Taller CEO Christophe Kolb makes his Forbes Technology Council debut with a piece on one of enterprise AI's least examined risks. The article makes a case for leaders whose AI systems have moved beyond personalization into something more consequential: building models of customer vulnerability from ordinary behavioral data. https://lnkd.in/eDwsc94J
Taller
Servicios y consultoría de TI
San Francisco, California 12.328 seguidores
Leading the agentic AI revolution in IT services and solutions.
Sobre nosotros
Taller is the enterprise accelerator for digital transformation, expertly orchestrating hybrid teams of senior specialists and AI agents under trusted oversight — the "humans in the loop" delivering unparalleled speed, scale, and strategic impact. Subscribe to our monthly newsletter covering the latest breakthroughs in enterprise AI: https://hubs.ly/Q03tqbNy0
- Sitio web
-
taller.us
Enlace externo para Taller
- Sector
- Servicios y consultoría de TI
- Tamaño de la empresa
- De 201 a 500 empleados
- Sede
- San Francisco, California
- Tipo
- De financiación privada
- Fundación
- 2008
- Especialidades
- Agentic AI, AI-Powered Solutions, AI Systems Engineering, Consulting Services, Custom Software Development, Digital Product Engineering, Data & Analytics, Cloud & DevOps, Business Intelligence & Data Science, DevSecOps, Enterprise Platforms y Robotic Process Automation
Ubicaciones
-
Principal
Cómo llegar
1 Sansome St.
Ste. 1400
San Francisco, California 94104, US
Empleados en Taller
Actualizaciones
-
There are two ways to scale AI: artisanal and industrial. Artisanal: isolated use cases, one-off tools, incremental adoption Industrial: multiple agents embedded across workflows, with the architecture, governance, and tech stack to match PwC reports that firms making the move have seen the cost of core operational workflows drop roughly 30%. https://lnkd.in/ezeGtRCm
-
-
Two weeks into Taller Academy: 161 active users, 1,900+ enrollments, 17 completions. And we're just getting started 🏁
Two weeks since we launched Taller Academy, and here’s where things stand today: → 161 active users → 1,900+ enrollments → 17 certificates already issued The number that gets me most isn’t the biggest one; it’s the 17. That’s 17 people who didn’t just enroll; they went through the whole thing and finished. That’s the part that’s hard to fake. What’s been surprising us internally is the momentum this is creating. There are moments where we stop and go, “wait, this is a little crazy”, but never because AI feels like it’s limiting what we can contribute. It’s the opposite. It gives us this shared sense that there’s no ceiling on what we can try, and every time the team pulls something off that’s better than what we imagined, it just fuels the next push. Two weeks in, the numbers tell me people are doing exactly that. More updates as the cohorts keep growing. Still just the beginning!
-
-
Most enterprises are confident they can deploy AI. Far fewer are ready for what scaling it actually demands. Deloitte's 2026 Global Technology Leadership Study surveyed more than 660 tech executives and found a telling gap: 81% say they can deploy and govern AI at scale today, yet nearly 75% acknowledge their operating model will need to change within the next 12 to 18 months to sustain progress. The operating model, not the technology, is at issue. The figure here shows what that could look like in practice. Human gates at every decision point, agents running execution, a feedback loop that spans the whole system. Read the full piece here: https://lnkd.in/g_WRvJwN
-
-
Two lines. Four quadrants. The illusion of order. A recent essay by Taller CEO Christophe Kolb — "The Transcendental Grid: Why Managers Keep Drawing Four Boxes" — interrogates our instinct for the 2x2. The matrix disguises the very thing it claims to organize: the messiness of human decision-making. Simplifying is real work. But the grid deserves harder questions: who chose the axes? Who drew the chart, and for what purpose? Read the essay: https://lnkd.in/g27mYRrt Learn what's inside each quadrant 👇
-
Most teams building on LLMs are still using deterministic instincts on a probabilistic system. Yet the mental models used for reliability — find the bug, fix the input, rerun the test — were built for predictable systems. LLMs generate every answer fresh, sampling from a probability distribution shaped by context, prompts, and retrieval. Control and governance in that environment are a different engineering problem. This visual by Taller CEO Christophe Kolb gives that problem an intuitive shape with a little help from Richard Feynman. It draws on Vitalii Oborskyi's "Uncertainty Architecture" to bring the core concepts to life: what it means to influence a probability distribution, how prompts and retrieval function as control signals, and how closed-loop feedback makes stochastic behavior more observable, bounded, and governable. Swipe to see the architecture that makes LLM-based systems more operationally controllable and trustworthy within defined boundaries. →
-
47% of tech leaders say high-performing human and AI agent teams will define the operating model of the future. Deloitte's 2026 Global Technology Leadership Study is direct about what that means: businesses must focus less on organizational structure and more on how leadership, capital, work, and risk are aligned around core business outcomes. Check out the full report here: https://lnkd.in/g_WRvJwN
-
-
Today we launched Taller Academy, a new internal learning platform built by Taller, for Taller. It’s one more way we’re helping our whole community build practical AI fluency together. Proud of the team behind this launch, and excited for what comes next.
Big day for us at Taller Today we officially launched Taller Academy, our internal AI learning platform for the whole Taller community. It started as a simple idea: we needed a better way to make AI trainings available to our people. Something practical, low-friction, a place to put courses so people could access them. It didn't stay simple for long. Somewhere along the way, "a place to put courses" turned into an actual experience, guided by our own AI, adaptive to where each person is starting from, built with the same care we'd put into any product we ship for a client. That shift happened because of an incredible dev and product team who took a training-access problem and turned it into something we're genuinely proud to put in front of our community. But if I'm honest, the part I'm most proud of is the mindset shift across the whole team. Somewhere in the last few months, we committed to treating AI as a skill to build, deliberately, together. Being "AI ready" means being the ones who shape how the technology gets used. That change in thinking is the real infrastructure behind everything we launched today. Today is proof of what happens when a team decides to be the architects of its own evolution. Proud of this team. This is just the beginning.
-
-
McKinsey analyzed nearly five hundred PE-backed companies and found that AI used only to cut costs produces nearly the same revenue multiple as no AI at all. Embedding AI in products reaches 20x. Building new AI-anchored business lines reaches 31x. Read the full analysis: https://lnkd.in/giSb2FfE
-
AI ate the work that made junior employees good at their jobs. That routine output — the memos, the first drafts, the manual pulls — was also where judgment developed. Automate it away and you gain efficiency today at the expense of undertrained employees tomorrow. Swipe for a breakdown of Christophe Kolb and Jim Caron's recent essay on task recomposition and junior work. → https://lnkd.in/ghAeJCk3