Corporate efficiency is undergoing its biggest reset since the SaaS boom. On October 22-24, 2026, Barclays and American Express announced coordinated cuts totaling 8,500 jobs, explicitly linking reductions to AI adoption — a signal of how growth is decoupling from headcount.
1. The Headcount Reset: Fewer People, More Output
- Barclays: Cutting about 5,000 jobs in 2026, or 5% of workforce, after already reducing 1,000 roles earlier. The bank reported taking £201 million in structural cost actions in Q3, part of a plan to reduce staff costs through greater use of technology. Around 60% of cuts will impact support functions such as developers.
- American Express: Eliminating approximately 5% of its global workforce, about 3,800 to 3,900 positions based on 76,800 employees at end of 2025, with a $336 million charge. CFO Christophe Le Caillec said the goal was to become leaner and prepare for AI by reducing layers before deploying agents.
- Broader wave: eBay cutting 1,000 jobs (9%), Netflix 1,000 (6%), Qualcomm performance reductions, and tech layoffs tracking toward 300,000+ this year.
As one CEO quantified: "Our engineering organization was 30% smaller than it was a year ago, but it delivered substantially more product output." Executives now frame AI as enabling slower hiring, not just mass layoffs, with companies managing 2-4x more workloads with roughly the same people.
2. From Experimentation to Scaling
EY's May 2026 global technology executive survey found 69% identify the shift from experimentation to scaling and production as the most significant change in AI strategy, while 76% expect AI investment to rise by at least 20% over next 12 months.
Companies are moving beyond simple copilots to agentic experimentation — agents that autonomously identify bottlenecks in marketing funnels or supply chains and implement fixes in real-time.
3. Pricing Shifts: The End of Per-Seat Software
Per-seat licensing breaks down when autonomous agents, not human employees, do the work.
- Traditional model: Subscription + seat = predictable cost, but one agent can do work of five people.
- New hybrid model: Teneo research tracking hundreds of US software vendors shows 78% have transitioned to a hybrid model in 2026 — about 30% fixed platform/seat fee plus 70% variable usage (tokens, credits, or outcomes).
- Real examples: Intercom's Fin AI agent charges $0.99 per successful resolution (outcome-based), Workday introduced Flex Credits in August 2026 for HR/Finance value-added actions, Salesforce Agentforce charges credits per action alongside per-user options. Microsoft said during Q3 2026 earnings that business apps are moving to "seats plus consumption" and nearly 60% of customers buy usage-based credits.
IDC expects subscription pricing to fall from 60% to 36% by 2035, while consumption rises from 11% to 33% and outcome-based from 3% to 24% — together 57% of pricing by 2035 vs 14% in 2025.
4. The Efficiency Trap to Avoid
92% of agentic AI deployments exceed cost expectations because a single customer-service task can trigger 4-8 API calls. The winning enterprises build cost controls into agent design, track cost-per-resolved-task during pilots, and define task boundaries before building.
Bottom Line: The corporate efficiency shift of 2026 is not just about cutting jobs. It is about re-architecting work itself — integrating AI agents into workflows, moving from seat-based to consumption-based economics, and measuring productivity per dollar, not per head.
You must be logged in to post a comment.