And Then There’s Reality

Big promises meet complicated systems and tight budgets. David Linthicum cuts through the hype to explain what works, what doesn't and why.


  • A computer generated image of three cloud shapes representing private, hybrid, and public clouds
    Image attribution tooltip
    Getty Images
    Image attribution tooltip

    Hybrid cloud won — now CIOs have to make it work

    David Linthicum argues that hybrid cloud isn't a temporary fix. CIOs need to design, fund and govern it as a permanent operating model.

    By David Linthicum • Nov. 9, 2026
  • Male IT Specialist Walks Between Row of Operational Server Racks in Data Center. Engineer Uses Tablet Computer for Maintenance.
    Image attribution tooltip
    Getty Images
    Image attribution tooltip

    FinOps starts with architecture, not accounting

    In his latest column, Dave Linthicum explains why cloud costs should be addressed when systems are designed, not after the bills arrive.

    By David Linthicum • Nov. 2, 2026
  • Employee in front of a laptop
    Image attribution tooltip
    Getty Images
    Image attribution tooltip

    IT skills shortage often a complexity problem in disguise

    Could your IT skills shortage be a problem of your own making? Dave Linthicum argues that companies sometimes make their technology jobs harder to fill than they need to be.

    By David Linthicum • Oct. 28, 2026
  • Agentic AI security illustration
    Image attribution tooltip
    Getty Images
    Image attribution tooltip

    The AI happy path is easy, but enterprise reality isn’t

    David Linthicum says the AI happy path ends where enterprise reality begins. CIOs must design for messy data, exceptions and failure before production.

    By David Linthicum • Oct. 19, 2026
  • Digital illustration of an AI chip shaped like a padlock, connected to glowing circuit pathways on a purple background.
    Image attribution tooltip
    Getty Images
    Image attribution tooltip

    Exception handling in AI: The silent killer of enterprise deployments

    AI systems can fail without throwing errors. CIOs need deterministic checks, escalation paths and telemetry to catch plausible but wrong outputs before they compound.

    By David Linthicum • Oct. 12, 2026
  • AI Agents and Prompt Automation Concept. Person using laptop with virtual prompt, workflow, coding, and AI icons. autonomous agent, task automation, smart integration, digital productivity, Chatbot,
    Image attribution tooltip
    Getty Images
    Image attribution tooltip

    AI makes workers faster -- and reviewers busier

    AI productivity gains can disappear when review, verification and exception handling pile up elsewhere. CIOs need to measure the downstream work, not just the initial speed.

    By David Linthicum • Oct. 5, 2026
  • An analyst uses a computer and dashboard for data business analysis and data management system with KPI and metrics connected to the database for technology finance, operations, sales, marketing.
    Image attribution tooltip
    Getty Images
    Image attribution tooltip

    AI productivity metrics are fooling enterprises

    Faster drafts, more code and higher adoption can look like AI productivity. David Linthicum argues CIOs should measure end-to-end outcomes, quality and hidden rework instead.

    By David Linthicum • Sept. 28, 2026
  • Revenue Operations Concept
    Image attribution tooltip
    Getty Images
    Image attribution tooltip

    The quiet rise of AI technical debt

    AI technical debt can hide in prompts, data, models and workflows. David Linthicum explains why CIOs need production discipline before pilots scale.

    By David Linthicum • Sept. 21, 2026
  • Ai agents illustrating artificial intelligence network concept
    Image attribution tooltip
    Getty Images
    Image attribution tooltip

    Why enterprise AI fails at the org chart, not the model layer

    AI models are easy to access. Organizational authority is not. David Linthicum explains why the org chart -- not the model layer -- keeps enterprise AI from delivering.

    By David Linthicum • Sept. 15, 2026
  • AI powers big data analysis and automation workflows, showcasing neural networks and data streams for business.
    Image attribution tooltip
    Getty Images
    Image attribution tooltip

    Who owns the workflow? Why enterprises are managing the wrong thing

    Enterprises manage applications well but often lack owners for cross-functional workflows. AI automation reveals that application ownership alone doesn't ensure business outcomes.

    By David Linthicum • July 30, 2026