Consulting

Less manual work. More system.

I help companies find which processes can be automated, put their data in order and leave it all running. No hundred-page reports: a clear assessment, priorities and, if you want, I build it.

Sound familiar?

Signs there is something to automate.

If you recognise two or more, there are almost certainly hours to win back every week.

  • Someone downloads a spreadsheet from one portal and pastes it into another every morning.
  • The monthly report takes days and, even so, nobody trusts the number.
  • Two departments present the same KPI with different figures.
  • Every client wants their information in a different format and at a different frequency.
  • There is a critical process only one person knows how to run.
  • You pay for AI tools, but nobody is quite sure what to use them for.

Consulting areas

Where I can help.

Six fronts that usually go together. You can start with just one.

01

Process and admin automation

I find the repetitive tasks that eat up hours (downloads, file matching, data entry, recurring deliveries) and turn them into processes that run on their own and tell you when something fails.

You getA process map prioritised by saving and difficulty, and the automated processes with their operating guide.

Done before: five daily manual processes turned into scheduled jobs, and an autonomous system that runs 17 times a day.

  • Inventory of manual tasks and an estimate of recoverable hours
  • Browser robots for portals without an API, even with SSO and two-factor
  • Flows between applications with n8n, Make or Python
  • Retries, alerts and a record of every run
  • Documentation so the process depends on no one
02

Data extraction from websites and APIs

If the information exists on a screen, it can be captured in an orderly, lawful way: prices, catalogues, order status, platform reports or supplier data.

You getA scheduled capture process, with history and quality checks.

Done before: portals behind SSO, OTP, MFA and anti-bot protection, with reliability raised from 30% to 90% or more.

  • Scraping and automation of web portals
  • Connection to third-party APIs (CRM, advertising, telephony)
  • Scheduling and incremental loads
  • Integrity checks on everything downloaded
  • A queryable history for trend analysis
03

Data modelling, management and storage

I sort out where each piece of data lives, how it relates to the rest and which source is the truth, so reports stop contradicting each other. And if the database is slow, I look for the cause with evidence, not guesswork.

You getA documented data model and an analytical base ready to query.

Done before: a critical query from over 90 s to 10 s, and a sales source of truth for some 11 clients.

  • Audit of sources, duplicates and inconsistencies
  • Model design (facts, dimensions, KPI definitions)
  • Storage in BigQuery, MySQL or PostgreSQL depending on volume and cost
  • Reconciliation between systems with traceable evidence
  • Quality rules and automated checks
  • Query and database server optimisation
04

Dashboards and reporting

We agree which questions the dashboard must answer and I build it to refresh on its own. I also rescue Power BI reports that have become slow or unmanageable.

You getA dashboard in use, with fresh data and agreed metrics.

Done before: five duplicated reports merged into one, and a complete B2B dashboard delivered in 4 days.

  • KPIs defined with the business before anything is drawn
  • Power BI (DAX and M) or licence-free web dashboards
  • Optimisation of slow models and duplicated measures
  • Recurring reports sent automatically
  • Training so your team can maintain it
05

Integrations with clients and systems

I connect your tools to each other and to your clients' systems so information travels on its own: from the data warehouse to the CRM, to the contact-centre dialer (OCM, Nextiva) or to the client's system.

You getIntegrations in production, monitored and documented.

Done before: integrations over SFTP, REST API, web service and webhooks with more than 15 platforms, with no duplicates.

  • Consuming and publishing data through APIs
  • Synchronising lists and statuses between systems
  • Recurring deliveries in the format each client requires
  • Duplicate and compliance controls (for example, exclusion lists)
  • Command-line tools for the operations team
06

AI applied to daily work

I use Claude Code every day as a development copilot, with my own rules and reviewers, and I know where it helps and where it does not. I help you bring AI assistants into your team's real tasks, with judgement and care over what data is shared.

You getUse cases in place and a team that knows how to use them.

Done before: an AI working framework for Power BI adopted day to day, and a GDPR-compliant AI usage policy.

  • Identifying use cases with a clear return
  • Classifying and extracting information from documents
  • Automatic generation of reports and summaries
  • Assistants over your own data and procedures
  • Hands-on training and safe-use guidelines

Formats

Four ways of working together.

Start small. If the assessment finds nothing worth doing, I will tell you.

One session

Assessment

A conversation to understand your process or your data problem and tell you what I would do.

  • Review of the case
  • Options and risks
  • Recommended next step
1–2 weeks

Audit and roadmap

I analyse processes, sources and tools, and deliver a plan prioritised by impact.

  • Process and data map
  • Estimated saving
  • Phased plan
By scope

Turnkey project

I build the solution end to end and leave it running and documented.

  • Design and implementation
  • Testing and go-live
  • Handover to the team
Monthly

Ongoing support

Recurring hours to evolve what has been built, answer questions and train the team.

  • Continuous improvement
  • Support and review
  • Training

Method

What it is like to work with me.

Small, frequent deliveries. You see something working early and decide whether we carry on.

  1. Understand

    I sit with whoever does the work today. The real process is almost never the one in the manual.

  2. Prioritise

    We rank by hours saved and difficulty. Highest return first.

  3. Build

    A working version quickly, with validations, retries and alerts from the start.

  4. Hand over

    Documentation and training so it runs without me.

Frequently asked questions

Before you write to me.

Do you work with the tools we already have?

Yes. I start from what you already use (Excel, Power BI, your CRM, your database) and only suggest changing something when the saving justifies it.

We have no technical team. Does it still make sense?

That is the most common case. Solutions come with a plain-language operating guide and automatic alerts, so anyone can tell whether everything is fine.

How do you handle the confidentiality of our data?

A confidentiality agreement from day one, minimum access and credentials always kept on your infrastructure. Before using any AI tool we agree together which data may pass through it and which may not.

What happens if the portal or the API changes and the process breaks?

It happens, which is why every automation carries validations and alerts: if something changes, it is caught on the next run instead of being discovered weeks later.

How much does it cost?

It depends on the scope. After the assessment you receive a fixed proposal with deliverables and deadlines, with no surprises.

Next step

Which process would you like to stop doing by hand?

Tell me the problem in two lines. I'll tell you honestly whether it can be automated, how, and what it would take.