My technology management operating model
I manage technology as an investment portfolio that connects business priorities, human capability, trusted data, responsible controls, reliable delivery, adoption and evidence.
The job
Technology and AI are useful only when they enable an organization to attain its business goals. My job is to help technologists understand what the business needs, help the business understand what technology makes possible, and keep both accountable for the outcome.
I lead across five responsibilities: represent technology in executive decisions; govern investment, risk and performance; build teams and organizational capability; create and scale useful innovation; and advocate for the customer, the workforce and the business value of technology.
The operating model
Every initiative must cross the whole chain. Skipping a link creates a pilot, a platform or a promise—not a durable outcome.
The integrated management model
Technology management works when business judgment, technical discipline and human leadership operate as one system.
Give technology a voice in enterprise decisions, define the outcome and accountable owner, then understand the actual workflows, systems, data, constraints and people involved.
Validate that the problem is consequential and owned. Fund conviction, test assumptions and make value, feasibility, readiness and risk visible enough to defend yes, no, sequence, scale or stop.
Name decision rights, owners and definitions. Build foundations that teams can secure, fund, operate and reuse—not architecture that succeeds only as a diagram.
Connect business and technology, innovation and risk, model and workflow, sponsor and user. Prove value quickly, then add the controls, support, skills and redesigned work needed for responsible scale.
Manage short and long term, innovation and continuity, leading and enabling. Build relationships as deliberately as systems, and cultivate judgment and ownership at every level.
Human commitments
Technical rigor and human dignity are not trade-offs. They strengthen each other.
No weaponized metrics, hidden logic or humiliation. Make reality visible in a way that helps people act.
Clear priorities, definitions, ownership and constraints protect teams from avoidable churn.
Limit work in progress, distinguish urgency from importance and preserve the capacity to think.
Delegate outcomes, make guardrails explicit, coach judgment and let people own the result.
A conviction library
They are grouped by leadership, operational reality and the lessons of experience—but they work as one system.
01“Forecast value, enabled value, validated outcomes and realized benefits are four different claims. When leaders present them as equivalent, credibility disappears long before the investment does.”
02“Most data problems are ownership and definition problems wearing a technology costume. A new platform cannot create trust in information nobody owns or consistently defines.”
03“Pilots rarely fail because the model cannot perform. They fail at the interfaces—between business and technology, innovation and risk, model and workflow, sponsor and user.”
04“The next phase of AI is not simply better models. It is governed adoption at enterprise scale—and that requires an operating model, not another technology program.”
05“Agents will change workflows much as cloud changed infrastructure. Organizations that understand their processes, decision rights, data and controls will adapt. Those that do not will automate their confusion.”
06“Responsible AI is not a policy document. It is business ownership, risk tiers, minimum guardrails, lifecycle controls and human accountability—operating every day.”
07“Many AI initiatives will be quietly retired over the next few years. The survivors will not be those with the greatest enthusiasm, but those with the clearest ownership, adoption and evidence.”
08“Upskilling is the AI strategy most organizations underfund. A capability people cannot confidently use creates little value, however impressive the launch event.”
09“Architecture creates value only when teams can build it, secure it, operate it, fund it and adopt it. A target state that fails any of those tests is still a diagram.”
10“The next five years of AI will not be won on capability alone. They will be won on adoption—and adoption begins by listening to the people whose work is expected to change.”
11“Enterprise AI proves itself on the plant floor, not in the boardroom. A strategy that cannot survive operational reality is still only a presentation.”
12“Any credible AI roadmap begins by mapping what is actually there: the workflows, systems, data, ownership and constraints—not what the documentation says should be there.”
13“In continuous operations, experimentation must respect continuity, safety and trust. Before committing technology, establish that the problem is real, consequential and owned.”
14“Industrial AI will be won by organizations that understand themselves first—their operations, decisions, data, constraints and people. The technology is rarely the hardest part.”
15“I have founded and sold three companies, and I have sat on the buying side as a CIO. Both chairs teach the same lesson: the most expensive technology mistakes are usually made before anyone writes code.”
16“I taught Technology Entrepreneurship for eight years. The hardest lesson was opportunity validation—distinguishing an interesting idea from a problem someone truly owns and will act to solve. Enterprise AI is relearning that lesson now.”
17“I treat technology as an investment portfolio, not a cost centre. That means funding conviction, testing assumptions and stopping initiatives that can no longer justify themselves. Without evidence, portfolio management becomes politics.”
18“I have governed dozens of AI opportunities across large organisations. The discipline is not simply picking winners. It is making value, feasibility, readiness and risk visible enough to defend both yes and no.”
19“Building an enterprise Data and AI operating system taught me that strategy is the easy artifact. The difficult work is underneath it: ownership, trusted data, decision rights, controls, adoption and evidence.”
See the frameworks behind the operating discipline or start a conversation about applying it.