Insource, the Tokyo Prime-listed corporate trainer, has given investors its own account of why it cut guidance. In a letter from its president, the company says it moved too early on advanced AI and put its sales staff in the wrong place.
On September 24, 2026, Insource lowered its outlook for the year to September 2026: sales from ¥16.0bn to ¥15.6bn, operating profit from ¥6.38bn to ¥6.15bn. It still expects growth over the prior year, but below its original plan.
Where demand fell short
The president says large companies bought heavily for company-wide basic generative-AI training, and that this wave ran its course in the year. The next stage, advanced AI use and AI agents, spread more slowly than Insource had assumed, while demand for basics such as Excel stayed strong.
The letter names three drags in the fourth quarter: weak wins in instructor-dispatch work from large companies, slower growth in big DX and tiered-training services, and slower uptake of its talent-development smart pack among large clients. Demand from mid-sized and smaller companies grew. The company adds that it raised sales headcount without a matching rise in profit.
What changes next year
Insource splits clients by size: 2,000 or more employees, 300 to 1,999, and 299 or fewer. For the largest, it says it will stop chasing deals with many salespeople and put content-development specialists in front of complex projects. Mid-sized and smaller clients become the growth focus, with the smart pack as the entry point to more services per customer.
The company says it will tighten hiring, shift toward subscriptions such as Leaf and AI BOAT and multi-year contracts, cut or exit low-margin businesses, and tie sales evaluation more closely to operating profit. It aims to restore double-digit profit growth in the year to September 2027. The letter gives no numerical guidance for that year.
These are management's own assessments of one supplier's year, not a measure of the wider AI-training market.