Whether an investment continues
I rebuilt portfolio reviews so every investment returns with evidence against its intended outcome. I bring the recommendation; a joint technology and business panel decides to persevere, pivot or stop.
Leadership · Roman Lobus
Leadership summary
My work connects investment choices with changes to how an organisation works. I shape portfolio decisions, redesign product operating models and lead AI adoption, staying involved through implementation.
Decisions I shape and carry through
Decisions and implementation.
Where AI goes first
I recommend concentrating AI where delivery is constrained and business value is highest, such as getting changes to customers sooner, rather than where a single task is easiest to speed up.
Whether a pilot earns wider investment
I judge AI pilots on review load, quality, end-to-end flow and adoption. When saved coding time reappeared in review, I changed the recommendation: fix the review step before scaling.
What an AI investment really costs
I model usage and consumption costs alongside licences when recommending AI tools and partners, because the invoice is only part of the cost.
Where AI may act and where a person decides
I co-designed the controls for AI-assisted delivery: review rules for AI-generated code, tool and agent permissions, and data and security policies.
What happens to released capacity
Time saved is not value until someone decides what it is for. I recommend clearing the bottleneck first, then putting capacity into backlog work and quality.
How roles and skills change
I led the changes AI brought to review and verification roles, AI skills training and team ways of working.
Who owns a benefit, and what leaders aim for
I redesigned the product lifecycle so each intended benefit has a named owner and is revisited after release, and coached product leaders to set outcome goals instead of output targets.
Selected work.
Making continuation a decision
Whether a funded investment should continue as planned, change direction or stop, and what evidence that call should rest on.
Following the time AI saves
Which AI tools and practices deserved wider investment, what evidence was strong enough to justify that step, and what had to change around them first.
Giving benefits an owner
Where benefits realisation belongs in the product lifecycle, and who is accountable for a benefit once the work has been delivered.
Contact
For questions about the work, a role or a project.