It is not about AI!
It is about having an excuse to fire people they don't like, i.e. those that question the mental state of the manglement.
It also pleases shareholders, apparently.
You'll use AI and like it too – if you work for PwC. Paul Griggs, US chief executive of the global professional services giant, has made clear there is no room at the corporation for AI skeptics. Speaking to the Financial Times, Griggs indicated that anyone who believed they had the "opportunity to opt out" of AI is "not going …
That, to me, is one of the critical AI questions.
If AI can do the coding, and AO can do the work of the management consultant, then why do I need the management consultancy at all?
Claude can very definitely code, yes it can. Denying that is like denying that night follows day. But, can it code to a sufficiently accurate level to replace each grade of engineer cost effectively while consuming both tokens and a "pilot" to give it requirements?
For that to be true we either need to move engineering closer to the business where they subsumed most of the product organisation, or chop engineering and massively skill up the product organisation. Otherwise we have all the humans and now tokens to buy too..
If the AI is going to tell the consultant what to tell me which way to go, why can I not just ask the AI and burn the consultancy?
Bye Bye McKinsey, PwC etc all, time to reap what you have sown for decades, it seems.
> Kudos to them for "eating their own dogfood."
I read this more like he's forcing his workforce to eat the crap some AI salespeople sold to him...
To quote Cory Doctorow: "AI can’t do your job, but an AI salesman can convince your boss to fire you and replace you with a chatbot that can’t do your job"
https://doctorow.medium.com/https-pluralistic-net-2025-03-18-asbestos-in-the-walls-government-by-spicy-autocomplete-ff437603809c
I mean, why would he continue to sell accountants per hour, a model that scales horribly, as one needs to add more and/or more expensive people to raise revenue and makes him somewhat responsible for results.
Instead, he can sell AI slop by subscription, a model that scales basically indefinitely and avoids responsibility.
Time will tell if customers accept AI generated garbage they can also get from every other player in this crowded market - at probably lower cost compared to PwC...
My grandson is a forensic accountant. A lot of his job is intuition. Seeing the woods for the trees, deciding where next to look. His boss (he works for a Wall St company) is looking at using an AI to replace a lot of his work. He says bring it on because he knows that a LLM set of rules can't match the deviousness of someone who is determined to hide millions and millions often almost in plain sight.
I'm sure that the likes of PWC C-Level execs will be laughing all the way to the offshore bank account with this move leaving the grunts that actually do the work high and dry.
That's exactly the point.
You need much experience to somewhat handle AI and correct all garbage generated in the output.
Once you eliminate experienced staff in your organization, you basically eliminate the capability to actually use AI at all.
Once you eliminate Junior staff, you eliminate the future experienced staff.
The only winning move is not to play ...
I can see the model for some types of activity (eg tax advice) where the customer is likely to ask the advisors for advice (and there is a large body of previous "knowledge") but consultancy where the consultant is expected to keep their eyes open and sell more consultancy (needed or not) it might not work as well.
"I can see the model for some types of activity (eg tax advice) where the customer is likely to ask the advisors for advice (and there is a large body of previous "knowledge") but consultancy where the consultant is expected to keep their eyes open and sell more consultancy (needed or not) it might not work as well."
In the US, the tax code is so complex that the IRS will not even guarantee that the advice their staff gives is valid and the taxpayer will be on the hook if it's found later that there was an issue with that advice. It could work, but only if the generated responses include references so one can read the official language. That would very helpful as it could find passages and story lines that might be an enormous slog to find otherwise for a particular situation.
From experience with some ex-Accenture people what will happen is those humans who are good at their job but get fired due to AI aversion but have good relationships with a portfolio of clients will form another company, let's call it CwP, and nick a lot of the customer base who are not wanting to trust ClippyPilot with their super important, legally required and deeply informative of their own investor happiness documentation.
Nose. Face. Spite.
" but have good relationships with a portfolio of clients will form another company, let's call it CwP"
Or they move to another company and the clients decide to follow the person rather than stay with 'the company".
If you run a company like a vending machine, your customers will treat you like a vending machine. Any Coke in a storm. I'm constantly bombarded with companies that want to sell me a "complete customer experience" package that has my clients interacting with a web site rather than me personally. If I did that, I'd be a machine and there would be no personal connection so no impediment to their going elsewhere and shopping my services and prices at the drop of a coin.
It probably doesn’t even need to be that sophisticated.
It could be a protected belief ? My philosophical belief is that AI/Robotics will be to the ruination of mankind. Plenty of source material for that - fact or fiction. ‘Merican’s love bonkers beliefs too - see current HHS Head.
“Under the Equality Act 2010 in the UK, "religion or belief" is a protected characteristic, meaning you cannot be discriminated against, harassed, or victimized because of your religious or philosophical beliefs”.
Here in Sweden, the government just had to withdraw a report from the Social department because someone who actually read the report discovered that the report's references had been completely made up. In light of that, the report's conclusions could hardly be trusted.
Will using a LLM save you time? Only if you don't double check every word it has written for hallucinations.
And the Lord spaketh saying
"Praise AI for AI will make us money without providing benefit to the customer"
"but the Government will bestow us with gifts and grants to build datacenters upon the lands."
And then he whispered
"and lo, we will be paid a second time to use a real person to fix the problems the AI leaves behind."
If I were required to use "AI" for my work, my first and only instruction to it would be for it to make it look like I'm using it enough to satisfy the AI usage requirement and then get on with my actual work.
Not sure how the higher-ups could complain about that; I'm just using AI to automate tiresome, pointless, and irksome tasks like using AI, for example.
Malicious compliance is still compliance, right?
If they are simply going to use AI, why bother to pay them to do so? If you want slop, you can use AI yourself. You hire a company for their skills and experience. There is no need to hire a company that simply embraces AI. It's like paying a travel agent to book you an AirBnB or paying someone to order you a book that is on Amazon.
There's a bit more to it than that.
See, most Big Wheel Executives already have plenty of folks to tell them what they want to hear -- they brought in their own gaggle of yes-people and lickspittles from the last place they fleeced, after all.
What the consultancies bring is the graft cycle; e.g. big sporting event box seat tickets, nepotism hires, favoritism favors, and so on, in exchange for a big fat consulting fee. At the end, consultancies also give the Big Bosses the standard consultancy playbook for cost-cutting, layoffs, etc. to juice the company stock price and thereby pump up the Executive Bonuses.
Then when the C*O types eventually finish pillaging and move on to the next place, they bring in the same consultant vultures from last time and start again.
The Executive-Consultant relationship isn't a client-customer sort of deal; it's more like co-conspirators.
"when we can just have Gemini do that for us for free?"
One would hope that what's done at PwC stays at PwC and if it isn't, there's a good chance of making them pay for sloppy data security. With Gemini, it will be freely available on the Dark Web at some point and no way to assign any blame. It also isn't likely to do the sort of analysis and reporting a large company using an outside accounting company would want/need. The sorts of things it might come up with are the sort of things that should be able to be seen through internal analysis.
With free, you get what you pay for.
AI cannot fail. Because if it does, it will drive many companies into ruin.
This is the reason why these C-level executives are so determined to maintain the highest level of excitement. "Use AI or quit"; "those who don't use AI will be out of the job market". It's easy to want to make a product mandatory when people have to pay you dozens of dollars per month for it.
The height of greed is wanting to make a product mandatory and forcing people to pay you dozens of dollars for it to a handful of Big Tech companies.
.... that it doesn't work/has issues/takes too many digital drugs, the people that have been sacked will have moved on to positions at other companies and won't be coming back.
It would have been a better move to have parallel operations in the basement replicating work being done by humans and comparing it with what an AI systems comes up with. Maybe the AI is "just as good" for a lower cost, but when it isn't, it could cause all sorts of lawsuits and regulatory inquiries. The magnitude of the problems when it isn't perfect can be important to look at. Some attorneys have lost their licenses for submitting AI generated documents without checking them first. I'm sure they saved loads of time and staff costs, but that one time........ Their new position will have them remembering to ask, "Do you want fries with that?"
"The unfair dismissal lawsuits incoming from here!"
The usual thing is for those companies to offer some compensation if the employee signs a waiver. I've had that happen before. It was insurance against unfair dismissal lawsuits as the cost to the company was far less than a day's engagement of a law firm.
A great way to get rid of anyone not a yes man and/or an empty suit. Except of course the AI true believers - yes men or not, this sorting does not apply to most of them...
Although the more competent of those who actually like AI may start to look for another job after idiotic policies like this are revealed, no matter if they fit their preferences or not. They can see how this is all going to go horribly wrong and get out of the way before the management stupidity is blamed on AI tools and evangelists ("we were fooled by our AI staff and have sacrificed 1200 of them to Mammon with a rusty knife, may the stock options go up again"")
"Staff at Accenture received a memo last month telling them to demonstrate "regular adoption" of AI services – with usage tracked "
Side note, I have a gig with the other half of ACN: Avanade.
My ACN "boss" made a comment about some data that needed analysing, lots of log files with nested JSON to unpack and redact stuff etc. In true PHB form, he wanted me to "download the logs and use co-pilot to analyse it". I guess he wants a promo!
/anon 'cos well.. obvs.
Company got PwC to run an audit on our disaster recovery processes. All they could come up with as an 'issue' was that our documentation was across half a dozen documents - all in the same location, each one about a specific aspect of the process, and all backed up in a least three different locations on- and offsite. Their 'recommendation' was that we *had* to put it all into one document.
Of course, when directors see an audit recommendation they go batshit crazy about having to fulfil it - rather than the more appropriate response of telling the auditors to piss off and stop being knobs.
"All they could come up with as an 'issue' was that our documentation was across half a dozen documents - all in the same location"
PwC felt they had to come up with something to justify their fee without considering that others can get things right. A bit of confirmation that another few sets of eyeballs thought it all looked good would have been fine. I expect there was some reason why the documentation was split into sections. I've seen all-encompassing documentation that is hard to get through as so much is not applicable. I keep my resumé in long form with all of my work experience, but I would trim it down to what's relevant for a particular job with maybe a one sentence divider to explain gaps so it wouldn't be thought I was smashing rocks with a sledge during those gaps. Education and other information is the same across any version I'd distribute. It's not like I wouldn't tell a company I hold a commercial radio operators license for some reason. Maybe that's not part of the position, but who knows if they might consider me for a better position that I don't know about.
Sounds like another Monday-morning epiphany from the top after skimming an AI hype piece. Meanwhile, the rest of us are left to actually make ‘AI-first’ work in client deliverables without enough training, clear guidelines, or proof it moves the needle beyond pilots. Good luck replacing the 'not paranoid enough' partners with people who magically figure it out overnight.
This reminds me of the increasingly desperate attempts by the Johnson government to identify the mythical Brexit dividends. Every government department was ordered to look down the back of the sofa for some benefit from Brexit.
My feeling is that if something is useful, people will use it of their own accord. Imagine if your company had to force you to use a smartphone on pain of being fired. You might conclude there was a problem with the product.
Smoke and Mirrors.
Classic case of a meeting where some body said, "using AI will save us millions in salaries and benefits". Yep bigger bonuses for the top dogs.
...and they can get rid of a few partners and staff who's faces don't fit as well, but that will be hidden under "restructuring the business to align with the new AI work flow model".
But the real kicker is the people forcing AI on the business/staff are the very people who will not be actually using AI on a day to day basis. No, because they don't actually know anything about AI and how it works or how it will actually work with in the business.