Almost half of the supply chain leaders Gartner surveyed have cut an average of 10.5% of their headcount in two years, and 66% of them say the cut fell on entry-level roles [1]. Days later, Kinaxis reported subscription revenue up 22% and $1,026.9 million of contracted work it has not billed yet [2]. Divide that backlog by the $483.9 million of yearly recurring revenue behind it and the average deal works out to roughly 2.1 years. Your silent 2.1-year contract is that cut, because the grade that left first is the one that learns when to overrule a forecast.

Two lines on one budget

A sales and operations planning (S&OP) upgrade is sold on forecast accuracy, and accuracy is the hardest number in the building to disprove. Nobody in a monthly review can show that the model would have missed the quarter. Payroll works the other way. Finance already audits it, so a team that sheds a tenth of its seats books a saving in the same period.

Gartner’s survey is narrower than the headline that travelled with it. The 47% covers 303 leaders at companies that already use AI, questioned between 5 May and 11 June [1]. The cut is one they credit mainly to AI, rather than an effect the study can isolate [1]. Where it landed is the part that matters: entry level, at more than double the rate of any other tier [1].

The firm’s own analyst draws the conclusion a vendor deck would leave out. Tess Frenzel, a director analyst in Gartner’s supply chain practice, says AI “is thinning the entry-level supply chain workforce faster than most chief supply chain officers are redesigning the talent pipeline”, a gap she expects to weaken the AI results those firms are buying [1]. A planning platform does not police itself at the exception layer, where a forecast is wrong for a reason no column holds.

The strongest respondents are already acting as if they know it. The top quarter by AI integration and business results is “elevating entry-level roles”, asking for more analytics skill and handing junior staff “less repetitive work and more consequential decisions” [1]. That is the opposite of a headcount target, and it costs money in the year it is made.

Kinaxis’s quarter is the other half of the same deal. Subscription revenue reached $112.3 million, total revenue $160.3 million, and the contracted backlog grew 21% on the year [2]. Professional services, the installation and integration work a buyer cannot do without, grew 14% to $42.2 million [2]. The company raised its full-year guidance, citing three consecutive quarters of outperformance [2].

What I keep circling is the timing. Put the contract beside the cut and the two become one story. That two-year cut adds up to about 5.4% a year, so the planning team shrinks at the same pace the deal runs on. Fewer people own more software for longer, and the platform will still bill when the bench that was meant to learn it has to be rebuilt.

By the numbers

  • 10.5% — Headcount cut: the two-year drop reported by 47% of the leaders surveyed, with 66% saying it landed at entry level [1].
  • 2.1 years — Implied contract term: $1,026.9 million of contracted work divided by $483.9 million of recurring revenue; Kinaxis publishes no average term, so this is a sum rather than a figure from the filing [2].
  • 70% — Manual planning effort removed: the average across 15 measured sites, all running the reporting vendor’s own engine, with 75% of forecast volume touchless at Mars [3].

What the seat line hides

The work that goes first is the work that teaches. Across the 15 sites in DemandForecast.ai’s report, manual planning effort fell 70% and planners were left reviewing the quarter of the volume where their judgment changes the outcome [3]. Exception review is a skill built by doing the dull version of it for a year or two.

  • Where accuracy is promised: a business case that assumes a few points of improvement should name who owns the overrides in year two, because that person is the input.
  • What the term implies: a long deal is a fair bet on a stable bench and a poor one on a team still cutting, since a rebuild takes longer than the early years of any contract.
  • How services grow: installation work up 14% while internal staff falls moves the integration skill onto a vendor invoice, a cost the seat saving never shows.

My read is that the accuracy fight is being won on the wrong document. The targets in an S&OP pack are written by the people who remain, and those people are now more senior. Whether a junior planner can still be trained to challenge a number is a staffing question, and it sits in the workforce plan rather than the technology plan (Planning AI Buys Capability, Not Autonomy).

The catch

Three limits hold this argument down. Kinaxis is one vendor reporting its own subscription line, so the growth describes a company and not a market. The term figure is a ratio of two published numbers rather than something the firm discloses. And Gartner’s 10.5% is an attribution by the leaders who answered, not a measured effect of AI on a planning team.

The fair counterweight is that the software may deserve the cut. A large drop in manual planning effort is real work gone, and a smaller team with a working model can beat a larger one still fighting spreadsheets. What I find hard to square is the order: the saving arrives in year one, while the cost of a thin bench arrives when the model meets a quarter nobody modelled.

Freight is the pressure behind all of it. Shipper spending rose more than 30% in the third quarter on only slight shipment growth, which is cost inflation arriving without demand behind it [4]. A planning budget under that pressure loses its seat line before it loses its platform.

At a glance

  • The shift: Planning platforms are now funded out of the planner payroll, and the entry level is where the cut lands.
  • The number: 10.5% of headcount gone in two years, against a contract that runs longer than the cut.
  • The risk: The bench that learns to overrule the model is cut first.

Related reading

Sources

[1] Gartner, “Gartner Survey Shows 47% of Supply Chains Reduced Headcount by an Average of 10.5% Over the Past Two Years Primarily Due to the Impact of AI”, 2026-10-28 — https://www.gartner.com/en/newsroom/press-releases/2026-10-28-gartner-survey-shows-47-percent-of-supply-chains-reduced-headcount-by-an-average-of-10-5-percent-over-the-past-two-years-primarily-due-to-the-impact-of-ai [2] Kinaxis Inc., “Kinaxis Inc. Reports Third Quarter 2026 Financial Results”, 2026-11-04 — https://investors.kinaxis.com/news-releases/news-release-details/2026/Kinaxis-Inc–Reports-Third-Quarter-2026-Financial-Results/default.aspx [3] DemandForecast.ai, “The Demand Forecasting Accuracy Report 2026”, October 2026 — https://demandforecast.ai/resources/forecast-accuracy-report [4] Supply Chain Now, “Sneak Preview: Q3 2026 U.S. Bank Freight Payment Index”, 2026-11-04 — https://supplychainnow.com/sneak-preview-q3-2026-us-bank-freight-payment-index-1656