Applied AI and AI agents to reduce repetitive work.
I help businesses identify repetitive, time-consuming or error-prone tasks and turn them into simpler, better-organised and more reliable processes.
I don't start with the tool. I start with the process: I understand how the work is done today, identify where AI can help safely and implement automations, assistants and agents integrated with the tools your business already uses.
AI only makes sense when it removes concrete work.
I work on situations such as requests, administrative tasks, scattered information, frequent replies, content preparation and internal workflows.
I don't recommend AI just because it is trendy. I first check whether there is a clear process, a verifiable result and a safe way to implement it. If the problem is simply connecting tools, see process automation.
What I deliver
Automations and agents with rules, limits and oversight.
I define the process, the data required, permissions and validations, and only then build the solution, integrated with your tools.
Process and rules
Definition of the process, data, permissions, exceptions and validations.
Automations and agents
Implementation integrated with email, forms, CRM and systems.
Classification and routing
Organising requests, messages and information.
Replies and content
Preparing replies, summaries and proposals, with human review.
Information organisation
Turning scattered data into useful information for the team.
Ongoing support and improvement
Documentation, adjustments and continuous improvement of the solution.
Where AI helps in practice.
Classify and route requests
Prepare initial replies
Organise scattered information
Create content, summaries or proposals
Support approval and validation workflows
Integrate tools and systems
Create internal assistants
Automate follow-up, alerts and records
I start with the process, not the tool.
Understand the process
I identify the task, who performs it, the systems involved and the intended result.
Assess the real potential
I distinguish what can be automated, supported by AI or decided by people.
Define rules and limits
I define data, permissions, exceptions and human oversight points.
Implement in stages
I build and test the solution in a controlled scope, starting with the essentials.
Document and support
I deliver documentation, explain the limits and support the solution.
AI with control, not AI without rules.
I don't recommend AI just because it is trendy.
I don't automate sensitive decisions without human oversight.
I define access, permissions and usage limits.
I protect data, confidentiality and boundaries between systems.
I document how it works and explain the limits and risks.
I keep people responsible for important decisions.
It is for you if...
You have repetitive tasks that take up your team's time.
You receive requests, emails or information through several channels.
You want to use AI but are worried about exposing data.
You have tried AI but don't know how to integrate it safely.
You want to support the team, not replace people.
You want a documented, controlled solution that is easy to evolve.
You want to reduce manual work without losing control.
Common questions.
Does AI replace employees?
No. The AI applied here is meant to reduce repetitive tasks and support the team, not replace it. Sensitive decisions are still made by people.
What kind of tasks can be automated?
Repetitive tasks with clear rules: classifying requests, preparing replies, organising information, summaries, proposals and updates between systems.
Does the AI access all company data?
No. Access is defined by scope: only the data needed for the task, with permissions, limits and a log of actions.
Are replies sent automatically?
It depends on the process. In sensitive cases, the reply is prepared and held for human review before being sent.
Which tools do you use?
It depends on what the company already uses. I integrate with email, forms, CRM, spreadsheets, databases and internal systems, with alternatives when an integration is not feasible.
What happens if the solution does not work as expected?
The solution is tested in a controlled scope before being used in production, with clear limits and human validation points. If it fails, it can be stopped or corrected without affecting the main process.
How does the project start?
With a technical assessment: I review the current process, identify what can be improved and explain honestly what makes sense to automate and what does not.