Skip to main content

Case Studies

Not a deck. Not a pitch. Real businesses.

Every AI consultant shows case studies. Most are written to make the consultant look good. These are written to show you what actually happens when a business operator, not a tech vendor, maps your processes and builds the AI system you need. Real clients. Real numbers. Real code running in production today.

60%

Fewer emergency callouts at Sovereign Water

15+

Weekly Fast Forwarder newsletters shipped from the Moving Adverts platform

32

Apps deployed across a single unified business OS

Client Engagements

How the work breaks down

Each engagement follows the same rhythm: business discovery, process audit, strategy roadmap, build and embed, train and govern. Below are three of the deepest implementations.

Case Study 01 · Featured

Sovereign Water

Commercial Water Treatment
QSR & Hospitality · UK / KSA / UAE

lockHybrid frontier + offline models
sensorsIoT integration
apps32 modular apps
scheduleOngoing partnership

Stack

React 19 Supabase Claude OpenAI Perplexity N8N WordPress Apify

Services delivered

  • check_circleConsulting
  • check_circleArchitecture
  • check_circleBuild
  • check_circleTraining
  • check_circleGovernance

At a glance

200+

Commercial sites under management

32

Modular apps in one platform

3

Geographies (UK, KSA, UAE)

Sovereign Water dashboard overview showing operations, sales, content and finance in one unified platform
errorThe challenge

Sovereign Water manages commercial water treatment for major QSR brands across the Middle East, with operations split across the UK, KSA and UAE. Over 200 sites, dozens of field engineers, and no unified digital oversight. The business was running on spreadsheets, WhatsApp groups and reactive callouts.

Worse, every other function was equally fragmented. Sales lived in one tool, consultants in another, content marketing in a third, financial tracking in a fourth. As the customer base grew across three geographies, the cost of running the operation was climbing faster than revenue.

tips_and_updatesWhat I built

A unified business operating system. Not three apps, not five apps. Thirty-two modular apps sharing a single PostgreSQL database, one auth layer, and row-level security that isolates data by client, entity and region.

  • checkOperations + IoT. Custom ERP covering asset management, site and contract tracking, field engineer dispatch. Sensors feed live operational data back to head office for predictive maintenance.
  • checkSales pipeline. Kanban deal tracking through six stages with probability-weighted forecasting, by month, team and region.
  • checkLead engine. AI-assisted prospect discovery, contact enrichment, and multi-channel outreach through email campaigns wired back into the sales pipeline.
  • checkContent cascade. Blog articles drafted in platform, published to WordPress, repurposed into LinkedIn and X posts, and aggregated from 60+ RSS sources with AI relevance scoring.
  • checkFinancial tracking. Creditor and debtor ledger, bank statement import, cashflow forecasting by entity, commission splits, and outstanding balances recalculated in real time.
  • checkConsultant CRM. Relationship manager for partner firms with activity logging, evidence file storage and warm/cold state tracking.
check_circleThe result

60%

Fewer emergency callouts

200+

Systems with real-time visibility

40%

Faster service response

One unified platform replaced what used to be five fragmented tools and a dozen spreadsheets. Sales, operations, finance, content and customer success all talk to each other now. Commercial data stays on Sovereign's own hardware via offline models, so major customer brands keep full data sovereignty.

Takeaway for you

If you are running a multi-site operation on spreadsheets, the biggest wins come from unifying the data, not chasing features. Map your workflows first. Then decide where AI and automation can replace human minutes. The tools come last.

Case Study 02

Moving Adverts

Creative Agency · UK
Featured deployment: CSN, Cargo Solutions Network

verifiedBrand voice enforcement
workspacesWhite-label multi-tenant
auto_awesomeAI content engine
rss_feedRSS-to-newsletter pipeline

Stack

React 19 Supabase Claude Sonnet 4.5 OpenAI Perplexity LinkedIn API X / Twitter API Mailchimp RSS Gen

Services delivered

  • check_circleConsulting
  • check_circleArchitecture
  • check_circleBuild
  • check_circleTraining
  • check_circleGovernance

At a glance

4

Content engines (Marketing, X, LinkedIn, Newsletter)

15+

Weekly "Fast Forwarder" newsletter editions published

100%

Brand voice enforced at generation time

X Social Engine managing the CSN X account with post creator, scheduler, calendar and feed
errorThe challenge

Moving Adverts is a UK creative agency building AI-powered marketing operations for its clients. They wanted to stop delivering AI as a billable hourly service and instead productise it: one white-label platform they could configure per client with that client's tone, brand, and channels baked in.

The first tenant was CSN, Cargo Solutions Network, a global freight and logistics brand that needed organic marketing, social automation and a weekly industry newsletter. Everything on-brand, everything scalable, and nothing that required the account team to babysit prompts.

tips_and_updatesWhat I built

A multi-tenant AI marketing platform with four content engines, each aware of the active client's brand profile so every generated word, image and post is on-brand by default.

  • checkOrganic marketing AI content engine. Generates blog articles, landing page copy and campaign assets from briefs or source URLs. Claude Sonnet 4.5 for drafting, Perplexity for live research, brand voice config enforced at every generation call.
  • checkSocial media automation. Cross-platform post generation, scheduling and publishing aligned to each client's brand guidelines and company tone. No more ten tabs per client.
  • checkX Social Engine. A dedicated module for X: post creator, scheduler, feed monitor and news feed integration. Drafts, scheduled and published counts surfaced per account, content generated from URLs or blog articles.
  • checkLinkedIn Engine. Post generation, engagement replies, approval workflow and calendar view tailored for LinkedIn's longer-form posting style.
  • checkAI Newsletter Builder + RSS feed engine. Ingests curated industry RSS feeds, clusters stories by theme, drafts newsletter sections with per-section regenerate and proof-read controls, then exports HTML and DOCX. CSN's Fast Forwarder weekly freight intelligence newsletter is now Volume 15.
  • checkBrand-safe AI image generation. Editorial photography prompts with style, lighting and subject presets so the images match each client's visual language. No stock lookalikes.
  • checkUnified analytics dashboard. Google Analytics connected per tenant. Active users, sessions, page views, top pages and sessions trend, all scoped to the current client.
check_circleThe result

4

Content engines in one platform

15+

Weekly newsletter editions shipped

100%

On-brand at generation time

Moving Adverts now has a productised AI marketing platform they can roll out to any new client as a configuration task, not a fresh build. CSN's organic marketing, X account, LinkedIn presence and weekly Fast Forwarder newsletter all flow through one system, with brand voice enforced before the agency ever sees the draft.

Takeaway for you

If you run an agency, AI is not just a productivity tool. It is a product. The margin is in building a platform your clients use under your brand, not in writing prompts by the hour. Codify the brand voice, embed the guardrails, and let the platform do the repetitive work.

Case Study 03

E8T

Hospitality White-Label SaaS
UK

workspacesTrue multi-tenant
apps15 apps, 5 departments
boltAI content pipeline
paletteDynamic branding

Stack

React 19 Supabase RLS Claude Haiku + Sonnet GPT-4 Vision Perplexity N8N LinkedIn + Meta APIs Mailchimp

Services delivered

  • check_circleConsulting
  • check_circleArchitecture
  • check_circleBuild
  • radio_button_uncheckedTraining (planned)
  • radio_button_uncheckedGovernance (planned)

At a glance

15

Apps across 5 departments

59+

Multi-tenant schema migrations

500+

Commits of iterative delivery

errorThe challenge

E8T needed a white-label business workspace their hospitality clients could rebrand and roll out without touching code. Existing options were either single-tenant apps bolted together or bloated enterprise platforms that required months of config.

The product also had to cover five departments at once, marketing, sales, HR, operations and finance, in one unified interface. Nothing on the market did that without forcing clients to stitch together six separate SaaS subscriptions.

tips_and_updatesWhat I built

A true multi-tenant platform built white-label from day one. Every piece of branding, app access, entity structure and permission is database-driven. Onboarding a new client is a configuration exercise, not a code change.

  • check15 apps, 5 departments. Marketing (content calendar, ideas, social, AI chat), Sales (CRM, pipeline, contacts, bookings), HR (employee directory, leave management), Operations (group to-do, stock), Finance (payments, VAT, cashflow).
  • checkAI content pipeline. N8N orchestrated Claude + GPT + Perplexity flow drafts blog articles with live research, generates brand-safe images, and suggests social posts with brand voice enforced per client.
  • checkContent approval workflow. Multi-stage review and audit trail so client sign-off is tracked before anything publishes.
  • checkSix role-based dashboards. Marketing, Sales, HR, Operations and Finance leaders each see their own real-time KPIs. Client admins see the roll-up.
  • checkDynamic branding. Logos, colours, even app labels pulled from the database per client. One codebase, infinite skins.
  • checkHard data isolation. Row-level security policies enforce client + entity boundaries at the database layer, not the application layer. Zero tolerance for cross-client leakage.
check_circleThe result

15

Apps in one workspace

59+

Migrations shipping multi-tenant schema

>500

Commits tracking iterative delivery

A single unified interface replaces five to ten fragmented SaaS subscriptions. New clients can be onboarded by configuration alone, with their own branding, their own data boundary, and their own subscription tier on premium apps.

Takeaway for you

If you are building a SaaS or internal platform you might ever want to deploy for a second team, multi-tenancy has to be in the foundation. Retrofitting it later is always more expensive than baking it in from day one. Database-enforced isolation beats application-layer checks every time.

Products & Platforms

What I have built for myself

Two products I own and operate. They are live, in production, and in some cases they power the client work above. You can think of them as proof that I eat my own cooking.

ClubHub Newcastle United hub with daily intelligence briefing, live stories, video breakdowns and creator desk Fixture threads for Newcastle United versus Leeds with pre-match, live and post-match community discussion rooms Match Centre listing live fixtures across competitions with kickoff times and live auto-updating data Live match detail page with lineups, stats and league standings table
sports_soccer

ClubHub

Premier League fan and creator platform

Sports Media · Self-owned

A multi-club fan platform covering all 20 Premier League clubs. Club-specific hubs with daily intelligence briefings, fixture-by-fixture community threads, a cross-competition Match Centre with live data, and an AI news pipeline that publishes the top stories to X automatically.

What's inside

  • checkClub-specific hubs for all 20 Premier League clubs with daily intelligence briefings
  • checkFixture threads with pre, live and post-match community rooms per fixture
  • checkMatch Centre across every competition with kickoff times and real-time auto-updates
  • checkLive match pages with lineups, stats and league standings tables
  • checkCreator Desk for instant fan posts straight to the club feed
  • checkAI news pipeline: 1000+ articles a day clustered and auto-published to @ClubhubPL on X

Stack

React 19 Supabase Edge Functions API-Football Stripe Recharts
RSS Gen feeds dashboard showing 200 feeds and 11,995 articles on the CLUBHUB account with health status and scraping activity Feed detail view with AI-assisted keyword filters, domain filters and feed URL configuration Source Hunter search interface for discovering RSS sources across news, YouTube, podcasts and X from a single search term Feed status and analysis panel showing extraction strategy, chronological ordering, images pulled and live scraping process log
rss_feed

RSS Gen

The content ingestion engine

Content Infrastructure · Self-owned

A source-discovery and web-scraping platform that turns any website into a queryable RSS feed. Currently managing hundreds of feeds and tens of thousands of articles across multiple client accounts. It is the hidden engine powering the content cascade in Sovereign Water, the news aggregator in Moving Adverts and the news pipeline in ClubHub.

What's inside

  • checkSource Hunter: one search term finds news, YouTube, podcast and X sources, auto-validates and saves them as feeds
  • checkMulti-tenant: switch between client accounts (ClubHub, CSN, Sovereign) to manage their feeds in isolation
  • checkFeed health monitoring: healthy, warning and error states tracked across every feed
  • checkAI-powered keyword filters: include / exclude rules with AI-assisted keyword suggestions, plus domain filters
  • checkThree-stage hybrid scraper with feed analysis showing extraction strategy, chronological order, and image availability
  • checkScale: 200+ feeds, 11,000+ articles on a single account, non-destructive stage variants so new site logic never breaks existing feeds

Stack

Node 18 Express Supabase RLS Perplexity GPT-4o mini Puppeteer Cheerio

Want the same kind of result for your business?

No pitch. No obligation. A free 15 minute call to see if it is worth us talking again.

Book my free 15 minute call