Speed breaks trust
Meta’s CTO, Andrew Bosworth, recently admitted the company “did an atrocious job” rolling out its new Applied AI organisation.
From failing to explain the vision, destabilising management structures, and undermining trust among the approximately 6,500 employees moved into the unit.
In a rapid pivot toward generative AI, Meta consolidated thousands of engineers and product managers into a single Applied AI division.
But the execution broke down: employees were shifted between teams, reporting lines changed frequently, and leadership failed to clearly communicate how roles, careers, and day‑to‑day work would evolve.
Bosworth acknowledged that this left teams “in the lurch” and eroded confidence in both leadership and future opportunities.
This isn’t just a Meta story, it is how tech organisations are reallocating talent at unprecedented speed, flattening hierarchies, and rewiring priorities around AI capability.
The pattern is consistent across industries: large-scale redeployments, layoffs in legacy roles, and aggressive investment in AI infrastructure, often outpacing the organisation’s ability to absorb the change.
WHY IT MATTERS
Meta didn’t fail on strategy, it failed on meaning. Learning a lesson many industries have learned the hard way.
Its people weren’t told how their expertise translated, how their status would shift, or how they could “win” in the new system. They broke the psychological contract.
The lesson is blunt: AI programmes succeed or fail on perceived fairness, identity continuity, and narrative clarity, not just org charts and capability builds. The things a good change manager would help you see and fix.
WHAT TO WATCH FOR
→ Corrective signals: smaller manager spans, fewer forced reshuffles, and more explicit career pathways tied to AI skills.
→ More overt messaging linking “AI capability” to individual advancement to rebuild alignment between effort and reward.
LIMITATIONS
This is from an internal memo surfaced during a period of visible employee unrest and talent competition. The leadership admissions may be both genuine reflection and strategic signalling. External reports suggest deeper dissatisfaction and low morale which may not be fully captured in executive narratives.
SOURCE
BESCI AI OPINION
Really? Really?
With all the resources available to it, with all the power of AI to help it and they make rookie errors that so many industries have made before.
Or have they?
To make an omelette you need to break a few eggs. Especially if you want to move quickly. There are trade-offs. There isn't time for consensus, or detailed role profiles.
The big question is whether they have destroyed the trust that their teams had in them. That may be irrecoverable.
