AI in Business: The Slides from Our ITALCAM Session
Where does your team lose time searching for information, chasing updates or reconciling figures? These everyday tasks were the starting point for Convolut’s presentation at the ITALCAM Business Breakfast.
Francesco di Cugno, Managing Director of Convolut GmbH, presented “KI im Unternehmen: Drei Wege zum praktischen Nutzen” — AI in business: three paths to practical value. We are sharing the complete slide deck for participants and anyone exploring where AI could support their organisation.
The presentation explores three ways to delegate work to AI: preparing recommendations, executing defined processes and investigating a goal within agreed boundaries. Practical examples include matching skills to project requirements, collecting status updates and examining differences in reporting data.
The discussion connects these examples to a practical question: which task is worth delegating first? A useful starting point is a clearly defined process, accessible information and a person responsible for reviewing the outcome. The slides offer a framework for weighing the potential benefit against feasibility and risk, while keeping decisions and accountability with people.
Contact us with your questions
Have questions about applying AI or improving your business processes? We will be happy to discuss your use case and explore where automation could help your team.
Three Ways to Delegate Work to AI
The presentation distinguishes three approaches. The right choice depends on the task, expected benefit and risk.
- Analysis and recommendations. AI compares information and prepares a recommendation, while a person makes the decision. The skills-matching example shows how project requirements, profiles and project histories can be used to suggest suitable candidates and explain strengths and gaps.
- Intelligent automation. A defined process provides the structure, while AI interprets language, follows up on missing information and coordinates tools within agreed rules. Examples include collecting project status updates, preparing management summaries and handling support requests with escalation to a person when needed.
- Goal-oriented investigation. An agent plans and tests hypotheses within a defined scope. The presentation introduces DeltaLens, a Convolut product, through a bank-reporting reconciliation example: investigating differences, gathering evidence and producing findings for human review, without changing source data.
The skills-matching and process-automation examples are described in the slides as being in development internally and with external users.
Set Clear Boundaries for Autonomy
Delegating more work requires explicit controls. The slides outline design requirements including access permissions, protection of personal data, a defined data perimeter, read-only analysis, approval checkpoints, traceable actions, cost and runtime limits, and stop conditions.
These are requirements to establish and verify for each use case. People remain responsible for setting objectives, reviewing results and deciding what action to take.
Choose a Focused First Use Case
Start by assessing three questions: how much time or quality could the use case improve, are the necessary data and systems accessible, and what would happen if the result were wrong?
A practical starting point has a bounded process, a measurable objective and a named person who reviews the outcome. Start small, demonstrate value and expand based on the evidence.
Thank you to ITALCAM for hosting the Business Breakfast and to everyone who joined the session.
HOW WE CAN HELP YOU FURTHER
Related Services
Data Analysis
Dashboard Design
Data Governance
Data Architecture
Calculation Engines for Finance
About the Author
