Darren Ferneyhough

United Kingdom (London area)

Darren Ferneyhough

AI Solutions Consultant

I turn business problems into working AI solutions. Nearly 20 years of enterprise consulting across government, financial services and global tech, now combined with hands-on delivery of production apps, data pipelines and agent workflows.

Claude & Claude CodeChatGPT & CodexMCP & agentsSalesforce AI Associate
~20Years in enterprise consulting
400+Logged AI agent work sessions
612Automated tests passing on one AI-built app
50,000+CRM records built and maintained
15+Production sites live

Figures checked November 2026.

Profile

About

I've spent nearly 20 years making enterprise platforms work for real organisations, including the Financial Conduct Authority, Google and the UK Cabinet Office. That work taught me how to run discovery with senior stakeholders, respect governance and hand over systems people can actually own.

I now apply the same discipline to AI. I use Claude, ChatGPT and their agent tools every day to design, build and run real systems - and I stay accountable for the result, from the first conversation to what happens in production.

Approach

How I deliver with AI

  • Start with the business problemEach build begins with who it serves and what has to change. Requirements are written down, decisions are recorded and scope is agreed before anything goes live.
  • Keep people in controlAnything public or hard to undo sits behind a human approval step. Social posts wait in an approval column, and new automations start in dry-run mode with a kill switch.
  • Check the output, not the confidenceAI drafts are treated as untrusted until verified. That means automated tests, source links, checks against fabricated data and code, not a model, for exact comparisons.
  • Measure cost and yieldI track what each method returns and what it costs, then change course on the evidence - such as replacing per-event automation with a daily batch sync when task costs stopped making sense.
  • Leave something you can ownRunbooks, handover notes and reusable playbooks are part of delivery, which is how I have handed over client platforms throughout my career.

Selected work

The AI work on this page is what I have built and run in my own ventures, so you can see real systems rather than slideware. The organisations that have contracted me are listed further down.

Flagship build - founder project

Nutrition Counts: a complete product launch, delivered with AI

A global nutrition app that scans food, plans targets and tracks intake on iPhone and Android. I led it from strategy to store submission through AI agents, with separate agent workstreams for the app, launch content and PR that shared briefs and facts.

291

App commits since Oct 2026

612

Automated tests, all passing

44,000+

Lines of TypeScript

5

Claude-powered API features

The app

A React Native and Expo app on Supabase, with product scoring built on published nutrition methods and nutrient data from CoFID and USDA. Server routes use Claude to read labels and menus, import recipes, create meal plans and power a coach. Now in the app store review process.

The website

A Next.js marketing site handling the waiting list, legal pages and app APIs, plus a dated comparison of competing apps with every claim tied to a checked source.

The launch content engine

A phased content plan with brand guardrails, feeding 110+ planned posts. Remotion templates render the video from data, a Google Sheet holds the calendar and approvals, and Zapier publishes only what has been approved.

The PR agent

A dedicated agent workstream researched a 68-contact media list and drafted the campaign plan, press release and tailored pitches from a single fact brief. Unconfirmed claims are marked VERIFY until checked.

Claude CodeClaude APIExpoReact NativeSupabaseNext.jsRemotionZapierVercel
Visit nutritioncounts.com

More case studies

Different problems, same discipline

Rapid delivery at scale

A portfolio of tailored sites for regulated professions

10

Audience-specific sites

5 weeks

First to tenth site

6

Checkout currencies

The challenge

A health-tech product needed its own conversion site for each UK professional audience, with messaging that respected different regulators, plus a separate multi-country consumer store.

What I did

  • Built a reusable Next.js template, then used AI to tailor research, copy and sourced evidence for each profession.
  • Integrated Stripe checkout, Google Calendar booking with Meet links, WhatsApp alerts and analytics.
  • Traced a booking failure to a missing Workspace delegation permission, fixed it and rolled fixes across every sibling site.
Claude CodeNext.jsStripeGoogle Calendar APITwilioGA4

Data engineering with AI agents

A verified CRM dataset of UK health and wellness professionals

53,600

Contacts given inferred location

2,455

Verified emails recovered

442

Duplicate merges completed

The challenge

Reliable market data was scattered across professional registers and directories, with inconsistent formats, duplicates and missing contact details.

What I did

  • Ran AI-assisted pipelines per profession, with cloud research waves and every CRM write made locally behind a lock and ledger.
  • Added checks that trace each email to a real source page naming that person, and used code rather than models for matching and de-duplication.
  • Measured the yield of each source so the next round went where it paid back, then captured the method as a reusable playbook.
Claude CodePythonFreshsales APICloud agentsGoogle Places

AI operations

Scheduled agents that keep the business running

294

Claude Code sessions logged

109

Codex sessions logged

6

Reusable agent skills written

The challenge

Routine operational work - email authentication reports, campaign data and failed deliveries - was slow to do by hand and expensive to automate event by event.

What I did

  • Set up a daily agent to read and act on DMARC reports, and a batch sync of email campaign data into the CRM that replaced costly per-event Zaps.
  • Turned delivery failures into CRM fixes, with typos corrected and unresolved addresses queued for enrichment.
  • Built a private read-only MCP worker on a local open-source model for low-risk summarising and drafting, keeping final decisions and sensitive data out of it.
Claude CodeChatGPTCodexMCPOllamaTailscalePython

Evidence-led research agent

Applying rules of evidence to thousands of family history records

1,174

People in the tree, up from 585

3,810

Hints in the starting queue

3,700+

Lines of audit log

The challenge

A personal project with a useful test built in: thousands of historic records, many of them ambiguous, where a confident wrong answer would quietly corrupt the whole family tree.

What I did

  • Gave the agent explicit rules: accept a record only if names, dates, places and relationships all agree, reject on clear mismatch and log anything ambiguous for review.
  • Kept a full written audit log, and investigated a claimed royal descent until the evidence closed it as undocumented.
  • Published a one-name study site showing deceased people only, keeping living relatives private.
Claude CodeBrowser agentsCustom scriptsAudit logging
Visit ferneyhough.com

Services

How I can help your organisation

AI opportunity discovery

Workshops with your teams to find where AI will pay back, and where it will not.

Rapid prototypes to production

Working software in weeks, with the tests, guardrails and documentation to keep it running.

Agents and workflow automation

Scheduled agents and integrations across your existing tools, with approval steps built in.

Data enrichment and quality

Pipelines that gather, verify, de-duplicate and load data, with a full audit trail.

AI on your Salesforce platform

Practical AI planning backed by deep Salesforce experience and Agentforce training.

Governance and adoption

Sensible rules for data, review and access, plus training so your people use it well.

Capabilities

Core skills & technical toolkit

AI Solution Design

Use-case discoveryAgent workflow designPrompt & context engineeringHuman-in-the-loop controlsModel selectionCost & yield tracking

AI-Assisted Delivery

Full-stack web appsiOS & Android appsAPIs & integrationsData pipelinesAutomated testingScheduled agents

Consulting & Leadership

Stakeholder workshopsC-level engagementRoadmaps & backlogsAgile deliveryGovernance & securityTraining & handover

AI tools

ClaudeClaude CodeClaude APIChatGPTCodexMCPOllamaAgentforcePrompt Builder

Platforms

Next.jsReact NativeExpoSupabaseVercelPythonStripeTwilioZapierRemotionSalesforceFreshsalesGitHub

Track record

Some of the organisations that have contracted me

  • Financial Conduct Authority
  • UK Cabinet Office
  • Department for Digital, Culture, Media & Sport
  • Google
  • eEnergy plc
  • Operational Risk Consortium

Credentials

Certifications & training

AI

Salesforce Certified AI Associate
Google Certified AI Professional
Google AI Essentials
Google AI for App Building
IBM Certified - AI Fundamentals

Platform, security & delivery

5x Salesforce Certified
ISC2 Certified in Cybersecurity
SC security clearance held on a 2023 contract
Google Foundations of Project Management
Gearset Salesforce DevOps Fundamentals

What people say

Recommendations

Excerpted from LinkedIn recommendations from clients and colleagues.

“Darren has been an excellent manager during our time together at the FCA. I've found him incredibly easy to work with. He is supportive without ever falling into micromanagement and he's allowed me to thrive in my role.”

Simon Owen

Project Manager at Cognizant

“You notice such people with no ego and like them immediately... The thing I was impressed the most is his willingness and openness to looking at the things differently.”

Alla Fadeeva

Product Owner, Agile Leader

“Darren is an incredible team player, and always puts the team ahead of his own needs. When he takes on a task he seeks out information, and does thorough research, so you can be assured that he knows his stuff.”

Kristján Kristjánsson

Business Consultant & Keynote Speaker

“Cool, calm, professional and a great motivator. Not many people can stand up in a room and give a flawless business pitch! But Darren manages to engage and be informative at the same time.”

Stuart Woodward

Graphic Designer/Director at Blast Factory Limited

“During this time I have been extremely impressed with his entrepreneurial skills, his on-stage presence (he is a fantastic public speaker) and his team spirit. Darren selflessly puts the needs of others first; providing unwavering help, support, advice, encouragement and technical skills.”

Pam Underdown

Business Growth Specialist

“Since Darren has been at UCFB i have found the CRM to be more user friendly and easier to use and will only continue to improve whilst he is working at UCFB.”

Jordan Hearne

Senior Regional Recruitment Manager, UCFB

Let's talk

Have a problem AI could solve?

Whether you need to find the right use case, prove an idea quickly or put a working solution into production - let's talk about what you're trying to achieve.

+44 7982 644911LinkedInUnited Kingdom (London area)
© 2026 Darren Ferneyhough. All rights reserved.Success Solutions Ltd