The Plan Was Ready on Day One. The Queue Was Not.
A SaaS company received a complete, task-by-task marketing plan. What happened next is a warning for every company still running marketing by hand.
“The entire plan could have been already executed at the moment we handed it to the team.”
ready to run
their own briefs
we found
A strong product, a capable team and a growth target that would not move
The company is an established SaaS business with a real in-house marketing department: content, SEO, design, video, paid media and outreach, each with its own specialist. Nobody here lacked talent or effort.
The numbers still were not moving. Organic rankings had slipped, paid campaigns were not producing leads and leadership wanted growth fast. They brought us in to find out why and to hand their team a plan they could run.
We are not naming them. The point of this case study is not the company. It is how common their situation is.
A complete operating plan, built with our agentic pipeline
Not a strategy deck. A working plan: every task with an owner, a date, a written brief and the files needed to finish it, including ready-to-upload technical files such as a rewritten llms.txt and video schema.
checklist items, each with an owner, a due date and a written brief
people, each with their own folder of briefs and working files
teams mapped on one deadline calendar
posts in a ready-to-produce launch campaign
None of these is fatal alone. Together, they describe a company moving at the speed of its slowest queue.
- Sign 1
Rankings fell after a platform migration
Moving to a new content platform reset years of search authority, and a later algorithm update deepened the drop. Recovery was happening one keyword at a time while the market kept moving.
- Sign 2
AI models got a sitemap, not an answer
The llms.txt file was generated by a CMS plugin as a bare list of links. It never said what the product does or who it serves, and the plugin rewrote the file on a schedule, erasing any hand fix.
- Sign 3
The best videos were invisible to machines
Customer videos had no voiceover, no transcript and no VideoObject schema. AI systems cannot watch video yet, so the strongest proof the company had did not exist to them.
- Sign 4
Local trust signals were missing
The Google Business Profile had no verified address and only a handful of reviews: a small detail to a visitor and a real signal to a search engine.
- Sign 5
The demo account could not produce marketing
The account marketing used for screenshots and product video was cluttered and had no sample data. The product was strong. The team could not show it.
- Sign 6
Paid channels changed before they could learn
Campaigns were paused on weekends, judged after a month and swapped for the next channel. No channel ran long enough to produce a real answer.
- Sign 7
A product launch kept slipping
A planned developer launch had already moved once, and every week of delay put its first-mover advantage at risk.
- Sign 8
Every fix waited in a queue
Items went from marketing review to development to IT, with logins and access requests in between. Nobody could see work in progress, so finished tasks looked missing and missing tasks looked done.
The handoff tax
The work in the plan was small. Writing a page takes hours. Updating a schema block takes minutes. The delay lived in the spaces between people: waiting for review, waiting for access, waiting for context and waiting on each other.
The plan entered the org chart, and every item started its own trip through marketing review, the development queue and IT. Work that was already in progress somewhere could not be seen from anywhere else.
“We can't see what you guys are actively working on. We only see the last, you know, screenshot, essentially, that you shared with us.”
The number that exposes this is fix-to-live time: the days between finding a problem and having the fix live. When a two-hour task takes three weeks, the task is not the problem. The system around it is.
What a 9.2 vulnerability teaches about marketing
On October 21, 2026, The Hacker News reported emergency patches for two PHP flaws, CVE-2026-109618 and CVE-2026-109619, each rated 9.2 on CVSS v4.0 and affecting nearly every PHP version back to 5.4. Exploit attempts were reported, and the PHP team's guidance was that only updating fully protects a server.
WordPress, which runs a large share of marketing sites, is built on PHP. We are not saying this company was affected. We are saying the habit that leaves a marketing fix in a queue for three weeks is the same habit that leaves a critical patch there too.
One costs you rankings. The other can cost you the company.
Fix it today. Run it next week. Test it by Friday.
Had we controlled the pipeline end to end, this is how the same plan would have run. It is our assessment of the work, not a reported result: the fixes were small, already written and ready to ship.
- Day one
Identify and fix
The rewritten llms.txt goes live and the plugin stops overwriting it. Video transcripts, summaries and schema are generated and published. The brand-defense search campaign is switched on.
- The next week
Execute everywhere at once
The 20-post launch campaign ships across every channel, already cut for each format: square for LinkedIn and Facebook, vertical for Reels and short-form video, widescreen for YouTube.
- End of that week
Test with real data
Tracking was in place before launch, so Friday produces answers. The next round of changes is chosen by what the audience did, not by what a meeting preferred.
AI agents handle the repeatable production work. People keep strategy, creative direction, brand voice and final approval. The difference is not effort. It is that the plan and the execution live in the same hands.
Boards do not read llms.txt files. They read the pipeline number.
This part is our point of view, not a statistic. When qualified leads lag one quarter, it is a marketing problem. When they lag three quarters in a row while competitors ship faster, it becomes a leadership problem, and the conversation moves to the corner office.
The market is already asking harder questions about speed. The satirical “Great Cooking” investigation of RSA Conference 2026 asked “How many RSA exhibitors could be replaced by a weekend of vibe-coding?” It is labeled satire and should be read that way, but buyers and investors now ask a version of that question about every company.
No one in this case study has lost a job. The value of a warning sign is that you see it while there is still time.
Seven things a leadership team can check this week
- 1
Patch the runtime today
If your site runs PHP, confirm it is on a fixed release and that the servers were restarted. Do not schedule it.
- 2
Read your llms.txt
If it is only a list of links, rewrite it to say what you do, who you serve and why you win, and stop any plugin from overwriting it.
- 3
Give every video words
Add a transcript, a short summary and VideoObject schema to every page that carries video.
- 4
Fix the trust basics
Verify your Google Business Profile and ask happy customers for reviews.
- 5
Let tests finish
Give each paid channel long enough to produce an answer before you switch.
- 6
Put all work on one board
Your team and your partners should see the same list, in real time.
- 7
Measure fix-to-live time
Track the days between finding a problem and shipping the fix. That number is your real speed.
Questions about this case study
Who is the client in this case study?
An established SaaS company with a capable in-house marketing team. We do not name clients in warning-sign case studies, and the details here are limited to what helps other teams recognize the same patterns.
What is fix-to-live time?
The number of days between identifying a problem and having the fix live for customers. It captures approvals, access, queues and handoffs, which makes it a more honest speed metric than output volume.
What is PHP2Shell?
The nickname for two critical PHP flaws, CVE-2026-109618 and CVE-2026-109619, patched on October 21, 2026. Both are rated 9.2 on CVSS v4.0, and updating to PHP 8.5.12, 8.4.27, 8.3.36 or 8.2.35 is the only complete fix.
Was this company affected by PHP2Shell?
We are not claiming it was. PHP2Shell appears here because it shows the same organizational habit: a critical fix that can be applied in minutes still waits when every change has to pass through a queue.
What does agentic execution mean for a marketing team?
One team owns the pipeline from finding a problem to shipping the fix. AI agents handle repeatable production work such as drafting, formatting, transcription, schema and reporting, while people keep strategy, creative direction and final approval.
If someone handed your team a complete plan tomorrow, how much would be live by Friday?
If the honest answer is “not much,” that is the problem worth fixing first. We run marketing pipelines end to end, from the audit to the live fix to the Friday test.