Martyn Rhisiart Jones

Madrid, Wednesday 25th February 2026.

Seven years on, the verdict is in, and it’s not flattering. Back in 2019, in a Brussels-fuelled rant channelling the incomparable Bob Hoffman, I declared Agile at Scale the next frontier of IT bullshit: immature, ill-conceived, supercilious, and emphatically not agile. It killed communication with jargon, turned criticism into heresy, and mangled history by ignoring what actually worked. I called it a cultish Frankenstein of Scientology zeal, Vatican pomp, and authoritarian quirks – all wrapped in vainglorious slide decks.

Was that just hot air, or did time prove the scepticism right? From the vantage of February 2026, the evidence leans heavily toward vindication. Agile at Scale (particularly the behemoth that is SAFe ) hasn’t detonated in spectacular fashion, but nor has it delivered the revolution it promised. Instead, it has plateaued, fragmented, and frequently underdelivered, drowned in bureaucracy, overcomplication, and cultural pushback. The hype has cooled; organisations have hybridised, adapted, or quietly moved on. Simplicity, it turns out, still wins.

First, the claim that it’s not agile. I hammered home the structural gaps: how do you truly handle requirements, change, quality, data, coherence, auditability, or non-linear events in these bloated frameworks? The answers were always evasive, revealing contradictions and business incompatibility.

In 2026, this critique stands tall. Scaled approaches often smuggle Waterfall governance back in under an “agile” veneer, prioritising ceremonies and predictability over genuine adaptability. SAFe, the market leader, draws relentless fire for bureaucratic layers that stifle innovation and drag decision-making to a crawl. High-profile misfires, think Nokia’s rigidity woes or US Bank’s “agile theatre” yielding zilch, echo those early warnings.

Data backs the pessimism. Recent analyses tie scaled transformations to persistent headaches: inconsistent processes, resistance to change, and feeble leadership buy-in. One eye-opening 2024 study suggested Agile projects overall faced dramatically higher failure odds when poorly applied, often due to weak requirements engineering – precisely the upfront gaps I flagged. While small-team Agile can boast solid success (around 75% in some benchmarks), scaling dilutes it badly: failure rates in large transformations hover at 47% or worse, thanks to misfits and cultural clashes.

The market has spoken. SAFe usage dipped sharply before rebounding to around 44% in recent surveys – but almost always as part of hybrids. A striking 74% of organisations now favour hybrid or homegrown models, blending SAFe elements with lighter alternatives like LeSS, Scrum@Scale, or bespoke setups. The old “universal method” dream? Dead. The goal has shifted from raw speed-at-scale to actual customer value – a tacit admission that velocity alone often sacrificed substance.

Then there’s the kryptonite to communication. I railed against the jargon supernova: “data by design”, “learning journey”, “lean-agile mindset”; empty labels pretending novelty, eroding shared meaning until requirements became unrecognisable mush.

This endures, painfully so. In scaled environments, layers of roles (ARTs, RTEs), ceremonies (PI planning), and artefacts pile on opacity. Terms get weaponised for credibility without substance, breeding “agile theatre” where rituals substitute for progress. Everyone nods along, pretending comprehension; no one delivers. The proof? Stalled value, unrecognisable needs, and undelivered promises. Yet the tide is turning: 2025–2026 reports show organisations ditching dogma for outcomes, with AI integration pushing toward autonomous, simpler delivery. Less is more – Mies van der Rohe would approve.

Finally, the cult dynamic. I likened Agile at Scale to a mega-religion blending Scientology intensity, Vatican liturgy, and sycophantic Führerprinzip; worshipping half-baked ideas, fixating on data-as-oil fetish, bullying doubters.

Hyperbolic? Perhaps, but it captured real behaviours. The certification industry (hundreds of millions), mandatory training, and dogmatic rollouts fostered sycophancy and heresy-hunting. Question the framework in a SAFe-heavy shop? Risk ostracism.By 2026, the cult has lost its grip. Surveys show diversification: fewer mandated frameworks, more principle-focused hybrids. High-profile failures spawned memes (“agile cancer”, “industrialised complexity”), and AI’s rise reframes agility around value over rigid process. Frameworks are now starting points, not holy writ.

That’s the reckoning. Agile at Scale didn’t collapse civilisation, but it didn’t transform it either. It bureaucratised what should have stayed nimble, complicated the simple, and evangelised where evidence should have ruled. Winners simplify relentlessly, question dogma, and chase outcomes over orthodoxy. In an AI-accelerated world, the lesson is sharper than ever: people who truly know simplify; those who don’t complicate – and bullshit to hide it.

Less dogma, more delivery. Simplicity still wins.

Many thanks for reading.

First pass… Bruxelles, 28^th May 2019