01
No end-to-end owner
Each team owns a fragment, but nobody owns the path from business need to operation and measurable value.
Enterprise AI transformation and enablement
Turn AI ambition into enterprise capability.
Enterprise Product Owner · AI transformation and enablement · Hands-on Builder
I work with leaders responsible for moving AI beyond pilots in complex organizations. I connect business, product, technology, operations, risk, and learning so promising use cases become governed, measurable capability that teams can run and improve.
Build the organization that can make AI useful.
Where AI gets stuck
A pilot can work and still go nowhere.
The model is only one part of the system. AI stalls when ownership, evidence, controls, and learning remain split across functions.
02
Control arrives late
Risk, security, privacy, compliance, or operations are asked to approve choices they had no chance to shape.
03
Training is generic
Everyone receives the same introduction while the people responsible for use cases, decisions, and deployment lack role-specific practice.
04
Evidence stops at the demo
The proof shows that the technology can respond. It does not show that the organization can operate, govern, measure, and improve it.
What I help change
From scattered AI activity to enterprise capability.
Most organizations do not need more AI ideas. They need a way to choose the right ones, connect the people responsible, prove value under real conditions, and build the capability to keep improving.
01
Enterprise AI capability and operating model
Clarify where AI should create value, who owns the decisions, how business and control functions participate, and what the organization needs in order to move from pilots to repeatable capability.
02
AI opportunity portfolio and product strategy
Separate opportunities ready for action from those that need evidence or strategic monitoring. Define the user and business value, evaluation criteria, control needs, and the smallest useful proof.
03
Integrated AI teams
Bring business, product, data, technology, operations, risk, security, and governance into the work early enough to shape the solution—not review it after the important decisions have already been made.
04
Enterprise AI learning and translator pathways
Build role-aware capability: responsible literacy for broad use, hybrid practitioners who can translate between functions, and access to the technical and governance depth needed to run AI well. I designed and launched a company-wide AI-enablement program for more than 100 people.
Selected impact
Evidence from systems people had to use.
Selected outcomes from AI products, operational systems, and capability-building work. The organizations and internal programs remain confidential.
81%+
AI-assisted resolution
Started and built a customer support AI pilot from the ground up—knowledge base, prompts, evaluation, and feedback loops—and improved resolution from under 50% to 81%+. The system also absorbed 14% demand growth without additional support headcount.
80%+
Customer satisfaction
Increased bot satisfaction from around 60% to 80%+, bringing it in line with human support.
30%
Less manual email handling
Automated a customer-email workflow, reducing manual volume by 30% while maintaining service quality.
$30K+
Annual savings
Delivered a systems consolidation that generated more than $30K in annual savings while improving operational continuity.
100+
Hours saved per quarter
Delivered a practical AI-assisted quality-control workflow that saves more than 100 hours each quarter.
$50K+
Annual savings
Built and delivered an AI-assisted quality system generating more than $50K in annual savings.
Decision stories
What changed beyond the first working prototype.
AI support that worked under real demand
I started and piloted a customer-support AI system, then built the evaluation and feedback loop needed to improve it under real demand. Resolution moved from under 50% to 81%+, satisfaction from around 60% to 80%+, and the system absorbed 14% demand growth without added support headcount.
Quality control that returned operating capacity
I shaped an AI-assisted quality-control workflow around the work people actually had to review. It returned more than 100 hours of operating capacity each quarter. The separate annual quality-system outcome remains outside this story.
AI enablement that became team capability
I designed and led an AI Champions community program for more than 100 learners. Champions have since built more than 50 AI and low-code automation workflows as capabilities inside their own teams. Together with their managers, they became accountable owners of the AI capabilities they were building in their domains and teams—turning enablement into local ownership and continuous improvement.
A focused place to start
Enterprise AI Capability Review
Map where AI ambition is being blocked by fragmented ownership, missing translator roles, late controls, weak evidence, or a training model disconnected from delivery. The review identifies the most important gaps, the opportunities ready for action, and the next decisions the organization needs to own.
Indicative outputs
- A capability and ownership map
- A deploy, prove, or monitor view of the opportunity portfolio
- Missing translator, product, technical, and control roles
- Evidence and governance requirements for the next stage
- A practical action plan
A good fit
- Live enterprise initiatives
- Fragmented AI portfolios
- Enablement problems connected to delivery
Not the offer
- Legal advice
- Independent model auditing
- Staff augmentation
- Generic AI training
How I work
Move from signal to capability.
-
01
Discover
Read the business problem, the work, the evidence, and the changes outside the organization that may matter.
-
02
Choose
Prioritize by value, maturity, control, and organizational readiness—not by novelty.
-
03
Connect
Bring the right business, product, technology, operations, risk, and learning roles into shared decisions from the outset.
-
04
Prove
Test the smallest useful capability in real conditions, with clear measures, limits, ownership, and a fallback path.
-
05
Embed
Turn evidence into operating capability through role clarity, enablement, governance, measurement, and continuous learning.
Evidence from operation informs the next decision.
About Luka
I work between the teams that need to agree.
I started on the customer support frontline, where process failures become visible long before they reach a strategy deck. From there, I initiated and piloted a customer support AI system, moved into AI operations, and grew into Enterprise Product Ownership. That path taught me to connect practical building with product judgment, adoption, and measurable value.
My work sits between the functions that must agree before enterprise AI becomes real: business and product, technology and data, and the people responsible for risk, governance, adoption, and learning. I also bootstrapped and continue to lead The AI Fellowship Madrid Chapter, creating practical spaces for people to learn, test ideas responsibly, and make AI useful beyond the slide deck.
- Based in Madrid, Spain
- Experience Long-standing experience in distributed, fully remote companies.
- Purpose Build what helps people do better work.
Selected thinking
Judgment in public, without the performance.
Three pieces on the cross-functional decisions that make enterprise AI useful.
Events
Fellowships, hosted sessions & appearances.
Hosted community launch · 18:30 · Madrid, Spain
The AI Fellowship Madrid Chapter Launch
The launch event for The AI Fellowship Madrid Chapter, bootstrapped and hosted by Luka Dujmovic at the Faculty of Computer Science, Complutense University of Madrid. A practical, human-first gathering to explore how AI can help people do better work.
A useful place to start
Where is your AI ambition getting stuck?
Bring an initiative, capability gap, or organizational tension. I can examine what is blocking progress, what evidence is missing, and whether the next move is to deploy, prove, connect, or wait.