LinkedIn Advisor Playbook¶

This page exists because Sush has 50+ blog posts that don't get pushed out enough. The problem isn't the content — it's distribution. This playbook captures the "thought-advisor mode" pattern that landed well on 12 May 2026: a customer question becomes a LinkedIn post becomes a blog click.

When to use this page

Read this at the end of any customer or T3 session to scan for a moment worth turning into a LinkedIn post. Also read it when Sush asks for "a LinkedIn post for this" or "thought-advisor mode". Cross-reference with Voice & Tone before drafting.

Why this works¶

A LinkedIn post that says "check out my new blog!" dies on arrival. A LinkedIn post that says "customer asked me this question, here's what I learned" travels — because:

It's not promotional. A question is generous; a brag is exhausting.
It's repeatable. Sush has customer sessions every week. That's 50+ posts a year without forcing anything.
It gives readers permission to not know. Niche topics travel further when they feel less lonely.
It uses the unfair advantage. Real customer conversations are content most people don't have access to.
It matches Sush's voice. The curious-intern, "for people like us", honesty-over-charm posture is genuinely his.

The blog is the destination. The customer question is the on-ramp.

The formula¶

Every post follows the same five-step structure. Keep it boring on purpose — the consistency is what makes it scan-able.

1. [HOOK]      — Customer's actual question OR Sush's honest admission OR a T3 pattern
2. [CONTEXT]   — One sentence of plain English. What is the thing?
3. [REFRAME]   — "Different jobs", "governing rule", the mental model
4. [VISUAL]    — Diagram, table screenshot, screenshot — the thing that makes it shareable
5. [LINK]      — Blog URL, framed as "full walkthrough" not "buy now"
6. [HASHTAGS]  — Two only. #MicrosoftCopilot #Microsoft365 by default.

Word count target: ~80 words of body text (the visual carries the rest). Tighter wins.

Three hook variants (rotate these)¶

Hook 1 — "Customer question this week"¶

Best for: posts pushing an existing blog after a customer session where one question came up.

Customer question this week: "we have Microsoft Designer — why would I use Adobe Express?"

Fair question. Different jobs. Brand Kit governs both.

When to use: A specific customer asked a specific question that maps to a blog you already have.

Hook 2 — "Honest admission" (the curious-intern voice)¶

Best for: posts about something Sush genuinely had to look up, learn, or had a "I didn't know that" moment with.

Honest admission: until a customer asked, I hadn't fully clocked what Adobe Express is inside Copilot.

(Turns out it's Adobe's free creative tool — think Canva — invokable as @Adobe Express in Copilot Chat.)

When to use: When the topic is niche and the reader is likely in the same boat. The admission lowers the trust barrier.

Hook 3 — "Every T3 session, this question keeps coming back"¶

Best for: posts about evergreen topics that recur across many customer sessions, not just one.

In every Train-the-Trainer session, one question keeps coming back: "how do we make Copilot use our brand?"

So I wrote it all down. Brand Kit, OAL, templates, Brand Checker, the lot.

When to use: Posts pushing a comprehensive guide. The pattern signals that this is a real, common need — not something Sush invented.

Voice guardrails¶

Cross-reference Voice & Tone — the brag-allergy rules apply doubly on LinkedIn.

✓ Do	− Don't
Open with a question or a confession	Open with "Excited to share..."
Plain English, intern-chair posture	Hype words: frontier, agentic, mission-critical, robust
"For people like us" framing	"For enterprise leaders looking to drive transformation"
Numbers as facts ("$30/user/month")	Numbers as authority ("1,200+ readers!")
Credit the customer or the conversation	"I'm proud to announce..."
One clear CTA (the blog link)	Multiple CTAs (link + book a call + DM me + ...)
`@Adobe Express` / `>` quoting the customer	Heavy formatting, ALL CAPS, 🔥💯🚀
Sentence-case headlines	"GAME-CHANGING Insights about M365 🤯"

If the draft sounds like a brand consultant, rewrite it as if Sush is texting a colleague.

Sizing rules¶

Body text: ~80 words. Going past 150 words tanks engagement.
First line is the hook. LinkedIn cuts after ~3 lines on mobile — the first line decides whether anyone reads more.
Blog link in the body, never the first line. LinkedIn deprioritises posts where the first line is an external link.
Line break every 1–2 sentences. Walls of text don't get read on LinkedIn.
Hashtags: 2 max. Three or more reads as spam. #MicrosoftCopilot #Microsoft365 is the default pair; rotate one for topic-specific (#BrandKit, #SharePoint, #PowerPoint).
No early external links. If you must link out, put it in the body, not line 1.

Visual rules¶

Every post needs one visual. The visual is what makes it scrollable-stoppable.

Topic shape	Visual to use
Decision between tools / paths	Decision tree diagram (Excalidraw → SVG → PNG via the scripts/ pipeline)
Comparison across N items	Table screenshot from the blog (clean, scannable)
Step-by-step process	Numbered list visual or PowerPoint slide screenshot
New feature walkthrough	Real UI screenshot with light annotation arrows
Architecture / how-it-works	Mermaid diagram screenshot or Excalidraw
Counter-intuitive insight	Big-text quote card with the punchy line

Always upload as PNG (LinkedIn favours image posts over link previews). The 2× retina version from the blog's /diagrams/ folder is usually right.

Always include alt text on the LinkedIn upload — accessibility plus a small SEO boost.

Cadence¶

Sustainable beats viral. Sush should not feel pressured to post on demand.

Target: 1 post per fortnight. Tied to customer sessions, not a calendar.
Trigger-based, not date-based. If no customer session produced a moment worth posting, skip the week. Forced posts read forced.
Bursts are fine. Two posts in the same week is okay if two genuine moments lined up. Three in one week is too many.
Never post the same blog twice within 90 days. Refresh the hook + visual if revisiting.
Best window for NZ audience: Tuesday–Thursday, 8–10am NZST or 4–6pm NZST. Microsoft people read LinkedIn at coffee and at end-of-day.

The trigger system (how I work this for Sush)¶

This is what I do automatically without being asked.

Every customer-session or T3-session journal entry I write:

Scan the entry for a moment. A moment is one of:
A specific customer question (→ Hook 1)
A "I didn't know that" or "had to look this up" admission (→ Hook 2)
A recurring pattern across multiple sessions (→ Hook 3)
A surprise reaction from the customer (e.g., "customers were surprised that..." → can become Hook 1 or 2)
Match the moment to a blog post in Sush's queue — does an existing blog cover what the customer asked? If yes, draft a post pushing that blog. If no, flag as a "blog post idea" instead.

Surface the draft to Sush in the session response, formatted as:

💡 Thought-advisor candidate
Moment: [the question / admission]
Blog this could push: [URL]
Draft post:
[80-word draft following the formula]

Sush decides: ship / tweak / pass.
If shipped, log it under Posted log so we don't double-post the same blog within 90 days.

This is opt-in. If Sush ignores three candidates in a row, scale back to suggesting once per customer-session entry instead of every entry.

The queue¶

A living list of blog posts ready to be pushed, each with the customer trigger that would justify the post. Sush adds rows; future-Copilot suggests posts when the trigger condition fires.

Format¶

Blog post (slug)	Hook variant	Trigger condition (when to push)	Suggested visual

Starter queue (12 May 2026)¶

Blog post	Hook variant	Trigger condition	Suggested visual
`microsoft-365-copilot-brand-kit-complete-guide`	Hook 1 or 2	Customer asks about brand assets, stock images, OAL, Brand Checker, or admits not knowing Adobe Express	Decision tree (already built)
`microsoft-365-copilot-licensed-complete-guide-for-trainers`	Hook 3	When teaching T3 and licence/feature confusion comes up	Licence matrix table screenshot
`microsoft-365-copilot-chat-complete-guide-for-trainers`	Hook 3	When a customer is starting their Copilot rollout journey	Foundations diagram
Practice-exam tool launches	Hook 1	When a customer asks how to prep for a cert affordably	Real-pricing comparison ($9 vs $79)
Copilot Readiness Checker	Hook 1	Customer asks "are we ready for Copilot?"	7-pillar wheel
Feature Matrix	Hook 3	Customer is comparing tiers (Chat vs Licensed vs E5 + Copilot)	Tier comparison table
Licensing Simplifier tool	Hook 2	"I had to read the licence docs three times before this clicked"	Tool screenshot

When Sush ships a post, move it to Posted log with date + LinkedIn URL.

Worked example — Brand Kit decision tree (12 May 2026)¶

The post that triggered this playbook. Documented in detail so future drafts can pattern-match.

Setup:

Customer session triggered Sush asking for a decision tree
Sush admitted he hadn't heard of Adobe Express → strong Hook 2 candidate
Existing comprehensive blog already published → ready to push
Built diagram in Excalidraw, exported SVG + PNG, embedded in blog

Final post (Option A — Hook 2, what Sush shipped):

Honest admission: until a customer asked, I hadn't fully clocked what Adobe Express is inside Copilot.

(Turns out it's Adobe's free creative tool — think Canva — invokable as @Adobe Express in Copilot Chat for animation and creative-depth designs.)

So: Adobe Express vs Microsoft Designer vs Brand Kit — which one when?

Drew the decision tree. Brand Kit governs both ↓

[diagram image]

Full walkthrough — templates, stock images, Brand Checker, admin setup, the lot — for IT and brand managers:

https://aguidetocloud.com/blog/microsoft-365-copilot-brand-kit-complete-guide

#MicrosoftCopilot #Microsoft365

Why it works:

Hook 2 = curious-intern voice, genuinely Sush
Parenthetical context = doesn't assume reader knows what Adobe Express is
"Different jobs. Brand Kit governs both." = the reframe in 5 words
↓ arrow = signals visual coming
Blog link in the body, not the first line
Two hashtags only
Word count: ~85 ✓

Posted log¶

Track every post we ship. Lets us avoid double-posting the same blog within 90 days and learn what hooks land.

Date	Blog pushed	Hook variant used	Post URL	Notes
12 May 2026	brand-kit-complete-guide	Hook 2 (honest admission)	Sush to fill in after posting	Decision tree as the visual. Adobe Express was the angle.

What NOT to post¶

There's a line. Some things stay off LinkedIn even if the engagement upside is real.

Anything Microsoft-confidential. WorkIQ insights, internal product timelines beyond what's on the public roadmap, customer names without explicit permission, tenant IDs, customer pain points that could identify the customer.
Anything that contradicts Microsoft's official messaging. Even if I'm right and a Microsoft doc is out of date, public LinkedIn isn't the place to litigate that. Use blog posts (which can be measured + revised) instead.
Posts about competing platforms that punch down. "Canva is fine but Microsoft Designer is better!" reads insecure. Sush's voice is comparative-honest, not tribal.
Self-stats. Subscriber counts, blog reader numbers, view counts, milestone celebrations. See the brag-allergy.
Hot takes that aren't your own. Sush's posts work because they're his honest read. Reposting someone else's spicy opinion isn't his voice.

For Sush, every time:

Download the PNG from the blog's /diagrams/ folder
Right-click the link in the figcaption → "Save As..."
New post on LinkedIn
Click the image/photo icon
Upload the PNG
Add alt text in the upload dialog (use the SVG's <title> text as a starting point)
Paste the body text (replace any [diagram] placeholder — the image carries that part)
Verify the first line is the hook, not the link
Verify two hashtags at the end
Post
Log it in Posted log with the LinkedIn URL

Future improvements¶

Things worth adding to this playbook over time:

A queue table that auto-populates from journal entries when I detect a customer question
A "topic exhaustion" tracker — when a blog has been pushed 3+ times in 12 months, retire it from the queue
LinkedIn Analytics integration to learn which hook variant performs best for Sush's audience
Templates for other surfaces (Twitter/X, Threads, Mastodon) if Sush ever expands beyond LinkedIn
A "comment reply" voice guide for when posts get engagement

Why this lives in the annex, not in `copilot-instructions.md`¶

Per the Learn-Doc-First Rule (set 10 May 2026, see ~/.copilot/copilot-instructions.md):

The rule in copilot-instructions.md is: "At end of any customer-session journal entry, scan for a thought-advisor candidate and surface a draft."
The depth (formulas, hook variants, voice guardrails, queue, posted log) lives here.

This keeps the always-loaded instructions thin and lets this annex grow as we learn what works.