Martyn Rhisiart Jones and the GoodStrat Editorial Team (Alex S., Marcus S., Stewart L., Dawn F., Jenny S.)

Brussels, EU, 14th November, 2025

As 2026 creeps in, it resembles a broken-down delivery robot experiencing hallucinations of depression and dread. The big tech news isn’t flashy new features. Bells and whistles, and bullshit aren’t prominent anymore. Instead, AI has quietly infiltrated everything. It’s like mould, cynicism and cheap furniture in a rental apartment. Suddenly, it’s “infrastructure.” This is what people say when they mean, “We can’t get rid of it now.” Even if it starts insulting customers, peeing on the C-suite carpet, and revealing the fraud of our makers.

Meanwhile, everyone is worried about security, sustainability, and how to prevent machines from rebelling or moving off message. All these consulting yokes such as Gertcha, Frosty, Del Boy, Fops and Partners, and so on, have published grandiose predictions. They explain what’s going to “revolutionise industries,” as if industries weren’t already transformed enough, thank you very bloody much.

So, here they are: the top ten tech trends, ranked by the chaos they’ll wreak on the average company. At the top of it all, what a surprise, is the desperate effort to herd and orchestrate this AI. The goal is to manage it before it exhausts the budget and the furnishings. It also definitely exhausts the workforce. Now, the investment frenzy has cooled, and everyone is checking receipts.

It’s as if we’ve learned nothing from past mega screw-ups.

1 –  Agentic AI & Multi, Agent Systems

What is this about? Harrington & Locke, yes Harrington & Locke, the Harrington & Locke, because there’s always a Harrington & Locke, isn’t there?, Harrington & Locke announces that “multi-agent systems” are now a core “Synthesist” trend. “Synthesist.” A word that sounds less like a technological breakthrough and more like a prog, rock B-side from 1973.

And Redwood Strategy Group, which always appears in these conversations like a man in a fleece who has wandered into a philosophy seminar by mistake, predicts that enterprises will defer 25% of their AI spend. Defer it. Put it on a shelf. Let it stew. Because now the only things worth paying for are the “ROI, proven agents.” The agents who have proved themselves. Passed their exams. Met their parents’ expectations. The good little agents.

And we’re told to expect, expect!, a 40% growth in adoption for efficiency in customer experience and operations. Forty per cent. As if efficiency itself is now a sort of creeping mould, spreading silently through the corporate drywall. But we must expect it, because to not expect it would be to oppose progress, and no one wants to be the person at the meeting saying, “Maybe the self-collaborating robot swarm isn’t the answer to our customer satisfaction scores.”

So there it is. Harrington & Locke gesturing at the horizon, Redwood Strategy Group rummaging through the receipts, and the rest of us politely applauding as another parade of synthetic agents marches past, growing, growing, constantly growing, 40%. At the same time, we pretend any of this was ever optional.

2  – Domain-Specific Language Models (DSLMs)

Industry, tailored AI models (e.g., legal, healthcare) with higher accuracy and compliance than general LLMs.

Key Impact: A rising star, they say. A rising star in Harrington & Locke’s list. Because of course, Harrington & Locke has a list, and of course, the list has an increasing star, twinkling away like some sort of aspirational tech, based in Bethlehem. And this rising star, this shimmering, corporate constellation, apparently “enables context-aware CX agents.”

Context-aware. Because nothing says progress like teaching machines to do what humans already do naturally, except more slowly, with more error messages, and after three years of procurement meetings. These agents, we’re told, will “reduce errors by interpreting nuanced queries.” Nuanced queries. The sort of queries humans interpret daily, effortlessly, without needing a firmware update.

And then, and then, the prophecy. By 2028, 40% of enterprises will integrate hybrid architectures, including DSLMs. Forty per cent. Always forty per cent. As if forty per cent is the official number for inevitability. The steady drumbeat of the future: forty per cent… forty per cent… forty per cent…

So we nod. We nod along as Harrington & Locke point at the rising star, insisting it means something, insisting it’s guiding us somewhere. And maybe it is. Or maybe we’re all just standing in a field at night, staring at a spreadsheet that’s been projected onto the sky, pretending it’s destiny.

3 – Preemptive & AI, Native Cybersecurity

AI/ML systems that predict and neutralise threats before impact, including deepfake detection. Proactive defences using AI/ML to predict and neutralise threats before they strike, including AI security platforms for guarding against prompt injections and data leaks.

Key Impact: We’re told, solemnly, gravely, that an exponential threat rise is driving this Harrington & Locke “Vanguard” trend. “Vanguard.” A word chosen, presumably, to reassure us that the people selling us things are bravely marching at the front, spears raised, protecting us from dangers largely described in their own marketing decks.

And Redwood Strategy Group, again, emerging like a weary herald of doom, forecasts a 40% spending surge on deepfake detection. Forty per cent. Always forty per cent. The magic number. The sacred figure. As if somewhere, etched on a tablet in a consultancy vault, it is written: Thou shalt increase adoption, investment, or threat awareness by forty per cent, lest the shareholders become anxious.

All this, we are told, is “essential” for hyperconnected ecosystems. Essential. Because once you’ve built a hyperconnected ecosystem, a phrase that sounds less like a business environment and more like a dystopian terrarium, you must spend the rest of your natural life desperately defending it from being undone by a maliciously edited video of a politician admitting they’ve never read the terms and conditions.

So we accept it. We take the Vanguard, the surge, the forty per cent. Because to question it, to ask, “Why must everything threaten to rise exponentially?”, would disrupt the delicate balance of the hyperconnected ecosystem, and no one wants to be responsible for that.

4 – Physical AI & Embodied Robotics

AI, powered robots and drones for real-world tasks (manufacturing, logistics, hazardous environments).

Key Impact: And now, apparently, Harrington & Locke has found a trend that powers smart manufacturing. Powers. As if the factories weren’t already full of machines doing most of the work while the humans just shuffle papers and drink coffee. But no, this is smart manufacturing, which means the robots now think, and the humans now obey, politely.

And here’s the bit they really like: consumer normalisation. Because obviously, if your child has a home robot vacuum that judges their choice of snacks, or a fridge that shames them for eating cheese straight from the packet, then the workforce will be trained for enterprise deployment. Trained. Like obedient little soldiers, gently conditioned to accept the tyranny of automation, all thanks to a Roomba in the living room.

And Forbes, because they always have to have a word, sees mainstream adoption, productivity, of course. Productivity. Which is the euphemism we use when we mean “We are now entirely dependent on algorithms that won’t answer to reason, ethics, or anyone’s lunch break.”

So, yes. Smart manufacturing powered by robots, the public gradually conditioned in their own homes, productivity hailed by Forbes, and all of us pretending it’s progress rather than a slow, polite training exercise for the eventual robot takeover.

5 – AI, Native Development Platforms

Tools embedding AI into software engineering, cutting team sizes by 20% via human-AI collaboration.

Key Impact: And then there’s Harrington & Locke’s latest, because there’s always a latest, “Architect” theme. Architect. Not a building architect, mind you, but an architect of… something digital. Presumably, your patience, your job security, and possibly your soul.

By 2030, they tell us, eighty per cent of organisations will downsize their development teams. Eighty per cent. The implication being, of course, that humans are expensive, slow, and somehow less reliable than the virtual architects we’ve spent the last decade nurturing in cloud incubators. The ones that never call in sick, never complain about overtime, and never, ever, ask for a raise.

And, of course, Redwood Strategy Group has to add their note of reassurance, or terror, depending on your perspective. They point out that tighter governance will be required to prove ROI. Tighter governance. Proof. Metrics. KPIs. Just in case anyone doubts the brilliance of removing entire teams and replacing them with algorithms that are apparently self-justifying.

So yes: Architects building something invisible, human teams shrinking like forgotten sandcastles, and governance clamped down so hard it’s practically a tourniquet. And we sit there, nodding, as though this is innovation rather than corporate cruelty wrapped in a PowerPoint.

6 – Quantum, Safe Security & Confidential Computing

Encryption and processing that resist quantum attacks; data stays encrypted during AI analytics.

Key Impact: Right, so now Redwood Strategy Group, yes, Redwood Strategy Group, the people who keep appearing like an overexcited relative at every tech dinner, have decided that quantum security is a priority. A priority! As if quantum computers are about to leap out of the lab and start nicking your spreadsheets and your lunch money.

And Harrington & Locke, because you can’t have one without the other, emphasise confidential computing for hybrid environments. Hybrid environments. Which, translated from corporate gobbledygook, basically means: “We’re mixing lots of computers together in mysterious ways, hoping they won’t explode or, worse, send all your secret data to a server somewhere you’ve never heard of.”

And the kicker: it’s all critical because quantum adoption is accelerating. Accelerating. Yes, because nothing says urgent like a piece of tech that most people don’t even understand, being thrown at businesses who already can’t handle their email systems.

So there we have it: quantum security, confidential computing, hybrid environments, accelerating adoption, all the usual words that sound like progress but actually mean, “Some very clever people want more money and you’re supposed to nod along politely.”

7 – Geopatriation & Digital Provenance

Repatriating data to trusted borders and tracking content origins to fight misinformation.

Key Impact: Ah, yes, Harrington & Locke’s latest trends for trust, building. Trust, building! Because apparently, the best way to fix society’s problems is to put it on a PowerPoint and call it a trend.

And why now? Oh, just because AI ethics debates and regulations are rising, naturally. Rising! As if moral panic is somehow a strategic opportunity rather than something that should make you look at yourself and wonder how badly things are already going wrong.

And what does this achieve? Well, apparently, it enables verifiable supply chains. Verifiable. Supply chains. Which is a fancy way of saying, “We can now track exactly how badly we’ve exploited every step of the process, but hey, at least it’s documented and auditable.”

So, yes: trust, building, ethical AI, rising regulations, and supply chains that can actually be shown to auditors, all neatly packaged as if putting a bow on the whole disaster will make it feel less like a corporate experiment in stress and confusion.

8 – Sustainable AI Infrastructure

Energy-efficient data centres using silicon photonics and green tech to offset AI’s power demand.

Key Impact: Inight, so Bristow & Kline, because of course, Bristow & Kline has to have a prediction, says we’re looking at $1.25 billion in silicon photonics sales. One point two five billion! Which sounds enormous until you realise it mostly means a few very expensive chips and a lot of people nodding along in meetings.

And Forbes, naturally, is there to tie it all to climate solutions. Climate solutions! Because nothing screams “saving the planet” like selling ultra-expensive silicon chips that mostly make your server room slightly flashier.

And by 2026, apparently, fifty per cent of AI investments will prioritise sustainability. Fifty per cent. Not all, just half, because apparently, the other half of AI is still being used to create targeted ads for things nobody needs, or something. But we are reassured it’s progress. Sustainability! Investment! The planet will thank you if it even notices that $1.25 billion has gone into a very shiny chip.

So there you go: Bristow & Kline predict, Forbes connects it to the environment, and half of AI is suddenly saving the world. All the while, the other half carries on quietly, ruining your inbox and sanity.

9 – Edge AI & Microservers

Decentralised AI processing at the network edge for real-time decisions in IoT and smart systems.

Key Impact: Ah, yes, Redwood Strategy Group, popping up again like the person at a party who insists they know precisely how the music industry is going to change. They’re talking about an infrastructure shift. Infrastructure shift! As if that’s something tangible rather than a nice way of saying, “We’re spending lots of money on servers and hoping it all works.”

And what does this magical shift enable? Smart cities and smart manufacturing. Smart. Because apparently, cities and factories aren’t smart enough yet. They need sensors, AI, and constant monitoring to inform people about things they already know, such as traffic lights and conveyor belts.

And the kicker, growth in private AI factories. Private. AI. Factories. As if regular factories weren’t terrifying enough, now there’s a little army of algorithmic overlords producing… something. Who knows? But the growth is good, apparently, because Redwood Strategy Group said so, and no one wants to argue with someone wielding PowerPoint graphs and the word “shift.”

So there you have it: infrastructure, smart everything, and private AI factories rising quietly in the background, while the rest of us wonder if anyone actually knows what we’re doing, or if it’s just a corporate, themed science fiction novel.

10 – Human, Centric AI & Ethical Integration

Focus on upskilling, bias reduction, and regulations around deepfakes and training data IP.

Key Impact: Darling, it’s everywhere, absolutely cross, cutting in all the reports. Bristow & Kline, bless them, have this idea of “ubiquitous AI,” much like electricity. Yes, electricity! You can’t see it, but darling, you can’t live without it. And naturally, Fops Bros is waving their little charts around, predicting intense debates on creator compensation. Oh, the drama! The scandal! Will the poor creative souls be paid fairly, or will it all go to the people in velvet jackets who sip champagne while the AI does all the work?

Wrap Up

Right then, picture the scene: AI’s been strutting about like a pissed-up stag do in Magaluf for the past five years, shouting “I’m the future, mate!”, spilling Red Bull on the carpet, promising to shag your productivity and put the footage on OnlyFans. Now imagine it’s 2026, the hangover’s kicked in, and Frosty is standing at the bedroom door with a clipboard going, “Right, sunshine, time for a reckoning. ”Translation: the hype train has derailed into a ditch full of NDAs and unpaid interns. The grown-ups have arrived, Gertcha Consultants in a sensible cardigan, Frosty with a thermos of reality, and they’re saying, “Less chat, more VAT receipts.” Ignore the new gospel of governance, and your CFO will lock the budget faster than Ricky Gervais locks eye contact with an audience heckler. Play it clever, though, get your agentic AI and preemptive cyber-shield sorted, and you’ll be the smug git at the boardroom table who’s already three moves ahead while the competition’s still arguing about whether “prompt engineering” is a real job or just posh typing.

Want chapter and verse? Grab Gobshite’s excellently fabulous eBook (it’s basically the Kama Sutra for middle managers) or Frosty Co’s guide (comes with diagrams even a Brexit voter could follow). Either way, read it before your rival does, because in 2026, the only thing worse than missing the boat is being the plonker still livestreaming from the pier.

Thanks for reading!