicians, performers, actors this all the time. become exceptional—because exceptional, authentic experiences and connections will be more valuable than ever.” But other commenters struck a more skeptical note, pointing out that “AI is scary, but I think if companies start replacing people, riots are going to happen and it won’t happen,” or even suggesting “A bill making any artificial account posing as a human being (bots) illegal is desperately needed, like yesterday.” Why should you care about this? For one reason: Cuban’s optimism stands out among much of the much bleaker forecasting about AI’s impact on the world. For example, an MIT professor recently issued a stern warning that we can stop AI from stealing everyone’s job—but only if we act really fast, to prevent AI from steamrolling over everything. If you can bring some of Cuban’s optimism into your workplace, that might lift worker’s worries about AI. Meanwhile you should take note of his warning and really plan for some serious re-skilling or up-skilling of your workforce to get them ready for the face-to-face future he predicts. Then you might be better placed to benefit from the AI revolution. BY KIT EATON @KITEATON

Why This IBM Exec Says AI Adoption Should Be Led by HR

HR is the natural choice to lead company-wide adoption of AI, according to Nickle LaMoreaux, senior vice president and chief human resources officer at IBM, who took to LinkedIn to make her case. She sat down Monday with LinkedIn chief people officer Teuila Hanson in the social-media platform’s latest episode of Conversations with CHROs, and Inc. got an exclusive first look. The two discussed issues that are keeping HR up at night. LaMoreaux said she believes HR should take the reins on AI adoption because the department is an expert on both skills and culture change. “AI is about the technology, but it is about a lot more than that. It is about willingness to change how you lead people through the different roles of managers and leaders,” LaMoreaux said. Although many companies choose to give this responsibility to leaders who deal with new technologies—chief product officers, head of engineering, line of business owner, etc.—LaMoreaux says these professionals are good at adopting tech to complete job-related tasks, but they lack the skills to ensure company-wide adoption. Hanson points out that when HR is handling benefit enrollment or performance management, they’re always thinking about how these processes will affect employees at different stages of their careers: the applicant, new hire, employees who want to be promoted, managers, and leaders. They need to consider how team dynamics will be affected by these company-wide processes, she said. “We’re the employee experience function, and so it’s sort of natural for us to be in this space and really think through how the change is going to hold,” Hanson said. On the subject of change, LaMoreaux said AI created a major culture shift in how IBM measures employee performance. It started with using data to measure skills—ones employees had, new incoming skills, and skills that were becoming irrelevant. Pairing this data with employee business performance initially resulted in unusual results that frustrated some managers—for example, some top performers were rated as ineligible for promotions because their skills were out of date. LaMoreaux said it was a “really, really difficult” transition, but managers started “to see first-hand how quickly these jobs were changing,” and that someone “could go from being a top performer one year to an average performer or even a low performer the next year.” But IBM felt it still wasn’t addressing the full picture. “We started to realize, in this age of AI, even that wasn’t enough, that the world is just moving so quickly that we actually had to start evaluating on this idea of behaviors,” LaMoreaux said. IBM now measures employees’ capacity for entrepreneurial spirit, curiosity, and being OK with failure while trying new skills. IBM measures whether employees are “OK with failure” to determine their resiliency, LaMoreaux said. She added that IBM’s HR team was inspired by how the company’s product and development teams treat feedback as a problem solving tool, rather than a reflection of employee performance. “When you get negative feedback, it’s not a bad thing. It means you have this feedback and you can pivot. So we kind of really learned a lot from their culture and ways of working,” LaMoreaux said. While there’s a lot of public concern over AI’s potential to replace jobs, LaMoreaux says that using it to identify certain behaviors helps guide employees toward their professional goals. She says anyone who has “ambition” and can “pivot with the changes” will be part of creating new jobs that work with AI. BY KAYLA WEBSTER

An AI-Powered Startup Can Now Perform This Important Task Better Than Doctors

A healthtech startup has used OpenAI’s technology to develop a new, customized AI model that they claim can outperform real physicians at generating accurate medical codes and notes. Ambience Healthcare, a San Francisco-based company that provides healthcare workers with an AI-powered platform, says that its new model is specifically designed to identify ICD-10 codes, which are used to classify diseases and medical conditions for billing and record-keeping purposes. By recording a doctor-patient conversation, the AI model can intuit the conditions being discussed and generate notes with accurate codes. From corporate giants like Amazon, which recently launched AI tools for medical practitioners, to startups like Suki AI, businesses large and small are creating AI-powered products aimed at helping doctors and clinicians spend less time on administrative work, like writing and filing medical notes. According to a 2024 study published by Google Cloud and The Harris Poll, clinicians spend an average of nearly 28 hours per week on administrative tasks, and 94 percent of them say their administrative workloads prevent them from spending more time helping patients. In a press release, Ambience chief medical officer William H. Morris said that its platform can be understood “as a scribe that fluently speaks both clinical language and the intricate healthcare billing rulebook from day one.” In addition to helping clinicians finalize their notes faster, he claims, it will also ensure that “revenue cycle teams receive cleaner, more accurate, and audit-ready charts.” Ambience’s platform also integrates natively with several popular electronic health record providers, such as Epic and Oracle. This means that after clinicians review and sign off on notes made in the platform, they’re immediately added to the patient’s record. In February 2024, Ambience raised $70 million in a series B round co-led by OpenAI’s startup fund, and the two firms started collaborating. In partnership with OpenAI’s startup solutions team, Ambience created a dataset of complex clinical cases, all labeled with accurate ICD-10 codes. By fine-tuning an OpenAI model with this high-quality proprietary data, Ambience was able to create a powerful medical coding model. In a test that pitted Ambience’s new model against 18 board-certified physicians, the company said that “Ambience’s AI demonstrated a 27 percent relative improvement in coding performance compared to the expert physician baseline.” These coding improvements could help medical practitioners ensure that they always get paid for the work they perform, and mitigate the risk of billing errors. ‍ BY BEN SHERRY @BENLUCASSHERRY

Here Come the Androids: Millions of Humanoid Robots Could Solve the Factory Labor Crunch

Humanoid robots, a.k.a. androids, could achieve a massive leap in automating many jobs because they fit into the complex work environments we’ve carefully shaped over the years around human bodies and how they work. Add the unarguably cool Star Wars/C-3PO sci-fi angle, and it’s easy to understand why people have tried for years to make these machines a reality. Now a report in industry news site Automotive News shows how wage costs could spur automakers to send more of these robots marching onto factory floors to help fill out their production line capacities—something that’s been difficult due to higher labor costs and worker shortages. The report sites recent analysis by U.K. consultancy firm IDTechEx that predicts that some 1.6 million humanoid robots could be working in the automotive factory sector by 2035. U.S. Bureau of Labor Statistics data show that in the manufacturing sector as a whole in 2024, employers spend $45.29 per hour on average on workers. But conversely, Automotive News says venture capital firm UP.Partners predicts the “labor” cost for humanoid robots would be around $1.29 per hour. That equates to under 3 percent of the cost of human labor—on an hourly basis at least. This statistic alone is a compelling reason why carmakers will adopt robots, especially when recent research shows most Americans don’t want to work in factories, even though they approve of the idea of boosting U.S. manufacturing. As the market now stands, factory labor shortages aren’t likely to diminish. But there’s another reason carmakers may be on the front line of adopting humanoid robot workers: Because they can use their expertise to make and sell robots to other people. Automotive News quotes David Kehr, president of humanoid robotics at the Germany-based drivetrain maker Schaeffler, who explained “We’re a motion technology company, and what do humanoids do? They utilize motion. This is in our DNA.” This stance echoes many bold predictions made by Tesla CEO Elon Musk, who’s staked the future of his EV company on humanoid robots, rather than cars—and has promised to test Tesla’s Optimus robot on factory floors in very large numbers. Other reports highlighted the presence of Chinese-made humanoid robots at the recent Shanghai Auto Show, with numerous automakers talking about billion-dollar-scale investments, vowing robots will soon actually be put to work in factories. But robots are expected to mechanically stride into other industries as well, and perhaps sooner than the general public might think. Industry news site TheRobotReport recently noted that Houston-based humanoid robot and AI developer Persona AI had just raised a $27-million funding round to help it develop robots to work on shipyards. The company said its machines would marry the high-precision actions of typical industrial robots with the kind of subtle dexterity that humans possess, making the robots suitable for the rugged and demanding environments of manufacturing ships. It also highlighted worker safety in its plans—essentially moving fragile humans from the riskiest environments. The company has already signed an agreement with HD Hyundai, the Korea-based manufacturer that holds about 10 percent of the global ship-building market share. It expects to be ready to sell humanoid robots for shipyard work in about 18 months. Meanwhile, speaking at the Computex 2025 trade show (which markets itself as the “World’s Largest AI Exhibition”) this week, Nvidia CEO Jensen Huang also painted a bold picture of humans working alongside robots. Describing data centers of the future as “AI factories”—because AI is now infrastructure “just like the internet, just like electricity”—Huang spoke of a future where people worked alongside AI agents and robots and machines fill job gaps left vacant by people. Why should you care about this? For one simple reason: Even if your company isn’t directly involved in the manufacturing industry, what’s happening there is a litmus test for almost any workplace in the future. Carmaking is already highly robotic, so the industry is ready to adopt the next-generation of robots in humanoid form, and once they’ve demonstrated their value in these scenarios, they’ll be adapted for many other jobs. If you’re building a long-term business plan for your company, you might want to think about how robots could help. BY KIT EATON @KITEATON