Fractional AI

    An AI consultant who joins your team and does the work

    I work onsite in your business for five to six months. I learn how your team works and build the automations that save it time. Then I train your people to run them.

    Book a call

    What is a fractional hire?

    A fractional hire is a consultant who joins your team. I work onsite for part of the week, for five to six months. I get to know the people, the systems and the jobs that eat the day.

    Most consultants hand you a strategy document. I do the work.

    The days are flexible. I do not work a fixed nine-to-five. You and I agree a pattern that suits your team, usually two to three days a week.

    A typical consultant

    • Interviews your team, then works from somewhere else.
    • Hands you a strategy document and leaves the build to your team.

    A fractional hire

    • Works onsite with your team and learns how the business runs.
    • Builds the automations and gets them into daily use.
    • Trains your people to run the systems.
    • Writes the documentation before the term ends.

    Why it works

    A fixed term instead of a permanent hire

    A full-time hire means recruitment, payroll, pension and a notice period. This is one monthly invoice for five to six months, with an agreed end date.

    Long enough to finish the build

    The first month maps the problems and delivers a visible win. The months after build the bigger fixes and get them used. A review at the halfway point checks the priorities still hold.

    Your team learns to run it

    Everything I build comes with training and documentation. When the term ends, the capability stays in the business.

    Breadth you cannot hire for

    One employee is usually strong at one thing. I bring automation, AI tools, workflow design, data pipelines and content systems. I have run a business, so I know where the hours go.

    From first call to handover

    The call and the discovery day are free.

    00. Free

    Dicky working at his laptop

    Discovery call

    Thirty minutes on how your business runs. I look for a fit. If I cannot help, I will say so.

    01. Onsite

    An impact and effort map with sticky notes, the kind a discovery day produces

    Discovery day

    I spend a day with your team. I talk to the people who do the work and map where the time goes. You keep the written report whether or not the engagement goes ahead.

    02. Within 48 hours

    A mock proposal page with builds ranked by impact, a timeline and a monthly fee

    Proposal

    The builds ranked by impact, with what each one saves and what it needs from your team. A plan for the five to six months, and the monthly fee.

    03. Five to six months

    Part of a 52-node n8n workflow, nodes coloured by type

    Build, train, hand over

    I work alongside your team on agreed days. The first automation goes live in month one. By the end, your team runs the systems and has the documentation to keep them going.

    What I typically fix

    These problems come up in almost every business I work with.

    Enquiries that take too long to answer

    A customer fills in a form and nobody replies for a day. By then they have called a competitor. I build systems that reply in seconds, qualify the lead and route it to the right person.

    Information typed twice

    Data goes from an email into a spreadsheet, then into another system, then onto an invoice. I connect the systems so the data moves once.

    Follow-up that does not happen

    Leads go cold because nobody chased them. Past customers sit in a database. I build sequences that keep the conversation going.

    Knowledge stuck in people's heads

    When one person is off, nobody knows the process. I build searchable knowledge bases so the answer is always there.

    Reports compiled by hand

    Someone spends hours each week pulling numbers into a spreadsheet. I automate it so the figures arrive where the team already looks.

    Marketing that never gets done

    Owners know they should post and email. Operations fill the day instead. I build content pipelines that draft the posts and emails for them.

    Examples of workflows I have built

    Admin

    Live

    An event hire company

    A folder for every job

    Saves about 3 hours a month

    Each enquiry gets its own shared folder, and the job's files copy in every hour.

    Before
    Site plans and paperwork were spread across inboxes and the hire system.
    Now
    Each enquiry gets a Google Drive folder. Files copy in each hour, and every crew shift links to it.

    n8n, Google Drive, Current RMS

    Marketing

    Built

    A gift ecommerce brand

    Blog post builder

    Saves about 12 hours a month

    A short brief becomes a full blog post, written section by section and published to the site.

    Before
    Each post took most of a day to outline, write, format and upload.
    Now
    A form brief produces an outline, a draft of each section and a published post, then the team gets a message.

    Make, Typeform, ChatGPT, Google Docs, Contentful, Slack

    Admin

    Live

    A car leasing firm

    Complaint evidence packs

    Saves about 6 hours a month

    One email with a registration returns the vehicle's full history in date order, each line linked to its source.

    Before
    A complaint meant hours in the mailbox, the customer folder and the service records.
    Now
    An email with the registration returns the history in date order and links each line to its source.

    n8n, Outlook, SharePoint

    Admin

    Live

    An event hire company

    Crew shifts and pick lists

    Saves about 5 hours a month

    Invoicing a job drafts the crew shifts, each with its pick list as a tickable task.

    Before
    Staff keyed each job into the shift app one at a time, and crews loaded from paper pick lists.
    Now
    Invoicing a job drafts four shifts within the hour. Each shift carries the pick list, and operations get a summary email.

    n8n, Connecteam, Current RMS

    Marketing

    Live

    My own business

    Daily AI content digest

    Saves about 12 hours a month

    Reads two AI newsletters each morning, summarises every new item and turns the best into post ideas.

    Before
    I read each newsletter in full and kept post ideas in my head.
    Now
    The workflow emails me one summary of every new item at 7am, tests up to three items a day for a post angle and files the good ones in my ideas list.

    n8n, Gmail, Claude API, Airtable

    Admin

    Live

    My own business

    Error alerts

    Saves about 3 hours a month

    Tells me the moment a workflow fails, with the step that broke and a link to the run.

    Before
    I missed failed workflows until someone pointed out the missing work.
    Now
    The workflow sends me the failed step, the error and a link to the run by email and Telegram.

    n8n, Gmail, Telegram

    Sales

    Live

    My own business

    Feedback and testimonials

    Saves about 2 hours a month

    Chases anyone who has not answered a feedback request and files every quote for approval.

    Before
    I asked for feedback by email and copied the replies into a spreadsheet.
    Now
    The workflow sends one reminder after three days, files each reply as an unpublished testimonial and tells me it is in.

    n8n, Resend, Airtable

    Admin

    Live

    A car leasing firm

    Fines and notices into the register

    Saves about 15 hours a month

    Scanned parking, speeding and toll notices update the register and get a drafted challenge.

    Before
    Staff opened, renamed, filed and retyped each notice, then wrote the challenge by hand.
    Now
    Staff scan the letters and send one word. The system updates the register and drafts the challenge, one email per letter.

    n8n, Claude API, Outlook, Excel

    Sales

    Live

    My own business

    Guide requests and website leads

    Saves about 3 hours a month

    Sends each guide the moment someone asks, and files every website enquiry as a lead, with an alert for each scorecard result.

    Before
    I sent each guide and logged each enquiry by hand.
    Now
    The workflow emails the guide, adds the person to my mailing list, files each enquiry in Airtable and alerts me to each scorecard result.

    n8n, Resend, Airtable

    Finance

    Live

    A car leasing firm

    Inspection reports into payment requests

    Saves about 5 hours a month

    Each vehicle inspection report becomes a draft request for payment, with a warning on odd figures.

    Before
    Staff copied each damage line and price into a bill template, line by line.
    Now
    Each report becomes a draft request for payment, with a red warning when a figure looks wrong.

    n8n, Claude API, Outlook

    Data

    Built

    A job seeker

    Job board scanner

    Saves about 20 hours a month

    Checks about fifteen job boards and feeds, scores each new role and files it in one list.

    Before
    The job seeker checked every board by hand, and the same role turned up several times.
    Now
    The scan adds new roles to one Airtable list each day, scored against what the job seeker wants.

    Make, RSS, ChatGPT, Airtable

    Data

    Live

    • Python
    • Apify
    • Claude API
    • Airtable

    My own business

    Job search system

    Saves about 20 hours a month

    Finds roles every week, scores each one against a profile and writes a research brief on the ones approved.

    Before
    Searching job boards and reading every advert in full.
    Now
    A weekly search scores each role. Approving one in Airtable sends a research brief to the inbox within the hour.

    Finance

    Live

    A car leasing firm

    Monthly admin charges

    Saves about 4 hours a month

    A month-end list of chargeable offences, which a person checks against payments before anyone is charged.

    Before
    Someone counted offences across a register of over 4,500 rows.
    Now
    A list arrives on the last day of each month and is checked against payments. Nothing is charged automatically.

    n8n, Excel

    Sales

    Live

    A car leasing firm

    New applications and affordability checks

    Saves about 15 hours a month

    The system files each application, lists the missing papers and writes a summary for a person to decide on.

    Before
    Someone read about thirty pages of bank statements per applicant and chased missing paperwork from memory.
    Now
    The system files each forwarded application, lists what is missing and writes a summary for a person to decide on.

    n8n, Claude API, SharePoint

    Sales

    Built

    • Python
    • Companies House API
    • Claude API
    • Airtable

    My own business

    Outreach research

    Saves about 10 hours a month

    Pulls local companies from Companies House and builds a digital snapshot report on each one.

    Before
    Researching each prospect by hand before a first email.
    Now
    One command collects a company's details, scores its online presence and renders a report page.

    Finance

    Live

    An event hire company

    Payment alerts and job tags

    Saves about 3 hours a month

    Each payment in the accounts software tags the job as paid and emails the team.

    Before
    Someone checked the bank and shared screenshots to tell sales a client had paid.
    Now
    Each payment emails the team an alert and tags the job as paid, deposit paid or part paid.

    n8n, Xero, Current RMS

    Data

    Built

    A gift ecommerce brand

    Product categoriser

    Saves about 17 hours a month

    Sorts new products into the shop's categories in batches and writes the result back.

    Before
    Someone read each product and picked its categories by hand.
    Now
    The workflow categorises products in batches and leaves anything it is unsure of for a person to check.

    Make, ChatGPT, Airtable

    Marketing

    Built

    A gift ecommerce brand

    Product description writer

    Saves about 13 hours a month

    Reads each product's details and images, then writes the copy and posts it to the shop.

    Before
    Someone wrote product copy one item at a time from the supplier's notes.
    Now
    The workflow drafts copy for each new product from the shop's own data, ready for a person to check.

    Make, Shopify, ChatGPT, Airtable

    Data

    Built

    A gift ecommerce brand

    Product link reader

    Saves about 8 hours a month

    Takes any product link and returns the name, price and images, from Amazon or a shop's own site.

    Before
    Someone copied product details from each web page by hand.
    Now
    A pasted link returns the product's details in a clean format, ready to use.

    Make, Apify, Make AI Toolkit

    Marketing

    Live

    My own business

    Short-form video analyser

    Saves about 6 hours a month

    Takes an Instagram or TikTok username, pulls the account's recent videos and breaks each one down shot by shot.

    Before
    I watched other creators' videos and noted each shot by hand.
    Now
    The workflow scrapes the account's latest videos, writes a shot list for each into Airtable and, on request, adds a hook analysis and messages me on Telegram.

    n8n, Apify, Gemini, OpenAI, Airtable, Telegram

    Marketing

    Built

    A gift ecommerce brand

    Social content scraper

    Saves about 12 hours a month

    Collects Instagram posts in the brand's space and has AI describe what each one shows and why it works.

    Before
    Someone scrolled Instagram and noted the posts worth copying in a spreadsheet.
    Now
    The scrape sorts posts into images, videos and carousels, each with a short write-up in Airtable.

    Make, Apify, ChatGPT, Airtable

    Sales

    Live

    An event hire company

    Website enquiries into the CRM

    Saves about 7 hours a month

    Each website enquiry becomes a customer record and a draft job in the hire system, with no retyping.

    Before
    Staff retyped every website enquiry into the hire system by hand.
    Now
    A customer record and a draft job appear within minutes. The system retries rejected enquiries three times each working day.

    n8n, Current RMS

    Marketing

    Live

    • Claude
    • Airtable
    • Python

    My own business

    Weekly content pipeline

    Saves about 24 hours a month

    Turns one weekly theme into research, an interview and a set of posts, checked for voice and facts.

    Before
    Every post started from a blank page.
    Now
    Three commands take a theme from research to a lead post and its cuts. A fact-checker and a voice check run before anything is published.
    See every workflow

    What it costs

    £3,500 to £4,500 a month

    Subject to discussion

    A month is four weeks of work. The figure depends on the length of the commitment and the days each week.

    The fee covers my time. Any new software runs on your own accounts. You pay for it, and you keep it when the term ends. I pick the cheapest tools that do the job. Most businesses spend £50 to £100 a month on them.

    FAQs

    See every workflow I have built

    See if this fits your business

    Thirty minutes on how your business runs and where the time goes.

    Open booking in a new tab

    Where is your team?

    Twelve questions, three minutes. You get a score, the rung your team is on, and the one change worth making this week.