Harvard Business Review with “Designing A Successful Agentic AI system”. Agentic AI systems are designed to execute workflows, make decisions, and coordinate activities across various departments, presenting significant challenges for organizations. Companies often respond to new technologies with minor adjustments, yet many fail to reorganize effectively, as highlighted in a 2018 article by Harvard Business Review, which noted that numerous high-profile digital transformations falter not due to a lack of innovation but because of insufficient structural adaptation. To harness the potential of Agentic AI, firms must establish workflows focused on outcomes and appoint mission owners who can guide both human and AI efforts. This approach requires breaking away from traditional silos and outdated methods of operation, emphasizing the need for cross-functional collaboration in increasingly complex environments.
And while I’m on that, DevOps with “What Automation Mistakes can Teach Us about AI.” The article outlines three major automation errors: pursuing quick wins without a cohesive strategy, automating in siloed systems, and establishing rigid foundations that hinder adaptability. To avoid these issues, organizations must ensure that AI is embedded within orchestrated and governable business processes from the outset, thus preventing the common failures of isolated automation tools and enhancing overall operational efficiency.
Inc with “Want AI Adoption to Actually Be Successful?”. A significant factor in the failure of artificial intelligence initiatives is the lack of training for senior leaders, who often make critical decisions about AI adoption without a practical understanding of the technology. A recent MIT report indicates that 95 percent of AI projects fail to deliver return on investment, largely due to this disconnect. As organizations invest millions in AI tools, executives are frequently excluded from hands-on learning, leading to fragmented execution and missed opportunities. To succeed, companies must ensure their leaders are not only tech-aware but also tech-immersed, understanding the direct impacts of AI on operations and employee dynamics.
Why do we care?
Big Ideas time — and today’s theme is pretty blunt: AI isn’t failing because the tech is bad. It’s failing because organizations aren’t built to use it.
Here are your questions to consider.
1. What business outcomes are we targeting — not what tools?
If the client can’t name the outcome, they’re not ready for AI.
2. Who owns the mission?
Not the tools. Not the project. The mission.
Is there a single owner accountable for both machine and human workflows?
3. Which cross-department workflows must be redesigned?
Where does data handoff break today?
Where do silos block automation?
4. What manual steps should remain human — and why?
Agentic AI requires clarity on when not to automate.
5. What governance and orchestration mechanisms already exist?
If none exist, the first phase is process stabilization, not AI deployment.
6. What leadership training is required before implementation?
Are senior leaders trained enough to make informed decisions?
7. How resilient is our workflow design?
Can we adapt when AI tools evolve monthly?
8. Do we have a viable proof-of-value path — not just a proof-of-concept?
POCs fail when isolated from the real workflow.
POVs succeed when tied to business outcomes.
9. What metrics will define success?
Cost savings? Cycle time reduction? Human error reduction?
These must be defined before deploying anything.
10. Are we ready to say no to AI projects that aren’t ready?
This is the single most important discipline MSPs can adopt.