One programme. Seven workstreams. Sequential, not parallel.
Slipstream is seven discrete workflows, each scoped and signed off independently. They share a single rhythm across the programme and a single principle that runs through every one: Augmented AI, not Replacement AI. Every output gets a named human's judgement on it.
Workstreams ship sequentially
Doing two concurrently is fine. Trying to ship three or four before the earlier ones are stable is not. The current rhythm: GIGI (01) is the umbrella vision; Diagnostic Agent (02) is the tool that powers it; TL Recovery (03) is the first use-case and the pilot for both — all three just published as draft ideas awaiting Sidd + Nadine sign-off on Stage 1. Comprehensive page (04) running active in parallel with The Blog Booster (05) being scoped. Dataflow (06) paused awaiting Shifa's response; Sales videos (07) paused awaiting Reham's brief re-clarification.
What's running, what's queued, what's paused.
A natural-language layer across GIG's entire stack — ad, Liferay, email, funnel, segment performance — built from a stack of twelve specialist agents sharing one anonymised data substrate and one dashboard everyone in the room talks to. The umbrella vision. "That is indeed a Game changer." Replaces ~15 individual analyses currently knitted by hand across Lucie, Ismail, Anthony, Shifa, Reham and the data team.
Open the GIGI lab →The first GIGI sub-agent and the tool that makes GIGI real. Reads GIG's policy, CC and marketing data; anonymises at the source; runs ten standing cuts every day; refreshes the dashboard. Sister project to TL Recovery (03) — TL Recovery is the analytical workstream, 02 is the tool that makes it repeatable. Stage 1 pilot lands inside 03 at marginal cost.
Open the Diagnostic Agent lab →End-to-end diagnosis of the dramatic post-February fall in TL Comprehensive bookings — ~$1M at risk of the 2026 achievable target. A combined Marketing + Call Centre approach with a shared data pool and one chart everyone in the room argues from. Three hypotheses (Shift · War · LCS) — to be ruled in, ruled out or quantified. First concrete use-case for GIGI + the Diagnostic Agent.
Open the TL Recovery lab →Five page directions for the car-insurance Comprehensive page on giggulf.ae — from paste-replace copy edits through to ground-up rebuild. Currently in lab review with Candace, Reham, Reba and April. Live at the lab hub below.
Open the Comprehensive lab →20 high-quality GEO-optimised blogs for giggulf.ae in the first month, then 8 per month sustained, targeting +2,000 incremental organic clicks per month. The biggest lever is the workflow itself — removing the process pain that's been the real bottleneck on cadence. Augmented AI principle: every blog has a named author and a 30-minute human-touch editorial pass.
Open The Blog Booster → Project mandateUnified marketing-data aggregator pulling GSC, GA4, Google Ads, Meta, Microsoft Ads, SEMRush and Dynamics CRM into one operational view for Shifa's team. Shifa engaged 2026-05-21 with her existing workbook spec; Pulse scoping pack delivered the same day — brief, interactive Power BI dashboard mock, 20-page research doc, SecOps + Compliance review, tech-team effort hypothesis, impact assessment placing Pulse at #2 in the Slipstream order. Sequenced as the data substrate for TL Recovery (03) and the Diagnostic Agent (02).
View Pulse — scoping pack →Testimonial and sales reels for GIG Gulf, character-driven via Higgsfield Soul (Arjun is the trained character). Pilot in flight at around 126 credits spent; distribution paused pending brief re-clarification with Reham.
Reel pilot work archivedRenumbered 2026-05-20
GIGI promoted to 01 as the umbrella vision under which the diagnostic + operational + customer-facing agents sit. Diagnostic Agent at 02 because it is the foundation tool. TL Recovery at 03 because it is the first concrete use-case proving 01 + 02 work. The four pre-existing workstreams (Comprehensive · Blog Booster · Dataflow · Sales videos) keep their relative order, shifted down to 04–07.
What I won't do, on any workstream.
Five rules that apply to every workstream, every time. Codified in our shared canon.
- Sequential, not parallel. Doing two workstreams at once is fine. Trying to ship three or four before the earlier ones are stable is not.
- Named individual per workstream. Every workstream has one accountable GIG person. No "GIG marketing team."
- No publishing on GIG's behalf without explicit per-asset sign-off. Articulate never publishes for GIG.
- Source materials confidential. No cross-pollination to other Articulate clients. What GIG shows me stays GIG's.
- GIG-owned, Articulate-credited. Every artefact is GIG's IP, with my name attached as the producer.