A practical sort of AI person.

Some work
needs you.
Some just needs
doing.

Hi, I’m Valentina. I help teams use AI for the repetitive bits, so people have more time for the work that needs their judgement.

Let me show you what I mean
A few things from my desk
WORKFLOW NOTES / VM

There has to be
a better way
to do this.

✓ Start with an actual task
✓ Let AI do the repetitive part
✓ Check the work yourself
PEOPLE STILL MAKE THE CALL
IDEAS WORTH TRYING / 001
Less copying.
More thinking.
Built around real workPeople check the resultsOne useful task at a time
A look inside the work

Things I’ve built.
And things worth trying.

A completed workflow, a community project, and an inbox prototype. Different problems, the same curiosity.

01 / My own workflow

I got tired of typing up lab reports.

One report used to mean 10–30 minutes of copying values into an Excel template. I built an LLM workflow to do the entry, then changed my job to checking its work.

Built & used
Explore the change
LAB REPORTEXAMPLE ONLY
SampleResult
Example A0.12 mg/kg
Example B0.08 mg/kg
Example C0.21 mg/kg
Read · copy · enter · repeat
REPORT TO EXCEL TEMPLATE

The numbers didn’t type themselves.

Read a value, find the right cell, type it in. Repeat until the report is in the template, ready to check contaminant levels.

10–30 minManual entry, per report
EXCEL TEMPLATEEXAMPLE ONLY
SamplePrepared result
Example A0.12 mg/kg
Example B0.08 mg/kg
Example C0.21 mg/kg
Upload · AI prepares · I review
PREPARED DATA, READY TO CHECK

Now I check, instead of copy.

I upload the report in seconds. The LLM prepares the Excel file, then I check the values and correct anything that needs attention.

1–2 minHuman review · AI processing time is additional
Based on my own experience, not a promise of client results.Human review stays part of the work.

This is an interactive explanation, not a live AI tool. The sample values are fictional. With AI, upload takes seconds and review takes around 1–2 minutes; processing time is additional.

02 / Inbox prototype

A calmer start
to the inbox.

Sort the messages. Flag what needs attention. Prepare a reply. Let a person decide what actually gets sent.

Open the project
Prototype / Sample data

The inbox, with a head start.

The idea: when someone starts their day, the sorting is already done. Messages are grouped, important requests are flagged, and draft replies are ready to check.

What AI does

Reads the incoming message, categorises it, and prepares a response from approved company information.

What a person does

Checks the draft, makes edits, and handles anything that needs context or care. Sending stays with the team.

Try the existing demo A prototype with sample data, not a client deployment. The demo opens on an external website.
03 / Discord community project

“Could we make
a bot for this?”

A functional roleplay bot for a community server. AI wrote the code. I directed the work and shaped the experience.

A little more about it
Completed / Community project

An idea for the server. A bot that works.

I built a functional roleplay bot for a Discord community using AI-assisted development.

My part in the project

I directed the work and shaped the experience. AI generated the code. My role was guiding that process towards the tool the community wanted.

Why it belongs here

AI isn’t only useful for documents. This project is an example of taking an idea and directing AI to help build something people can use.

The bot runs in its community server. There is no public bot demo on this page.
Valentina MattssonCurious about what could work better.
The person behind the workflows

I started with a task
I wanted off my plate.

Manually entering lab-report data showed me how much of a working day can disappear into repetitive tasks. Building an AI workflow for it changed how I approached the work.

That’s what I want to help others explore: practical ways to use AI, grounded in an actual need. I bring hands-on experience directing AI tools and building workflows, with people checking the results and making the decisions.

Where we could start

You don’t need
a grand AI strategy.
Just a good first task.

Tell me where the time goes. We can work out whether AI would actually help.

Find the right starting point

Walk through your current workflow, identify repetitive steps, and choose a small project with a useful outcome. We agree what AI should do and what a person needs to check.

Get documents into a useful shape

Prepare structured data or spreadsheets from documents, with review steps and a way to handle exceptions. The lab-report project is one example of this approach.

Make the inbox less of a project

Explore sorting, priority flags, and reply drafts based on approved information. Your team reviews the suggestions and approves what gets sent.

Turn meeting notes into next steps

Use a meeting transcript to draft a recap, priorities, and actions. With participant consent, the organiser checks the details and confirms owners before sharing.

Build a small tool. Help people use it.

Try a focused bot or prototype, then create simple instructions and review habits so the team knows what to do with it. Scope and tools depend on the task.

A few things you might be wondering.

Do we need to know anything about AI?

No. We can start with the way you work today and a task you’d like to make easier. You don’t need a technical brief or a particular tool in mind.

Will AI send emails or make decisions on its own?

The workflows centre on assistance: preparing data, highlighting information, and drafting responses. You or your team review and approve the work. Any automation beyond that would need to be explicitly agreed.

What about sensitive company information?

Before choosing a tool, we need to understand what information the workflow uses and your organisation’s requirements. Tool access, data handling, and what may be shared should be agreed before using company data.

Can we start with just one task?

Yes. One repetitive task gives us something concrete to test, improve, and assess before considering a wider rollout.

A conversation is a good start

What would you love
to stop doing manually?

Tell me about the task. What you do now, where it gets frustrating, and what you’d rather spend that time on.

Say hello