FRACTIONAL AI OFFICER

Most AI projects
never ship. Yours will.

Assessment, roadmap, and hands-on implementation. I work embedded in your team to build AI systems that run in production, not in a slide deck.

FRACTIONAL AI OFFICER  ·  PROCESS AUTOMATION  ·  N8N  ·  MICROSOFT CRM  ·  WHATSAPP BUSINESS API  ·  LLM INTEGRATION

Most AI initiatives fail not because the technology is wrong, but because no one owns the gap between a working prototype and a system running in production. Strategy without execution is the most expensive kind of consulting.

I work as a Fractional AI Officer: embedded in your team for the duration of the engagement, responsible for identifying the right automation opportunities, designing the architecture, and implementing the systems. From assessment to deployment.

AI Opportunity Assessment

Map where AI creates real leverage in your operations or product. Prioritized by impact and feasibility, not by what's trendy. Now a fixed-price product: €450, one week, guaranteed.

Implementation Roadmap

A prioritized plan with tools, timelines, and build-vs-buy decisions. No generic framework, specific to your stack and your team's capacity.

Fractional Execution

Hands-on implementation using the right stack: n8n for automation workflows, CRM integrations, WhatsApp Business API, and custom LLM pipelines. I own the build, you own the system.

Quick fit check

Would an assessment find 5 hours in your week?

Four questions, thirty seconds. An honest first read before you spend anything.

How much of your week goes to work a template could do?

How repeatable is that work?

Where does the information live?

Who would own the change once it exists?

Indicative only. The assessment measures your actual hours rather than estimating them.

01 / Start here

The AI Opportunity Assessment.
Your week has 5 hidden hours.

Fixed price · One week · No commitment beyond it

A structured diagnosis of where your business loses time, and the exact tools that get it back. I interview you, I do the research, you get a prioritized action plan. Like a doctor's visit for your operations: you leave with a prescription, not a proposal.

€450 flat, all included
1

Discovery interview

A structured conversation about your typical week. We find where time actually leaks: email, scheduling, quotes, reporting, follow-ups.

45 min · video or in person

2

Analysis

I cross your bottlenecks against hundreds of proven AI and automation tools, filtered by what I have seen run in production. No experiments on your time.

zero effort on your side

3

The report

3 to 7 recommendations, each with the tool, monthly cost, setup time, and hours saved. Plus an effort-vs-impact map, a 4-day quick start plan, and your net ROI in euros.

yours to keep

4

Review call

We walk through every recommendation together. You decide what to switch on yourself and what, if anything, to hand over to me.

30 min · video call

The guarantee

If the assessment does not find at least 5 hours per week your business can reclaim, you get a full refund. Worst case, you have lost 75 minutes of your time. Best case, you get a working week back every month.

In person in Barcelona · Remote across Europe

The report is a diagnosis, not a sales pitch. Most recommendations are off-the-shelf tools you can switch on yourself in a day. Where a recommendation needs real engineering, custom workflows, integrations, LLM pipelines, that is a separate engagement: the Fractional Execution work described below.

02 / Build

Then it gets built.
By the person who scoped it.

The assessment ends with a prescription. Most of it you fill yourself, and that is the point. What is left is the work that needs an engineer: the integrations, the custom workflows, the pipelines that touch your real data. That is this. It is a separate engagement, priced separately, and you are free to take the report and give the build to someone else.

Per project

A defined scope with an end date: process audit, automation architecture, build, handover. The dental clinic below was three months, from audit to a live system handling 3,000 patients.

TYPICALLY 2 TO 4 MONTHS

Monthly retainer

Set days each week inside your team for as long as the roadmap runs. For companies with a sequence of things to automate rather than one workflow, and nobody in-house to own the architecture.

ONGOING, FIXED DAYS

OWNERSHIP

Everything runs on your accounts and your stack, documented. I build it, you own it. If we stop working together nothing switches off, and there is no licence of mine left to keep paying.

n8n · Microsoft CRM · WhatsApp Business API · custom LLM pipelines

Talk about a build Most builds start from an assessment. They do not have to.

03 / Fit

This is for you if

You are building an AI-native product and need a technical partner who ships, not a consultant who presents options.
You run an SME and know automation could save your team hours per week, but you do not know where to start or which tools to trust.
You have tried AI tools but nothing has made it to production. Prototypes exist, real systems do not.
You want to automate a specific process: customer communication, data extraction, internal workflows, or lead qualification.

04 / In action

Automation that runs
in production.

Case study · Clinica Dentale · Italy · 2024

-20% No-show rate
8 min Saved per patient
3,000 Patients automated
3 mo Full deployment
Problem A dental clinic with 3,000 active patients was managing check-in, appointment reminders, and post-visit follow-up entirely by hand. Staff were spending 8+ minutes per patient on administrative coordination. No-show rates were eating into revenue and scheduling efficiency.
Action 3-month engagement covering process audit, automation architecture, and full implementation. Built automated patient communication flows using WhatsApp Business API, integrated with Microsoft CRM, and orchestrated with n8n. Covered appointment confirmations, reminders, post-visit follow-up sequences, and reactivation of inactive patients.
Outcome No-show rate down 20%. 8 minutes of staff time saved per patient interaction. All 3,000 patients enrolled in automated sequences. The system runs without manual intervention.

3

months to full deployment

assessment to production

3,000

patients on automated flows

zero manual intervention

-20%

no-show reduction

measurable from month one

FREQUENTLY ASKED QUESTIONS

How do you measure the ROI of an AI automation project?

Start with hours, not euros. Measure how long a process takes today, across how many people and how many times a month, before you change anything. That baseline is the only honest denominator, and almost nobody has it. After automation, measure the same process the same way. The difference in hours, multiplied by a loaded hourly cost, minus tool subscriptions and setup time, is your net return. Anything else, including model accuracy and volume of tasks processed, is a vanity metric.

How soon should automation show results?

If a workflow automation has not paid back the time spent building it within 90 days, it was probably the wrong workflow. Off-the-shelf automation of a repetitive, well-defined process usually shows measurable hours back in the first two to three weeks. Anything promising returns on a 12-month horizon is either a genuine engineering project, which should be scoped and priced as one, or a way of deferring the moment when someone checks whether it worked.

What should we automate first?

The process that is boring, frequent, rule-based, and currently done by your most expensive person. Frequency matters more than complexity: a five-minute task done forty times a week is worth more than a two-hour task done monthly. Avoid starting with anything customer-facing or anything where a mistake is expensive to unwind. The first automation should be one where failure is visible immediately and costs nothing, because the point of the first one is to build confidence, not to save the most time.

Is it better to use one AI partner end-to-end, or separate firms for strategy and build?

Separating them removes a conflict of interest. A firm that diagnoses and then builds has an incentive to find problems that require building. That is why this assessment is deliberately structured the other way round: the report names off-the-shelf tools you can switch on yourself, and engineering work is a separate engagement you are free to give to someone else. For a small company, one partner end-to-end is usually faster and cheaper. Just make sure the diagnosis was not written by the person quoting for the build.

How do I know whether an AI consultant is actually qualified?

Ask what they would refuse to automate, and why. Anyone who answers "nothing" is selling. Ask for a process they assessed and recommended against changing. Ask how they measure a baseline, and whether their fee depends on you buying a build afterwards. Credentials help, but the useful signal is whether they can describe a failure in specific terms. Vague enthusiasm about transformation is the tell, in this field more than most.

What happens if the assessment does not find 5 hours a week?

You get a full refund. The guarantee is on the assessment identifying at least 5 hours per week of recoverable time, so the downside is 75 minutes of your time: a 45-minute discovery interview and a 30-minute review call. It has to work that way, because a fixed-price diagnosis is only credible if the person selling it carries the risk of finding nothing.

Ready to get your 5 hours back?

Start with the €450 assessment, or with a free 30-minute discovery call.