Latest / MBA Training AI / AI spend per employee: why the falling metric misleads

Transcript

AI spend per employee: why the falling metric misleads · MBA Training AI

• 0:00
  Host: Welcome to the Leaders Insights Podcast. Today's episode AI spend per employee is falling.
• 0:11
  Host: Efficiency win or adoption stall? Enterprise AI spend per employee dropped in August 2026.
• 0:17
  Host: Everyone's calling it either the efficiency dividend or the beginning of the crash. You
• 0:21
  Host: think both camps are wrong? Convince me.
• 0:24
  Expert: They're both looking at a ratio and pretending it's a story. Spend per employee is spend
• 0:29
  Expert: divided by headcount. If either number moves, the ratio moves. And right now, both are moving
• 0:34
  Host: at once. So the number went down and told you almost nothing.
• 0:38
  Host: Break it apart then. What's actually happening underneath?
• 0:41
  Expert: Two things. First, prices collapsed. The cost per million tokens, a token being roughly three
• 0:48
  Expert: quarters of a word the model reads or writes, has fallen by more than half at the frontier
• 0:52
  Expert: labs since early 2025. So you can do the same work for less money. That's not a slowdown.
• 0:59
  Expert: That's deflation. And the second thing?
• 1:02
  Expert: Headcount grew at the firms doing the most AI. When your denominator, the number of employees,
• 1:08
  Expert: swells because you're hiring. Spend per person drops even if total spend climbs. The top
• 1:13
  Expert: quartile of adopters increased absolute AI budgets in August. The per head figure fell
• 1:19
  Expert: anyway. It's an arithmetic mirage in…
• 1:22
  Host: Google AI put out numbers suggesting productivity gains are driving the efficiency. You buy that?
• 1:28
  Expert: Google AI says teams using their tool ship features around 30% faster. But they sell
• 1:33
  Expert: the models, so that figure has a thumb on the scale. Cross-check it against O'Reilly
• 1:38
  Expert: Radar, which surveys practitioners who don't have anything to sell you. O'Reilly's
• 1:44
  Expert: read is messier. Real gains in coding and drafting. Near zero gains in workflows, nobody
• 1:50
  Expert: bothered to redesign. The tool got cheaper. The habits didn't change.
• 1:55
  Host: So the falling number could be hiding a genuine adoption stall in some places.
• 1:59
  Expert: Exactly. And that's the part both camps miss. Averages lie. If 10 teams cut spend
• 2:05
  Expert: because they gave up quietly, and 10 teams cut spend because they got ruthlessly efficient,
• 2:10
  Expert: the average looks identical. Same number, opposite meaning. You have to look at usage,
• 2:15
  Expert: not dollars.
• 2:16
  Host: How does a manager tell the difference from inside their own company?
• 2:19
  Expert: Look at active seats versus paid seats. If you bought 500 licenses and 120 people log
• 2:25
  Expert: in weekly, your spend per employee looks lean and healthy because 380 people abandoned the
• 2:31
  Expert: thing. That's the stall wearing efficiency's clothing. Cheap because unused is not a win.
• 2:37
  Host: That's the uncomfortable version. Give me a company example that's real.
• 2:41
  Expert: Klarna made noise in 2024 about replacing customer service headcount with AI, then
• 2:47
  Expert: quietly rehired humans when quality slipped. The lesson wasn't AI failed. It was that
• 2:53
  Expert: they cut cost before they fixed the workflow. So the savings evaporated the moment customers
• 2:58
  Expert: complained. Cost per employee looked brilliant for about two quarters. Then it didn't.
• 3:04
  Host: So what does a smart spend per employee drop actually look like?
• 3:08
  Expert: It looks boring. Total spend flatter up, usage up, and the per head number falling
• 3:14
  Expert: because the price per unit of work dropped and people are using more of it, not less.
• 3:20
  Expert: Rising volume, falling unit cost. That's a company eating the deflation and putting
• 3:24
  Expert: it to work.
• 3:26
  Host: And the dangerous drop?
• 3:27
  Expert: Spend down, usage down, and someone in finance calling it discipline. That's not discipline.
• 3:34
  Expert: That's a pilot project dying in a spreadsheet where nobody has to watch it.
• 3:38
  Host: Give me the dry version of that.
• 3:40
  Expert: A dead subscription is the most cost efficient subscription you'll ever own. Zero usage,
• 3:45
  Expert: predictable bill, terrible outcome.
• 3:48
  Host: So the headline metric is basically useless.
• 3:51
  Expert: It's worse than useless because it's confident. It gives aboard a single tidy number that
• 3:56
  Expert: feels like insight and hides whether your investment is compounding or rotting. If
• 4:01
  Expert: you only track dollars per head, you've optimized for the one thing that can't
• 4:05
  Expert: tell you if the strategy works.
• 4:08
  Host: And what should someone track instead starting tomorrow?
• 4:11
  Expert: One number. Weekly active users divided by paid licenses per team tracked over time.
• 4:18
  Expert: If that ratio climbs while your bill holds steady, you're winning on efficiency.
• 4:23
  Expert: If the ratio falls, you're not saving money. You're abandoning the tool in slow motion.
• 4:28
  Expert: And the falling spend is the receipt.
• 4:30
  Host: Usage over seats, not dollars over heads. That's the takeaway.
• 4:34
  Expert: That's the whole game.
• 4:35
  Host: Sources for today's episode, the decoder, Google AI vendor, AI lab, TechCrunch AI,
• 4:42
  Host: K nuggets, ours, Technica AI, O'Reilly radar.
• 4:46
  Host: That's your five. The full AI track, structure, no fluff, is at MBA-training.com.