For development, operations and DevOps teams
AI coding assistants that work for the whole IT team
Most IT teams already use AI assistants, each person in their own way. With my own team and a client’s whole IT department, I tried several approaches, and ran into several pitfalls, before we found a way of working that lasts. I help your teams get there faster: Claude Code set up for the team, clear rules for access, data and review, and training on your own tasks.
Typical situations
AI assistants spread faster than the rules for using them.
Everyone uses AI, each in their own way
Different tools, personal accounts and personal settings: results vary from person to person, and nobody can say which company data went where.
The assistant can do more than it should
Nobody has defined what the assistant may read, run or change, so credentials and production systems are only a command away.
Reviews can’t keep up
AI-assisted changes arrive faster and larger, while review and testing stay as they were.
Pilots that never become practice
A few licences and a few enthusiasts, but no shared way of working, so the rest of the team never follows.
Training that stays abstract
General AI courses that never touch the team’s real tasks, repositories or infrastructure.
Case study: AI adoption across an IT department
A European technology company (a current client, not named)
Since July 2025 · DevOps team lead
I lead a DevOps team at the company and am closely involved in how AI is used across its IT department. Together with operations and development, we tried several approaches and ran into several pitfalls before we found a way of working that suits the teams and the company. Claude Code is now a shared, everyday tool.
The starting point
AI assistants were in use, but each team and each engineer worked with them differently, with no common rules for access, data or the review of AI-assisted changes.
What I did
- Led the introduction of Claude Code in the DevOps team
- Worked with operations and development on a common approach for the whole IT department
- Tried approaches in practice, kept what worked and dropped what did not
- Shared rules for access, data and the review of AI-assisted changes
- Trained the team on its own tasks
How the work is organised
Three phases, each with a written result. You decide after every phase whether to continue.
Assessment
Where AI already helps, and where it adds risk.
- Conversations with the teams and a look at how AI is used today
- Use cases in development, operations and DevOps, ranked by benefit and risk
- Data and access: which systems and data the assistant may reach
- A written recommendation: setup, rules and the order of the rollout
Setup and rules
Tools configured for the team, not for one person.
- Claude Code on a Team or Enterprise plan, or through Amazon Bedrock, with team-wide settings
- Shared project instructions, permission rules and review rules
- Credentials and production systems kept out of the assistant’s reach
- A short, practical usage policy for IT
Training and follow-up
People who use the tools well.
- Hands-on sessions built on the team’s own tasks
- Separate sessions for development, operations and DevOps where useful
- Guidelines and examples the team keeps using
- A follow-up review after the first weeks
Scope and working model
Configuration, not custom development
I set up and configure the tools and the rules around them. I don’t build custom AI applications or agents.
Data protection in mind
Each recommendation states which data may go into which tool, and on which plan.
Location and hours
Based in Alicante, Spain, working remotely with teams across Europe and beyond, on Central European Time (CET/CEST). Sessions in English or Spanish.
Confidentiality and access
A non-disclosure agreement is in place before any system access. Access is least-privilege and time-bound, and no production change is made without your agreement.
Outside the scope
- Software development: building AI applications or agents, or changing your product code. That work stays with your developers; I provide the setup, the rules and the training.
- Legal advice on data protection or AI regulation.
I work as an independent consultant on AWS infrastructure, Kubernetes, CI/CD and the team-wide use of AI coding assistants. See experience
Frequently asked questions
Which AI tools do you work with?
Claude Code and Claude. That is where my day-to-day experience comes from, so that is what I set up and teach.
Is this only for developers?
No. Operations and DevOps teams work with code too: infrastructure code, scripts, pipelines, troubleshooting and documentation. The rules for access and production systems matter even more there.
Can our data stay in the EU?
Claude Code can also run through Amazon Bedrock in your own AWS account, with requests kept to EU Regions. The assessment compares the options for your data and your requirements.
Do you build AI applications or agents?
No. I configure and introduce the tools and train the teams; building custom applications stays with your developers.
Our developers already use AI. Why a team setup?
Shared instructions and rules make the results consistent and reviewable, keep credentials and production systems out of the assistant’s reach, and let new team members start with what already works.
Is the training available in Spanish?
Yes. Sessions are held in English or Spanish, remotely.
How much work is it?
That depends on the number of teams and tools. The written recommendation after the assessment states the effort of each step before you commit to it.
Schedule a consultation
Describe briefly which IT teams use or want to use AI, how they use it today, and what you would like to change. Please do not send passwords or confidential data through this form.