Daybreak - After Deloitte's blunder, the Big 4 may learn a new AI rule: being wrong first beats being right late

Episode Date: October 8, 2025

When Deloitte refunded part of the A$439,000 it was paid by the Australian government for a report riddled with AI-generated errors, it seemed like the perfect moment to slow down. Instead, the firm doubled down and announced a global rollout of Anthropic’s Claude to nearly half a million employees. That decision captures the strange new logic shaping the Big 4 consulting companies. PwC, EY, KPMG, and Deloitte are no longer just using AI, they are performing it.Audit and tax work has slowed, regulation is tightening, and growth now depends on signalling technological boldness. In this new credibility economy, hesitation looks worse than failure. A mistake is no longer a crisis; it has become proof that you are early. But as every firm rushes to prove its AI edge, sameness is setting in, and the next real differentiator may not be accuracy at all.What could it be then?Tune in.

Hi, this is Rohan Dharma Kumar. If you've heard any of the Ken's podcasts, you've probably heard me, my interruptions, my analogies, and my contrarian takes on most topics. And you might rightly be wondering why am I interrupting this episode too. It's for a special announcement. For the last few months, I and Sita Raman Ganeshan, my colleague and the Ken's deputy editor, have been working on an ambitious new podcast. It's called Intermission.

We want to tell the secret sauce stories of India's greatest companies. Stories of how they were born, how they fought to survive, how they build their organizations and culture, how they manage to innovate and thrive over decades, and most importantly, how they're poised today. To do that, Sita and I have been reading books, poring over reports, going through financial statements, digging up archives, and talking to dozens of people. And if that wasn't enough, we also decided to throw in video into the mix. Yes, you heard that right. Intermission has also had to find its footing in the world of multi-camera shoots in professional studios, laborious editing, and extensive post-production.

Sita and I are still reeling from the intensity of our first studio recording. Intermission launches on March 23rd. To get alert, as soon as we release our first video. episode, please follow intermission on Spotify and Apple Podcast or subscribe to the Ken's YouTube channel. You can find all of the links at the ken.com slash I am. With that, back to your episode. It all began as corporate embarrassments often do, you know, with a quiet correction and a partial refund. Earlier this year, Deloitte was paid $440,000 Australian dollars by the government of Australia to produce an independent assurance report on the country's welfare compliance system.

But the final report came back with some problems, not disagreements over policy or methods, but something stranger. You see, parts of that documents cited non-existent academic papers included fabricated court quotes and featured footnotes that led to nowhere. The culprit? As you might have guessed, generative AI. Deloitte later confirmed that its team had used an AI tool to help draft sections of that report and that the model had indeed, in the polite phrasing of bureaucrats, introduced factual errors. So, earlier this week, Deloitte updated the report and agreed to refund a part of the fee that the Australian government had given it. Now, for most companies, this might have been a pause moment, you know, a chance to be.

to slow down, retrain themes, recheck policies and so on. But not for Deloitte. Because within days of that refund, the company announced its largest ever technology rollout. Anthropics Claude, an AI chatbot, would now be available to nearly half a million Deloitte employees worldwide. So you will probably understand when I say that the timing of all of this feels almost defiant,

as if the public embarrassment had merely cleared the path for more AI, not less. And this is what makes the story unusual. Because Deloitte is not just any other firm. It is one of the world's biggest consulting companies, a company whose entire business model rests on trust. It is the institution you call to make sure you haven't made a mistake, you know? It feels like a remarkable moment of convergence. The big four firms that spent near,

early a century selling their judgment are now suddenly competing on speed, because they cannot differentiate on price or process anymore, so they differentiate on whose algorithm looks smarter. For example, PWC has struck a deal with OpenAI and Microsoft, becoming one of OpenAI's first enterprise resellers and one of the largest corporate users of Azure. EY has launched EY.I.I.I. It's in-house AI platform backed by more than a billion dollars in investment, and it is also now building sector-specific agentic systems with Nvidia. KPMG, meanwhile, has gone all in with Microsoft's Azure AI, creating its own multi-agent automation platform called Workbench. So AI is not just another

consulting tool anymore. It is becoming the core of what consulting is. And that is the irony. The companies that build global empires on caution and credibility are now reorganizing themselves around a technology that is known to make things up. Welcome to Daybreak, a business podcast from the Ken. I'm your host, Nick Dha Sharma, and I don't chase the news cycle. Instead, every day of the week, my colleague Rachel Varghees and I will come to you with one business story that is worth understanding and worth your time. Today is Thursday, the 9th of October. What does this reorganization actually look like? Because it is one thing to talk about consulting's big AI pivot,

but another to see how it's quietly reshaping the mechanics of the business itself. Inside the big four, that shift is everywhere. The traditional engines, which was auditing and taxes, have slowed down. Regulations are tighter, automation is squeezing margins, and clients have become reluctant to pay a premium for predictability. The dependable work still pays the bills, but it no longer delivers growth. And that is why transformation has become the new consulting product. Digital overhaul, strategy resets, AI roadmaps, projects that can be pitched as forward-looking,

not backward auditing. Nothing embodies that promise better right now than artificial intelligence. But here's the irony. For most of these firms, doing a bit of those firms doing. AI is not just about dramatic efficiency gains. It is about optics. Clients want proof that their advisors are living the future that they're selling. As one consultant told the financial times, clients want to see that we use the tech that we're selling. That expectation has flipped, and it is not enough to know AI anymore. You have to perform it. And this also seeps inwards

into these companies, AI becomes a kind of an internal status marker. You could say the corporate equivalent of a new suit. Partners compete to sponsor pilot projects, practice heads jostle for ownership of firms' AI centers of excellence. You know, the slides almost right themselves. Responsible innovation, augmented expertise, trust in transformation. Across the industry, the choreography is strikingly similar. Let me give you a a sense of how. PWC, like I told you, has become Open AI's largest enterprise customer and first reseller of ChartGPT Enterprise. It is integrating it across its U.S. and UK workforces and also selling it to clients. It is all a part of the $1 billion AI investment plan.

EY has poured over a billion dollars into its own AI platform, EY.comi.aI, which is meant to fuse automation and advisory work. KPMG has aligned tightly with Microsoft's Azure AI, building internal multi-agent tools under its workbench platform. And you already know about Deloitte's deal with Anthropic. So you see, every firm now has a model partner. It has internal AI teams, pilot dashboards, and slides about responsible AI. But the real differentiator is not model accuracy.

It is who can launch the fastest, loudest, and. and with the least appearance of hesitation. And all of this posturing has created a new kind of an arms race, one where credibility, not competence, will decide who wins. And that race is just beginning. Stay tuned. There was a time when a consulting firm admitting to factual errors would have been a small corporate earthquake.

Clearly, not anymore. Because when every major consultancy is shouting, we are AI powered, the safest place to stand is somewhere in that motion. But here is the other side of that same coin. When every firm shouts, we are AI powered, the landscape also flattens. Everyone uses the same cloud platforms, the same third-party models, and the same disclaimers about human-in-the-loop checks. The competing narratives then begin to converge and become nearly indistinguishable from one another. But the sameness does not dull competition. It in fact intensifies it. Because once everyone is working with the same tools, the race shifts from who

uses AI best to who looks like they use it best. Deloitte's refund in that sense becomes a useful mistake. A small, costly reminder that visibility often matters more than precision. Like Deloitte's embarrassment in Australia did not derail its global AI rollout. If anything, it seemed to validate it. So clearly, in this new credibility economy, if I may call it that, being wrong first is safer than being right late. So going by the same logic then, the next phase of this race will not be about speed at all. It will be about governance. Who can prove that they are using AI responsibly?

regulators are catching up. The European Union's AI Act will soon require disclosure of generative AI use in high-risk professional services, while the UK and Australia are drafting parallel guidance for algorithmic accountability. Clients, especially in the government and finance, are beginning to demand audit trails, which model generated this analysis, who verified it, how was it stored? So the next differentiator will not be accuracy, it will be traceability. Firms will compete on compliance tax, documentation and explainability, a return in a strange way to their auditing rules. Because after a year of racing to adopt AI, the big four may discover what clients now want

is not speed, it is proof. And that is the irony that closes this loop. The same companies that once solds. certainty through human judgment, now have to rebuild it, line by line, log by log, inside the machines that they have chosen to trust. Daybreak is produced from the newsroom of the Ken, India's first subscriber-focused business news platform. What you're listening to is just a small sample of a subscriber-only offerings and a full subscription offers daily, long-form feature stories, newsletters and a whole bunch

of premium podcasts. To subscribe, head to the Ken.com and click on the red subscribe button. on the top of the website. Today's episode was hosted and produced by my colleagues Niktha Sharma and edited by Rajiv CN.