02 · ongoing

An outside brain — for the AI-era web.

I take one or two advisor engagements per quarter. A fractional presence in your team — not a delivery resource. Weekly rhythm, async access, sprints when the moment calls for it.

Web strategyCRO + experimentsAI skills + MCPsMigrationsMulti-market ops
the leverage

Most fractionals show up with experience. I show up with experience plus a working AI playbook — skills, MCPs and context repos your team can adopt. The leverage compounds; the judgment calls stay yours.

what I cover

The areas I work in.

Not a generalist consultant. The work is focused on the web and how modern teams build and operate it.

01
Web strategy

Where to focus web investment, what to ignore, how to prioritise the roadmap. Second opinion on direction, not just execution.

02
Conversion, A/B testing + optimisation

Experiment roadmaps, test prioritisation, reading results without fooling yourself. From velocity testing to full-funnel analysis — and how AI fits into a modern CRO stack.

03
Migration architecture

CMS, platform, stack migrations — vendor selection, architecture decisions, scope management. I've done this at scale.

04
AI tooling + agent ops

Where AI agents genuinely save leverage vs. where they add noise. What a Claude Code-powered workflow actually looks like in practice. How to wire agentic automation into web ops — not as a demo, but as a durable operating model.

05
Team and hiring

Second opinion on web team structure, agency relationships, and hiring decisions. Especially useful if you're scaling the web org for the first time.

06
Localisation + multilingual ops

Running a website across multiple languages and markets — translation workflows, CMS architecture for i18n, QA pipelines that hold up when you add a locale.

the playbook

What I bring beyond opinions.

Most advisors show up with experience and a Notion doc. I show up with a working AI operating layer — already tuned for web work. I help your team pick what fits, wire it in, and run it — taught by walking through the real systems and live examples I use myself, not slide decks. The leverage compounds. The team gets faster, sharper, and more efficient every cycle.

01skills
Pre-built workflows your team adopts.

A growing set of Claude Code skills tuned for web work — audits, QA passes, A/B test scaffolding, content drafts, performance and SEO checks. I help your team pick the right ones, adapt them to your stack, and weave them into the rhythm.

02MCPs
Claude wired into your real stack.

MCPs plug Claude into your CMS, analytics, search console, repo, CRM. Agents stop guessing and start reading your real data — drafting against actual content, opening real PRs, pulling live numbers. I advise on which to wire, in what order, for what payoff.

03context
A company memory agents can use.

Brand voice, ICP, product context, design system, decisions log — structured into a repo your agents read on every task. The compounding asset most teams don't realize they're missing. I help you start one, or audit the one you have.

the playbook in production

Battle-tested at Pipedrive.

I lead the web org at Pipedrive — a global SaaS used by 100K+ businesses across 175 countries. The same operating model I bring to advisor engagements is the one I run there.

Pipedrive
Head of Web · 2021 – now
scale
24
languages live
500+
A/B tests run
20%+
lift on winning tests
1,000s
pages across markets
web partner for
engineeringdesignbrandproduct marketingcontentlocalizationanalyticsgrowth
operating layer · in production
  • A/B testing & CRO
    Skills draft and iterate variants against real company context — funnel pain, ICP, brand voice — while MCPs pull live experiment + analytics data. Variants ship without engineers in the loop.
    skills + MCPs
  • Multi-market QA
    Pre-deploy QA across locales runs as a skill — copy, broken paths, tracking, schema, broken links — before anything ships to a single market.
    skills
  • Web publishing at scale
    MCPs connect Claude to the CMS so review, edits and updates run inside the agent loop — not 50 manual tabs across 24 locales.
    MCPs
  • Company context repo
    Brand, product, ICP, decisions and design system maintained as a structured context source agents read on every task across the org.
    context
how it works

Shape-to-fit, not a fixed deliverable.

Every engagement is different. The default is a weekly sync and always-on async access — but the mix shifts to what's useful. Sprint-heavy when you're in a crunch. Lighter when you just need a thinking partner on call.

Minimum three months. Not because of a contract clause — because the useful work takes time to build context and actually move things.

01
Reach out

Drop a message via the contact page. Brief context on what you're working on and what you're looking for.

02
Intro call

A 30-min call to see if there's a fit. No prep required. I'll ask a few questions, you ask yours.

03
Scope the engagement

If it makes sense, we shape the engagement. Rhythm, focus areas, how async access works.

04
Start

Three-month minimum to start. Renews on mutual agreement.

who this is for

Good fit. Bad fit.

Good fit
  • Founder or VP making web decisions without a dedicated Head of Web
  • Growing a web team and need an outside perspective on direction
  • Mid-market company navigating a platform migration or AI tooling shift
Not the right fit
  • Looking for hands-on delivery or a development resource
  • Need a fixed project with a scope doc and sign-off milestones
  • Not yet invested in web as a growth channel