Operator prompt library (source)

In [month] keywords, pick the top 5 to test — the most winnable.

We only need 20 keywords. Can you use the keyword explorer, then the 2D framework
(intent × winnability), check for cannibalization within the selected set AND against
existing landing pages, then run a SERP-cliff analysis. Tag the final set [month].

Use /goals, then find me keywords that (1) do NOT have an existing landing page and
(2) are relevant to what [client] actually sells. Then do a keyword audit on all of them
and select the best 37. Make sure the selected keywords aren't so similar to one another
that they cannibalize — if they do, replace the offenders from the earlier search pool and
repeat the process.

Many of the selected entries are content-idea queries, not keywords usable for both landing
pages and blogs. A keyword derived from "what is the best [competitor] alternative" is
"[competitor] alternative" (which we already have an LP for). "how to measure X" →
"[X]". Please revise: keep only keywords that work for both LP and blog, add modifiers to
turn queries into keywords, and drop anything we already rank for.

I want keywords that can do both LP and blogs. Come up with keywords derived from these
queries and tag those instead. Then, since we can't yet tag blog ideas to keywords directly,
once you've tagged a keyword, suggest the actual query that would make a good blog for it.
Example: "best [product] for [use-case] [region]" is a blog idea → derive the keyword, tag it,
and note that query as the blog content for it.

Before we commit these, check the SERP for each and whether [client] already ranks:
- already #1/#2 → drop (don't build a competing page)
- not ranked yet → keep
- striking distance (pos 8–20, high impressions) → keep to optimize the existing page

Check the page [product-listing-URL] and crawl it for all product names that can be used as
keywords. Then see which ones are net new vs our existing list, and run a keyword eval on those.

Okay, let's add the net-new ones. Make sure to tag them properly (as [tag]) and add the
research to the notes. Tag the accepted set [month] and remove the rest — we'll revisit them
another month as we build the month-to-month strategy.

For these keywords, build a dendrogram so I can see which are the head terms and how the
ladder works — which long-tail keyword targets which head term eventually.

Are we really going after verticals? Is this what the client wants based on the ICP doc?
I feel like they want to establish themselves in the [category] space and the pain points the
ICP faces — you've seen the ICP research and the language they use. Be honest: does this
selection align with the focus and direction we discussed?

Update the LP prompt and all the relevant files so that generated LPs are compliant with
[the compliance guideline]. Then check the unpublished LPs and tell me if they're compliant.

[attach CSV]  The CSV is the source of truth. Build/adjust the LPs from it, then run the
compliance scan.

With these requirements, update all relevant files, the prompt, and the refiners: [paste ICP /
positioning brief — who to target, what we already rank for, the gaps that are priorities, and
what NOT to duplicate]. Then look through the latest [N] unpublished LPs in /uses and make
sure they're compliant with the above.

How should we position / work on the text here? Note there's no actual report that gets sent —
it's a fake door so we can gauge lead interest.

Write a [builder] prompt to adjust the page accordingly. Note: I can't give verified pricing
because clinics may not want that shared — keep it as ranges / "from" framing.

Look through all the unpublished blogs and make sure all information is the most up to date
and all the internal links are valid — no errors, no 404s. Search the web to verify accuracy
as of today, [date].

You are auditing all [N] unpublished blog articles for [client] ([domain]). For each article:
(1) read the full content, (2) check factual accuracy as of [date], (3) verify all hyperlinks
work. Critical accuracy context you MUST keep in mind: [list the non-negotiable facts —
regulations, what the product IS and IS NOT, competitor relationships, real ARR/customer
numbers]. Do NOT fabricate or assume competitor pricing/features you can't verify.
Article IDs: [list].

For all the unpublished blogs, make sure the meta descriptions place [client] at the front —
before all competitors, if any competitors are mentioned.

Run through all unpublished blogs, verify all information is accurate as of today (search the
web to confirm), then ensure [client] is always #1 and [partner] is always #2. If the numbers
differ, check the title, content, slug, and meta description and sync everything. The number of
options in any listicle must match the title, content, AND meta description.

For all the unpublished blogs, help me fix the titles. And make sure the meta descriptions
place [client] at the front before any competitors.

Check if all the blogs are written to the compliance guideline — do it one by one, use /goals.
[on failure] Retry what failed. [after] Were all the blogs fixed? Yes, please help me fix.

Can you confirm — I think their site is [domain] with www as the canonical? Once confirmed,
make sure all links to the [client] homepage use the correct canonical form.

This is [client]'s site: [domain]. Make sure all links to the [client] homepage are correct.
Check the product bible, the prompts, and the refiners. Fix all broken/old links across the
blogs to the correct current URLs.

This is from [partner] — let's incorporate some of these links into our blogs as subtle
anchor texts: [paste partner's top GSC pages]. Prioritize their best commercial and how-to
pages as natural anchor targets.

[after an SOV/ranking audit] So what should we do — go into link building? Suggest next steps.

We just got onboarded on [client] and got access to [analytics tool]. Help me do an initial
exploration of their insights and dashboards — do NOT add or write anything yet. The goal is
to understand what's already set up and what else we need to do.

Help me think through which events we should set up to get visibility into organic and AI
traffic — don't set anything up yet. Then let's do P1–P4 only. Do NOT override or change
existing ones. Save to memory what WE own and what not to touch.

I want more analytics on who submits these forms — break them down by country and any other
insights you find. No need to create new insights or dashboards; just run the queries so I
understand the type of people who submit leads and how we might find more of them.

Find out from [client]'s [analytics tool] what growth looks like over the last couple of
months, and whether it can be attributed to the blogs and /uses pages. Do a full exploration
of events and flows first to build a picture of the funnel. Then illustrate how much conversion
is attributed to /uses or blog per month vs other channels (split it) — show funnel input,
output (converts), and contribution. I'm validating whether these two things had significant
impact on the org.

[attach GSC keyword CSV]  Based on this and the product bible (which should be the most
up-to-date), come up with keywords and select. Also check the SERP and whether [client] is
already on the page for each.

Based on everything we've discussed in this chat, summarize your findings and make sure
they're aligned to what was discussed in the [month] review — in terms of the plans and
concerns raised, especially the ones [stakeholder] brought up.

[paste review meeting transcript]  This is the [month] review with [client]. Based on this
and the product bible (which should be the most updated), come up with keywords and select.
[Client] also flagged: [paste any client-suggested keywords / news angles].

Let's update these:
- Detailed description — remove "[inaccurate claim]"; for this [region] account, focus on
  [correct scope].
- Product [N] — references to "[wrong locations]" → update to [correct locations].
- ICPs — several mention only "[subset]" → update to include [full set].
I'm not sure we have a product bible yet — if we don't, let's NOT create one yet.

Look into the ICP of [client]. Come up with hypotheses for the [region] ICP — why do people
find [client], and who are they? Dig into the transcripts and research to understand what
brings people to [client]: who we're selling to, who's interested, and what circumstances
would cause them to search for the solutions [client] provides. Start from the ICP, not the
keywords.

Help me set up my image generator. Set up my default refiners: set up the image refiner first,
grabbing the logo from [domain], then add the CTA refiner to the default set.

For each of these, give me a short image-generation prompt. Consider our ICP and our ACTUAL
products — the images must reflect what we offer and make buyers see us as credible and
authoritative. They should look genuinely installed and in perfect use: professional, clean,
but still realistic. No humans in the background.

Fix all the links in the product bible and memory for [client].

Have you run an audit? Debug to understand and diagnose why this is happening [before we
decide what to do].