Accenture publishes its full-year results today. If your firm bills by the hour, pay attention... Accenture is the biggest firm in the world that bills by the hour, and in June the market gave its verdict on that model in a single day.
On 18 June the company lowered the top of its revenue guidance, from 3 to 5% growth to 3 to 4%. Bookings were solid, margins went up, earnings per share rose 9%. The shares fell 18% anyway, the biggest one-day fall in Accenture's history, to less than half their high of the past year. EPAM and Cognizant, two of its largest listed rivals, had each lost half their value or more from their January highs by the summer, and India's Nifty IT index was down 29% for the year by midsummer.
A one-point trim to the top of a guidance range doesn't cause that on its own. Investors were worried about the whole industry, for a simple reason. These companies sell people's time, and AI is making the time shorter.
The most successful services business ever built
Accenture is the most successful professional services business there has ever been, and it got there by industrialising enterprise transformation. Almost 800,000 people, huge delivery centres in India and the Philippines, multi-year programmes staffed by pyramids of analysts and consultants, all priced on the hours and the heads involved.
Every technology wave before this one made Accenture bigger. ERP, the internet, cloud, mobile... each created more systems to build, migrate, integrate and maintain, and more people to bill for the work. Accenture's revenue today is roughly seven times what it was in 2000, and that is why.
AI is the first wave that cuts the human effort inside the work itself. Clients may well want more transformation than ever. They will need fewer hours of Accenture to get it. And when revenue is a function of hours, cheaper hours mean less revenue, unless the price changes with them.
Julie Sweet, Accenture's chief executive, said on the June earnings call that "AI will be a tailwind for us and our industry as it scales." I think she is right, for the firms that work out how to charge for results instead of effort. The 18% fall in Accenture's shares that same day was investors doubting that an 800,000-person company can make that change fast enough.
The number Accenture stopped publishing
Until December, Accenture reported its AI bookings and revenue separately. The last figures it gave were $2.2 billion of AI bookings and about $1.1 billion of AI revenue in a quarter where total revenue was $18.7 billion. Then management said it would stop breaking AI out, because AI was now embedded across its bigger programmes.
I understand the reasoning. But that was the one number that answered the question investors care about, which is how much new revenue arrives for every dollar of old work that AI removes. Stop publishing it and people assume the answer is ugly. Judging by the share price, that is what they assumed.
Others in the industry are more direct about it. The chairman of TCS told shareholders in June that within three years the company expects to have as many AI agents as human employees, and acknowledged, for the first time, that TCS will hire fewer people than it used to. Every agent that does the work of a billable person is a person you no longer invoice for, so that is as much a statement about revenue as about technology. The pyramid is being shortened from the bottom, the same way it is in law and accounting, for the same reason.
Who is doing it differently
While Accenture defends its model, two kinds of competitors have built a different one.
First, the consultancies that moved early. McKinsey says about a third of its revenue now comes from underwriting client outcomes rather than billing for advice, and it runs something like 25,000 agents alongside 40,000 people. That only works in a business paid for results. In a business paid for hours, every agent you deploy shrinks the bill.
Second, the labs. Anthropic co-founded Ode, a services firm, with Blackstone and Goldman. OpenAI has its own services venture and now puts its engineers inside client companies to build custom tools. That is Accenture's core business, done by the company that makes the model, with the model maker's margins behind it.
Accenture still has things neither of them have. Relationships so deeply embedded in the world's largest companies that moving away from it is painful, a record 129 quarterly bookings above $100 million last year, and a long track record of making technology actually work inside organisations where change is hard. That is a strong position. It is also the same position a very good hourly law firm holds, and it gets worth less every quarter the pricing stays the same.
What I'll be looking at in the results
Not the revenue line. It will grow 3 or 4% and people will argue about whether that is a slowdown or a floor.
Headcount against revenue. If Accenture grows revenue while its workforce shrinks, the change has started, and revenue per employee is the number to track from here. If headcount grows alongside revenue, nothing has changed yet.
Any mention of pricing on outcomes. If Accenture starts talking about contracts priced on results rather than on teams, that is the real shift. The investor day on 14 October is where it would come up, if it comes up at all. The competition from the labs won't appear in a guidance cut either. It will appear in who wins the next big programme.
I don't believe Accenture is going anywhere, whatever today's numbers say. But the question behind the June sell-off is the right one, and it applies to every firm that sells time. Can you earn more as your clients need fewer of your people? Accenture is the biggest company in the world that has to answer it.
How long does the billable hour have left? Hit reply, I read everything, and the sharpest replies shape the next edition.
See you next time.