Home

Services

SEO Services Social Media Website Development AI Automation About Reviews Portfolio Blog Contact Free SEO Audit
AI & ML Automation Portfolio

Automation We Built for Ourselves First

Everything here runs on our own client work before it is offered to anybody. That is deliberate — the fastest way to find out whether an automation genuinely saves time is to depend on it yourself for a year. What is not here is a client logo grid: this practice is newer than the other three, and the page says so.

Every figure on this page is attributed — client, Search Console, or our own property.

✓ Google Certified SEO Experts ★★★★★ Trustpilot 5.0 ★★★★★ Google 4.9
An abstract network of connected nodes representing automation
Systems in daily use3
Content written by automation0
Before You Read The Numbers

Where every figure on this page actually comes from

Most agency portfolios show percentages with nothing behind them. Every result below is tagged with its source, and anything we could not attribute to one of these three is simply not shown as a number.

Client-stated — the client said it publicly, on the record Search Console — measured in a property you or they can open Our own site — fully inspectable, including the mistakes
AI & ML Automation Work

What we have built and rely on

The internal systems run on live client engagements, including this site’s own migration. None of them is a demo built for this page.

Abstract network of connected nodes representing automation Internal build · Monitoring
4AI engines monitored
Ghalib Ashrafi Runs continuously

AI-visibility monitoring, built in-house

Tracking how often and how accurately a brand is surfaced by ChatGPT, Perplexity and Google AI Overviews — not just where it ranks on Google. It runs underneath every SEO engagement here rather than being sold as an add-on.

0Extra cost to clients
LLM monitoringAEO / GEOReporting
Internal build — runs on our own client engagements
Colleagues reviewing data together at a desk Internal build · Analysis
0Words written by automation
Ghalib Ashrafi Used on every audit

Clustering and crawl analysis at a scale humans will not do

Thousands of queries grouped by meaning rather than matching words, and crawl and log files read for patterns that never surface row by row. The writing is deliberately not automated — Google is explicit that mass-produced content aimed at rankings is a spam violation.

24Pages in our own migration audit
Topic clusteringCrawl analysisPipelines
Internal build — used on our own audits, including this site’s migration
A professional services team working in a bright office Professional services · UK
68%Fewer manual hours
A UK professional services firm 4-week build

Enquiry triage that stopped losing weekend leads

Enquiries were landing in a shared inbox and sitting there until Monday. We scored and routed them automatically to the right person, with a human checkpoint before anything reached the client — the bottleneck was routing, not writing.

9 minMedian response time
100%Human-checked before send
Lead routingWorkflow automationIntegrations
Search Console — client property, measured across the engagement
The Line We Draw

What we deliberately do not automate

Every system above is research, analysis or monitoring. Automation points a strategist at the right problem; a person then decides what to do about it. That division is not a stylistic preference — it is the difference between a tool that compounds and a liability that arrives with the next algorithm update.

We do not mass-produce content. Google is explicit that using automation to generate content at scale for the purpose of manipulating rankings is a spam violation, and enforcement has only got sharper. Building a client’s visibility on something that can be removed in a single update is not a service, it is a deferred bill.

The other thing we will not do is start from the tool. Most AI projects fail because software got chosen before anyone named the problem, which produces a subscription rather than a result. We work backwards: find the repetitive, expensive workflow, check whether the data behind it is clean enough to trust, and only then decide whether AI is the right answer.

Sometimes it is not. A spreadsheet, a form, or deleting a process outright beats a model more often than anyone selling models will admit.

  • Analysis, not authorship — automation finds the problem. A senior strategist decides what to do about it.
  • No content at scale — Google names mass-produced content as a spam violation, and the penalty lands on your domain rather than ours.
  • Problem before tool — we find the expensive workflow first. Picking software first produces a subscription, not a result.
  • Data quality checked first — an automation running on unreliable data produces confident nonsense faster than a human ever could.
  • We will say no — if the honest answer is a spreadsheet or deleting the process, that is what you will hear on the call.
Colleagues reviewing data together on a screen

Almost every failed AI project we have been asked to rescue started the same way — somebody bought a tool, then went looking for a problem it could solve.

GA
Ghalib Ashrafi Founder & Digital Strategist · 12+ years across search, social & web
3Systems in daily use
0Content written by automation
4Services they run underneath
12+Years of process to automate
What You Actually Get

What an automation engagement looks like

Four stages, and the first one frequently ends with us recommending that you do not build anything at all.

Bottleneck audit

We map where the hours actually go, which is rarely where people assume. This is the stage where most bad projects get stopped before they cost anything.

Data readiness check

Whether the data behind the workflow is clean, complete and accessible enough to be trusted. If it is not, that becomes the project instead.

Build

A working system on your own data, integrated with tools you already pay for rather than adding another dashboard nobody opens.

Handover & monitoring

Documentation, ownership and a way to tell when it stops working — because an automation that fails quietly is worse than having none.

Human checkpoints by design

Every output that reaches a customer passes a person first. That is a design rule in the architecture, not a phase that gets dropped when things get busy.

Your accounts, your data

Keys, models, repositories and data stay in your name and stay yours if we part ways.

What these clients said

Named people at named companies — the same quotes the results above are attributed to.

All of them, in full, on our reviews page.

Tell Us Where the Hours Go

No tool pitch. Describe the workflow that eats your week and we will tell you whether automation is the right answer — including when it is not.

Ready to find out what is actually worth automating?

Free 30-minute audit · No obligation · Response within 24 hours

Get in touch with Ghalib Ashrafi HQ

We're here to help you grow your business. Reach out today.

Google Map embed — this page's HQ address
Phone / WhatsApp: +92 343 2653224
Email: info@ghalibashrafi.com
Hours: Mon–Sat, 10am–7pm (PKT)
Still deciding? Get a free SEO audit — no obligation. Get Free Audit