! This page is part of a project defined by the brief, the knowledge base and the constraints. Owned by Candace, Reham, Reba and April. Change those — and this changes.
The Blog Booster · Slipstream 03 · for Candace · GIG Gulf

The Blog Booster.
Accelerating insight into engagement.

Five compounding ways we turn GIG's data, expertise and reputation into reader engagement — every week, with the system getting smarter as it runs. Twenty blogs in month one. Eight a month thereafter.

For Candace Tay · GIG Gulf From Articulate AI Date 19 May 2026 Status Discovery complete · live for your direction
The five levers

Five ways we accelerate insight into engagement.

Each one is a path from data and judgement already in GIG's possession to engagement already on the table. The biggest single lever is the first one — the workflow itself. AI is the supporting cast, not the headline.

01 — The workflow does the work — not you, not me, not the AI

The single biggest reason cadence has slipped isn't lack of strategy, lack of talent, or lack of tools. It's the friction in the existing process — the ten-email threads, the lost drafts, the duplicate approvals, the version drift between Word and Drive and Liferay. This pipeline is built around removing that friction. Specifically:

  • Blogs are uploaded as structured data, not manually retyped into Liferay. One source of truth, one machine-readable handoff to Mark and Sucheta. Copy-paste between systems goes to zero.
  • Fresh topic ideas land every week. The backlog is refreshed from real GSC signal every Monday morning — automatically. You don't have to brainstorm; you have to choose.
  • No single person stops the flow. If Candace is travelling, the pipeline still moves — pre-approved templates ship, named authors fill their slots, the backlog stays warm. The system tolerates absence by design.
  • Approvals are routed, not chased. When a blog needs fact-and-quote sign-off, you see a clean draft with a clear decision — yes, edit, or no. No "did you see my email last week."
  • Every artefact lives in one place. Draft, edits, approvals, performance data, repurposed assets — one project view, no version hunting.

Remove the friction and the cadence holds itself. Everything else on this list compounds on top of that.

02 — Read every signal in the data you already own

Search Console, GA4, the funnel data from your quote engine, the inbound questions to MyGIG — the signal is already there. We pull it weekly into one operational view that tells us which queries are on the cusp of page-one rankings, which posts are driving funnel conversion, and which competitor blogs are eating ground we should own. Within ninety days we also know which posts drove quote-starts, which produced policies, which only produced traffic. Stop guessing what to write. See what's earning attention — and what's earning revenue.

03 — Codify GIG's voice once, use it everywhere

Will (the brand-voice persona in our team) extracts the tone and patterns from your top-twenty existing blogs and turns it into a working voice profile. Every AI draft then sounds like GIG, not like generic insurance copy. The same profile applies to email, social, paid-search, broker comms — one investment, every channel. Voice consistency is one of the strongest brand-trust signals there is.

04 — Thirty minutes of human judgement on every blog

Every blog gets a 30-minute editorial elevation on top of the AI draft. An editor's note from a UAE-resident voice, a pull quote, a local insight only a Dubai expat would catch, one named source from Khaleej Times or RTA or the Central Bank of the UAE. This is the step that turns "an article" into "a piece worth sharing." Non-negotiable.

05 — Be the named answer when customers ask AI

The shift nobody in UAE insurance is yet preparing for: half of "search" in 2026 is happening inside ChatGPT, Perplexity, Google SGE and Claude. We optimise GIG content for those answer engines as well as for Google. When a UAE customer asks "best car insurance with traffic fines" — to Google or to AI — GIG is the named response.

Slipstream · project family

Part of the Slipstream project

This blog pipeline sits alongside two other workstreams we've already started together — the Comprehensive page redesign and the Dataflow project. One name, one rhythm, one team. See the Slipstream home →

Why now

The blog is already working. Focus is the bottleneck.

Search Console tells a clear story. Across the last sixteen months, the blog section of giggulf.ae has earned 43,300 organic clicks from 4.21 million impressions, at an average CTR of 1% and an average ranking position of 9.2. When you pushed it, it really worked. The clean uplift from September 2025 onwards lines up exactly with the month you increased publishing volume.

43.3K
Organic clicks · 16 months · blog
4.21M
Impressions · same window
9.2
Average ranking position
+2,000
Target additional clicks per month

When you didn't push it, it didn't work — and the reason isn't strategy.

The reason for the gaps in the cadence isn't that GIG ran out of ideas, ran out of brand fit, or stopped believing in content. The reason is that the process to ship a blog is painful.

Painful for me. Painful for Candace. Painful for Kevin. Painful for Sucheta. Ten-email threads to land one paragraph of approval. Versions floating between Word, Drive, Liferay drafts. Approvals waiting on people who already approved a near-identical version last month. Drafts that go cold because they fell to the bottom of an inbox. Posts that ship and nobody knows because the launch email got buried.

That's not anyone's fault. It's the system. The system was built when blog cadence was monthly and approvals could afford to drift. Now we're trying to run weekly cadence on a monthly-cadence process. It can't hold.

+2,000 incremental clicks/month — roughly a 40% lift — is achievable. Not by working harder. By removing the points where the process eats the week. The pipeline you'll see below is built around that single observation.

What it is

A pipeline that gets smarter every week.

Six steps that run in a loop, not a line. Each cycle's performance data shapes next cycle's topics — so we stop guessing and start tuning.

1. Signal in — what's working, what's missing

We pull Search Console weekly. Identify the queries where you rank between positions 5–15 (the "almost there" page-one zone), the queries with high impressions and low CTR, and the queries your competitors rank for that you don't yet.

2. Topic backlog — refreshed every Monday

The signal feeds a topic backlog. Every Monday morning we review and re-rank. You can add ideas at any time and they go into the same backlog. The backlog never has fewer than 30 candidates.

3. Draft — AI does the heavy lifting

AI produces a 700–1,200 word draft built on your existing tone of voice, with links to your product pages and your funnels. Source citations included (Khaleej Times, RTA, Central Bank of the UAE, where relevant). Tables and structured data when the topic supports it.

4. Human touch — 30 minutes to elevate

Every blog gets a 30-minute editorial pass on top of the AI draft. This is where the personality goes in — an editor's note from a UAE-resident voice, a pull quote, a local insight, a sharper headline. This step is non-negotiable.

5. Brand voice + your sign-off

The draft goes through a brand-voice check (consistency with the existing GIG voice), then to you for fact and quote approval. You get a clear yes/no with a deadline — typically 48 hours per blog. If you have edits, they come back through the same loop once.

6. Publish + measure

Mark handles the upload to Liferay. We track each blog's performance for 30 days post-publish — clicks, ranking trajectory, time-on-page, and most importantly which blogs sent visitors into a quote or claim funnel. That data feeds Step 1 of the next cycle.

Already in flight

This isn't a plan on paper — it's a project in motion.

I've spent this week running the discovery work so the proposal is grounded in your data, not in assumptions. Here's where I am:

✓ Done

Search Console data pulled

16 months of blog performance data analysed. Headline numbers above. Top-query breakdown landing tomorrow with the per-page ranking detail.

✓ Done

Blog archive audited

All ~100 published blogs catalogued by category, date, and topic. Travel Insurance is the dominant category (~45 blogs, mostly destination-led). Car Insurance is the second strongest (~15 blogs). The September 2025 publishing spike is what drove the GSC uplift — proves volume works.

✓ Done

Blog template analysed

I've taken apart the current template and identified the elements I want to keep (FAQ accordion, social share, related-blogs sidebar, regulatory footer) and the elements that are missing (author byline, visible date, sources/references section, structured-data tables). I'll send Mark a one-page brief on the additions.

→ In progress

Tone-of-voice extraction

Pulling the writing patterns from your top-20 existing blogs so the AI drafts sound like GIG, not like generic insurance content.

→ In progress

Competitor landscape

Auditing Sukoon, Tawuniya, Salama, Orient and Watania (UAE) and Lemonade, Hippo, Aviva (global) blog operations. Where they're winning, where they're not, where GIG can own ground they've ignored.

○ Queued

20 topic ideas for you to react to

Comes to you by end of next week. You react with thumbs up / down / "add this angle." The 20 that survive become the month-one slate.

Proof — we fix the process

The life of one blog. And what you do in it.

One blog, five working days, end-to-end. Below is the actual journey — who does what, when, on which device, in which channel. Your touchpoints are highlighted. The aim is brutal: your total time involvement per blog is roughly 8–12 minutes, across three short touches, none of them requiring you to open a new tool, log into a new system, or read a long email.

Day 0 · Monday morning · the weekly topic refresh

What happens automatically: the system pulls GSC + competitor signal overnight Sunday → Monday. A re-ranked backlog of 30+ topics is surfaced; the 5 strongest new candidates for the week are pulled to the top.

Your touch · ~3 minutes · on your phone, in WhatsApp: at 09:00 Monday you get one WhatsApp message — five topic lines, each one sentence. "Topic 1: Cycling rules update for 2026 — Bahrain reciprocity? 👍 / 👎 / reframe." You reply 👍, 👎, or a voice note saying *"yes but make it specifically about families with kids."* That's it. Backlog locked for the week.

What Candace does NOT do: open Word. Open Drive. Read a strategy memo. Brainstorm topics. Sit on a Zoom for an hour.

Days 1–2 · Tuesday and Wednesday · drafting + voice + human elevation

What happens behind the scenes: AI draft built using GIG's voice profile + the topic brief. Will reviews for voice consistency. Editor does the 30-minute human-touch elevation — editor's note, pull quote, one named UAE source, structured-data table where it fits. Hero image generated or selected.

Your touch: none. You don't see this stage. You're free.

Day 3 · Thursday morning · fact & quote sign-off

What happens: the draft is ready. You get one WhatsApp message OR one email (your pick — we set it to whichever you actually open faster). One link. Tap or click it.

Your touch · ~4–6 minutes · on phone or work laptop: the link opens a clean reading view of the blog — no logins, no Liferay, no Word. Three short callouts at the top: "Please verify (a) AED 850 entry price still current; (b) the quote attributed to you; (c) the claim-time statistic." You can:

  • Tap a 👍 button → approves all three callouts in one tap, blog moves forward.
  • Type a one-line edit in the comment box → edit applied automatically, draft re-renders.
  • Voice-note it via WhatsApp reply → AI transcribes, applies the edit, sends you the revised version. Voice works on phone, on work laptop's WhatsApp Web, anywhere.
  • Tap 👎 → blog goes back to draft with your reason logged.

If your work laptop blocks WhatsApp Web (some corporate IT policies do): the same link works in email. One tap from your inbox. The system adapts to whatever IT lets through.

Day 4 · Friday · final preview & publish

What happens: all edits applied. Final preview rendered exactly as it will appear on giggulf.ae. Mark gets a structured-data package via API — no copy-paste, no re-formatting, no version drift. The blog goes from our draft system to Liferay as machine-readable data.

Your touch · ~2 minutes · optional final glance: if you want a final look, you get a preview link. One tap, scan, 👍. If you trust the loop, you skip this — the Day 3 sign-off was the binding decision.

Publish: Mark schedules or publishes. You receive a confirmation message with the live URL.

Days 4–34 · the loop closes

What happens automatically: performance tracked for 30 days post-publish — clicks, ranking trajectory, time-on-page, funnel attribution (did this blog send people into a quote or a policy?). Repurposed derivatives (LinkedIn carousel, email digest slot, social posts) are produced in the 24 hours after publish.

Your touch · ~3 minutes per week · a single WhatsApp digest every Friday: three lines. *"This week's published blogs: A drove 340 clicks · B drove 120 clicks · C drove 80 clicks + 5 quote starts. Next week's topic queue: [link]."* You react if you want; you don't have to.

Candace's total time per blog

8–12 minutes. Three short touches. No new tools.

Topic approval Monday morning (~3 min on phone). Fact-and-quote sign-off Thursday morning (~4–6 min via WhatsApp or email link). Optional final glance Friday (~2 min). Weekly performance digest (~3 min total across the week).

The system adapts to you, not the other way round. WhatsApp is the primary channel — voice notes, single-tap approvals, no logins. Email is the fallback for whenever WhatsApp Web is blocked on your work laptop, or when you'd rather read than tap. Either way, the action you take is always a single tap or a 30-second voice note. Never a thread.

For comparison — the current shape of shipping one blog

Today, by your own description and Kevin's, Sucheta's and Mark's, shipping one blog can involve 5–10 emails, multiple drafts in Word and Drive, version drift between systems, approvals that wait 3–7 days, and posts that occasionally ship without anyone being clear who said the final yes. The Blog Booster's commitment is to compress that to one topic-approval tap, one fact-check link, and a final-glance option — at your speed, on your phone, in your time. The pipeline absorbs the friction so you don't.

When things land in your inbox

What happens next, and when.

Week 1 — this week
Discovery + tone-of-voice + competitor audit
Mostly done. Output: this microsite + the discovery summary + the competitor read.
Week 2 — next Monday
20 topic ideas land in your inbox
You react: thumbs up, thumbs down, or "add this angle." Aim for a 48-hour turnaround.
Week 2 — same week
Standards locked: structure, categories, author voices, CTAs
I send you the working standards doc for sign-off. One review, not several.
Week 3
Pilot — first 3 blogs drafted end-to-end
Three blogs run through the full pipeline so we catch friction with the template, with Mark's upload, and with your review cadence before we scale.
Week 3 — pilot review
You review the 3 pilots + sign off on scale
If yes, we go to 20. If you want to adjust the formula, we adjust before scaling — cheaper to learn on 3 than on 20.
Weeks 4 – 6
17 more blogs to hit the month-one slate of 20
Cadence: roughly 3 per week. You see each one for fact/quote sign-off.
Week 7 onwards
Sustained cadence — 8 blogs per month
Plus your own articles, plus weekly topic-backlog refresh, plus monthly performance review with you.
Your part

Three things I need from you to keep moving.

1. An author list

I'd like at least three named authors at GIG who can each own one or two LOBs — Motor, Travel, Home, Health, SME. Each gets their own byline on relevant blogs. It spreads the load, builds trust with your readers, and lets us write in a voice that feels human.

You're an obvious seventh author — your own articles publish under your own name on whatever cadence suits you. I'm an obvious sixth — I appear as the strategist / "in conversation with" voice, not as a content lane.

What I need: a name, a role at GIG, the LOB they'd own, and one or two sentences describing how they speak (formal, warm, technical, conversational). I'll handle the rest.

2. Approval on the first blog-draft list

When the 20 ideas land in your inbox next week, you go through them and say yes / no / "yes, but reframe to X" per row. Five minutes total if you're decisive. The ones that survive get drafted.

What I need: a 48-hour turnaround on the topic list. After that, 48 hours on each blog's facts and quotes. That's the SLA the system runs at.

3. Your own ideas, anytime

You almost certainly have angles I'd never think of — the customer call last week, the underwriting trend you spotted, the partner conversation that revealed a gap. Those go straight to the top of the backlog.

What I need: a short message — WhatsApp, email, voice note, whatever's fastest. One sentence is enough. "We should write about X because Y." I'll take it from there and bring it back to you as a topic card for your sign-off.

How this gets built

Augmented AI. Not Replacement AI.

I'm not building this from scratch. The drafting layer uses best-in-class AI tools (Claude, brand-voice plugins, SEO tooling, GEO-optimisation tools for the Perplexity / ChatGPT-search era). The CMS layer is your Liferay system, operated by Sucheta. The publishing layer is Mark's domain.

What I add is the integration — the order in which those tools fire, the 30-minute human-touch step, the brand-voice consistency check, the GIG-specific judgement about what to publish and what to skip, the measurement loop that closes the system. That's where Articulate earns its keep.

The principle that runs the whole pipeline: every person — you, the authors, Mark, Sucheta, Will, me — uses their judgement, experience and critical eye on the content, not on the email about the content. The AI handles the chase, the formatting, the version control. The humans do the work only humans can do.

PROJECT MANDATE

The full tools, case studies and augmentation principle

A separate page lists every tool I'd use, what it actually does, with a named example of an improvement it has made elsewhere. Plus three insurance + fintech case studies (Lemonade, Wise, Aviva UK) and three places this investment compounds beyond blog.

Read the project mandate →
Everything in one place

Documents and links.

Next move

Tell me what you think, and I'll send the 20 ideas Monday.

Reply to this with one of three answers:

  1. Yes, proceed — I send 20 ideas Monday, with names of three authors from you in parallel.
  2. Yes, but reframe — quick call this week to tune the shape, then we go.
  3. Not now — we close this and focus on the Comprehensive page only.

I'd rather you say not-now than say yes-out-of-politeness. The pipeline only works if you've got the bandwidth for the 48-hour review cadence, and I'd rather build something smaller you'll actually use than something bigger that stalls in your inbox.

— Anthony, Articulate AI · 19 May 2026 · Dubai