WRITER

If you're looking at this list and wondering where to actually start, you're not alone. Most marketing teams have 10-15 AI opportunities mapped out but spend months agreeing on priorities. We created a toolkit that helps teams cut through that and identify their top workflows quickly. Get the framework 👉 https://lnkd.in/eJ3miGN4

"I looked at my operation and thought, this isn't an opportunity to just keep adding additional headcount." That realization changed everything for Bradley Lane at EE. What came next: content creation time cut by 92%. Quality jumped from 30% to 100%. Product catalog scaled from 60 to 4,500+. He walks through the entire playbook alongside our Chief Customer Officer, Mina Alaghband on Jan 21 https://lnkd.in/epiQD-wP

Most AI vendors will sell you the platform and wish you luck. At WRITER, we think that's exactly where the real work begins. The gap between "we bought AI" and "AI transformed our business" is enormous. How do you reskill teams? Redesign workflows? Scale across 10,000 employees securely? These are the questions most vendors don’t want to touch. At WRITER, we know these aren't your problems to solve alone. They're our shared mission. That's why we’re absolutely thrilled to announce that Mina Alaghband has joined WRITER as our first-ever Chief Customer Officer! 🙌 Mina's spent years as a Partner at McKinsey & Company, guiding enterprise leaders through precisely the kind of transformations most companies only dream of completing. She doesn't just know what it takes to go from pilot to enduring scale; she's proven it. Now, she's bringing that expertise to our customers as they build the talent, frameworks, and operational muscle to win in the age of agentic AI. Welcome to WRITER, Mina!

2026 will bring a brutal leadership reckoning. Most C-suite leaders still aren’t getting ANY agentic AI done in the enterprise. Not because the technology isn’t ready (it is) but because leadership hasn’t caught up. AI is moving at warp speed but the playbook at the top has barely moved at all. The first generation of AI leaders already had its sink-or-swim moment. But three years in, the era of learning as you go is over. If you’re not moving to production-grade, BUSINESS-REDEFINING systems now, you’re not just behind. You’re at real risk of never catching up. The window is closing FAST. And here's the quiet shift, the real game-changer: Companies are now realizing they have to go 3 or 4 levels deep into their organizations to find the leaders who will actually drive the agentic enterprise revolution. We’re talking about the unsung heroes in operations, marketing, GTM – the ones inside the workflows, living with the friction every single day. Close enough to the work to know what has to be rebuilt, not “AI-enabled.” This is the year we’ll see a new class of leaders emerge. These aren't the folks content to just patch over your broken, 20-year-old processes with AI. They’re the ones ripping out organizational muscle memory altogether. They’re redesigning operations for human-agent collaboration by default: re-architecting how decisions get made, how work flows, and how careers are shaped in an agentic world. I wrote more on this leadership shift for the World Economic Forum, ahead of their annual meeting in Davos: https://lnkd.in/eJ5dg2y9 This is the number one dialogue enterprise leaders should be having right now. And Davos is where it will start. If you’ll be there, find me!

WRITER is heading to the World Economic Forum, Davos! 🏔️ Join us on January 19-23, as we tackle the biggest question facing enterprise leaders: How do we build organizations that are not only more powerful and efficient, but profoundly more human in the age of AI? What we're bringing to the conversation: → The end of scarcity-based strategy and the rise of the limitless enterprise → How to replace bureaucracy with radical simplicity → Building trust infrastructure for human-agent collaboration → The new leadership model for AI-native organizations https://lnkd.in/e5Ydbtsv

Over the holidays, everyone started talking about context graphs — the missing infrastructure layer that's keeping AI stuck at incremental gains. We've been building it for over a year. We called it the orchestration graph. Here's what we learned: Coding transformed first because the infrastructure was already there. Git commits, code reviews, CI/CD pipelines — the "context graph" practically built itself. Marketing is different. You can't write on-brand messaging without understanding brand evolution. You can't approve competitive claims without knowing legal precedent. You can't scale campaigns without capturing what worked and why. The reasoning exists — it's documented in briefs, review threads, post-campaign analyses. The problem? It's scattered across Google Docs, Slack, Asana. When people leave, it disappears. That's what Enterprise Brain solves. And we’re deploying it in our own marketing team first. Real results to follow. Our CMO Diego Lomanto shares his full thoughts on why context graphs for marketing require a different approach ↓ https://lnkd.in/edkkc-CX

Scaling from 60 to 4,500 products online sounds impossible, right? Not when you have the right AI strategy. Bradley Lane, Head of Product for New Markets at EE, is sitting down with our team to share exactly how his team did it — and the results will blow your mind. → Content creation time cut by 90% (3 hours to 15 minutes per product) → Quality standards jumped from 30% to 100% → Online products scaled from 60 to 4,500+ …all without having to scale the headcount! Bradley's bringing the real playbook: the metrics that matter, the change management strategies that worked, and his vision for what's coming next. https://lnkd.in/edqNcZjs

"AI slop." Diego Lomanto (CMO, WRITER) and Andrew Strickman (CMO, New American Funding) didn't shy away from the hard truth in their recent conversation: everyone has access to AI now — so how do you actually differentiate? Their answer? You don't avoid AI. You collaborate with it. Andrew shared how his 85-person marketing team uses WRITER daily — not to replace creativity, but to amplify it. From scaling their award-winning "Hell Yeah" campaign across 25 social handles to ensuring compliance in a highly regulated industry, AI became the productivity engine that freed his team to focus on what matters: the human spark. The philosophy: "Use AI for scale and speed. Funnel those gains to the places where humans make a real difference." What a masterclass in strategic AI adoption! 👏 https://lnkd.in/emA_7up8

A theme is emerging as we head into 2026: the real bottleneck in large models isn’t parameter count — it’s signal propagation and memory coherence across depth. For years, we treated skip connections and residuals as “solved.” They helped us train deep networks, but they also masked a deeper issue: as models grow deeper, layers increasingly pass signals forward without meaningfully transforming or retaining them. Depth becomes structural, not computational. This is why recent work on hyper-connections is so important. Improving how information routes, compounds, and survives across layers isn’t an optimization detail — it’s an architectural reset. And it’s notable that 2026 is starting with DeepSeek putting real weight behind this direction. Palmyra-Creative LLM runs at ~122B parameters with ~140 layers. That depth raised eyebrows early on, but depth itself wasn’t the problem. Memory was. Once memory is treated as a first-class architectural concern — not just longer context windows, not just more tokens — depth stops being fragile and starts compounding capability. For context: - Palmyra-Creative: ~122B params, ~140 layers - DeepSeek v3: ~685 params, ~61 layers - GPT-4: ~1.8T params distributed across ~120 layers These models make very different tradeoffs between width, depth, and signal flow. What’s becoming clear is that raw scale alone doesn’t solve reasoning or reliability. How information propagates across layers matters just as much — if not more. Hyper-connections point in the same direction we learned the hard way: - Depth is valuable only if signal doesn’t collapse - Memory must persist across layers, not reset every few steps - More layers should mean more computation, not more identity passes If this trajectory holds, 2026 won’t just be about bigger models. It will be about architectures that can actually use depth effectively. That’s when real step-function gains tend to happen. Paper: https://lnkd.in/ezMNiRQh Palmyra-Creative: https://lnkd.in/e9pHACW8

How do you get people to trust AI? Simple: transparency. In the newest episode of the #HumansofAI podcast, Rowan Reynolds (General Counsel at WRITER) explains why understanding every single step — from how outputs are generated to where information is pulled from — is the only way to build reliable AI systems. Whether you're deploying simple tools or complex agentic flows, the principle is the same: trust comes from understanding. And that's the foundation of human-centric AI. Worth a listen if you're building (or using) AI in any capacity. 🎧