The complete open toolkit for building an AI-augmented marketing function. Not to replace your team. To help them move faster, structure ideas better, produce more, and turn insight into action.
The best CMOs won't be replaced by AI. They'll be augmented by well-designed AI systems. And the ones who build those systems first will hold an advantage their competitors can't quickly copy.
That's the whole premise of this guide. AI in marketing is past the experiment phase. The question is no longer whether to use it. The question is whether you use it as a toy or build it into how your function runs.
A CMO with a system beats a CMO with a subscription.
Most marketing teams today use AI the way they'd use a search bar. One-off prompts, no shared context, no standards, output nobody trusts. The result looks like activity and produces almost nothing. The teams pulling ahead do something different: they treat AI as infrastructure. Shared context, defined workflows, quality bars, and clear rules for what stays human.
This is the same playbook we build for clients. I’m sharing it because the CMOs who act early win, and most of what’s written on this topic is theory.
This guide shows you how to build that infrastructure with AI, chapter by chapter. Strategy first, then briefs, then workflows, then content, then measurement. It ends with the exact prompts to start today.
Chapter 02
The new role of the CMO in 2026.
The job is shifting from operator to growth architect. You still own brand strategy, editorial calendars, campaigns, content teams, and performance reporting. But the way you run them is changing in four specific ways.
01
From campaign manager to systems builder
The old job was shipping the next campaign. The new job is building the system that ships campaigns: repeatable processes where AI handles the production load and your team handles judgment. Build once, run many times.
02
From gut calls to data-informed decisions
AI makes it cheap to analyze what used to be expensive: every sales call, every support ticket, every competitor move. The CMOs who feed that signal into decisions will out-position the ones still deciding by instinct alone.
03
From team output to team throughput
Headcount stopped being the growth lever. Throughput per person is the new metric. A five-person team with a real AI system outproduces a fifteen-person team without one, and does it with tighter quality control.
04
From tool user to AI partner
Occasional prompting gets occasional results. The shift is treating AI like a senior hire: give it your positioning, your voice, your customer research, and your standards, then hold it to them. Context in, quality out.
Where is your team today? Four stages.
Most teams sit at stage one or two and believe they are at three. The gap between stages is never the tool. It is always the system around it.
Stage 1
Experimenting
Individuals prompt on their own. No shared context, no standards, results nobody trusts.
Stage 2
Adopting
A few tools are official. Output rises, quality wobbles, and nobody can prove the gain.
Stage 3
Systematizing
Shared workspaces, written SOPs, human gates, baselines. Workflows run the same way every time.
Stage 4
Compounding
Every new workflow builds on the last. Throughput per person climbs quarter over quarter. Competitors cannot copy it fast.
Chapter 03
Why AI is changing marketing.
Every AI model can write a passable social post. The real shift is different: modern AI can hold your entire business in its head while it works. Here's where that delivers the most value.
Strategy & Planning
A thinking partner with full context
Load your positioning docs, ICP research, win-loss notes, and last quarter's results into a Claude Project. Now every strategy conversation starts from your reality, not a blank page.
Research & Intelligence
Analysis that used to take weeks
Competitor teardowns, market scans, and voice-of-customer synthesis from hundreds of pages of raw input. The work that justified an agency retainer now runs in an afternoon.
Content Creation
Drafts that sound like you
With a voice guide and examples in project knowledge, AI produces first drafts your editors polish instead of rewrite. The difference between the two is your entire content budget.
Campaigns & Activation
Concept to variants in one session
Campaign concepts, messaging angles, channel adaptations, and test variants generated against your actual strategy. More shots on goal without more headcount.
Analytics & Insights
Reporting that explains itself
Drop in the raw exports and get the narrative: what moved, why it likely moved, and what to do next. Your team stops assembling reports and starts acting on them.
Team Enablement
Your standards, everywhere
Brand rules, messaging hierarchy, and quality bars live in a shared workspace. New hires produce on-brand work in week one because the institutional knowledge is in the system, not in someone's head.
Keep The System Growing
Get the next guides and systems first.
We publish new playbooks, prompts, and working systems as we build them for clients. Drop your email and they land in your inbox before they land anywhere else.
You’re in. The next system lands in your inbox.
Chapter 04
Building a marketing strategy with AI.
Strategy is where most teams use AI least and where it pays most. The process below turns AI from a writing tool into a strategy partner. Run it once a quarter.
01
Load the context
Create a dedicated Claude Project. Add your positioning doc, ICP definitions, pricing, win-loss notes, recent sales call transcripts, and last quarter's performance. Garbage in, garbage out works in reverse too: real context in, real strategy out.
02
Pressure-test your positioning
Ask the model to argue against your positioning from the perspective of your top three competitors and your most skeptical buyer. Where the argument lands, you have work to do. Better to hear it from the model than from the market.
03
Map the buyer's real questions
Have AI extract the questions, objections, and hesitations that appear across your sales calls and lost-deal notes. Your strategy should answer what buyers actually ask, not what the company wants to say.
04
Build the message hierarchy
One core narrative, three supporting pillars, and proof points for each. AI drafts it from your context; you make the calls. This hierarchy becomes the source every brief and campaign pulls from.
05
Turn it into a quarterly plan
Campaigns, content themes, channels, owners, and success metrics, each traced back to a pillar. If an initiative doesn't map to the hierarchy, it doesn't make the plan.
Chapter 05
Creating briefs that don't get redone.
Redone work is the most expensive line item in marketing, and it almost always traces back to a weak brief. Vague inputs produce vague outputs, from people and from AI alike. Fix the brief and you fix the rework.
The one-line job
What this asset must make the audience think, feel, or do. One sentence. If you can't write it, the project isn't ready to brief.
Audience and moment
Who sees this and where they are in the buying journey. A first-touch ad and a late-stage case study video have nothing in common except a logo.
The message pillar
Which pillar from your hierarchy this asset carries, and the specific proof points it should use. This is what keeps a hundred assets sounding like one company.
What good looks like
Two or three reference examples with a note on why each works. Examples do more than a page of adjectives ever will.
What to avoid
Banned claims, tired angles, competitor language, and anything legal flagged last time. Guardrails up front beat corrections after.
Format and constraints
Length, channel specs, deadline, and mandatory inclusions. The boring section that prevents the expensive mistakes.
Brief builder
Fill in the six elements. Get a one-page brief you can copy straight into your workflow.
Name the thing and where it lives. Example: 90-second customer story video for the pricing page.
One sentence: what should the audience think, feel, or do after this? Specific beats broad. "Make late-stage buyers confident that switching is low-risk" works. "Raise awareness" doesn’t.
Who sees it, and where they are: first touch, actively comparing, or ready to buy. Example: Ops leaders comparing three vendors.
The single core message this asset carries, plus the evidence behind it. Example: fastest onboarding in the category, proven by the 3-week Acme rollout.
One or two reference pieces and why each works. Links or names are fine. Example: Slack customer stories, because they use specific numbers and real names.
Claims legal flagged, angles you have worn out, competitor language. Example: no ROI percentages we cannot source.
Channel specs, length, deadline, and mandatory inclusions. Example: LinkedIn vertical video, under 60 seconds, due March 3, logo end-card required.
Where AI fits: keep a brief template in a Claude Project alongside your message hierarchy. Give the AI a two-line request and let it draft the full brief, pulling the right pillar, proof points, and guardrails automatically. Your team reviews and sharpens instead of starting from zero. The brief in chapter 10 does exactly this.
Chapter 06
Automating the workflows eating your team's week.
Your team's calendar is full of work that is repeated, rule-based, and joyless. That work is where automation pays first. The rule: automate the workflow, keep the judgment human.
01
Find the repeats
Ask each person on your team one question: what do you do every single week that follows the same steps? Weekly reports, content repurposing, meeting notes into action items, QA passes, status updates. That list is your automation backlog.
02
Write the SOP once
For each workflow, document the steps, inputs, output format, and quality bar in plain language. Put that document in a dedicated Claude Project. You've now trained a system instead of a person, and it never forgets the process.
03
Connect the data
Use connectors to give the AI direct access to the sources the workflow needs: your drive, your calendar, your analytics exports. Every manual copy-paste you remove is a failure point removed with it.
04
Keep a human at the gate
Automated drafts, human approval. Nothing publishes, sends, or ships without a named owner signing off. This single rule is what separates an AI-augmented team from an AI liability.
05
Measure the reclaimed hours
Time each workflow before and after. Reclaimed hours are the first ROI number you'll report, and the easiest one to defend. Chapter 08 shows you what to do with it.
01
Plan
The SOP and the brief set the job. Both live in the project so the system starts from the same place every time.
02
Draft
AI produces the first version against your context and quality bar.
03
Approve
A named human edits and signs off. Nothing skips this step.
04
Ship
Publish, send, or hand off. Connectors remove the copy-paste.
05
Measure
Log the hours and the outcome. That data feeds the next plan, and the ROI report in chapter 08.
Chapter 07
Accelerating content and campaigns.
Speed without a system produces more mediocre content faster. The goal is a content engine: one strong idea in, a full channel spread out, every piece on-message and in your voice.
Foundation
Put your voice in the system
Build a voice guide: tone rules, sentence rhythm, banned words, and five examples of your best work with notes on why they work. Add it to your content Project. Every draft now starts on-voice.
Pillar → Spread
One asset becomes twelve
Start with one pillar asset: a video, a report, a keynote. AI turns it into posts, emails, scripts, and ad variants, each adapted to its channel instead of pasted across them.
Calendar
Plan a quarter in a session
Feed the AI your message hierarchy and launch dates. Get back a full editorial calendar with themes, formats, and owners, mapped to pillars so nothing drifts off-strategy.
Campaigns
Concept wider, choose better
Generate ten campaign concepts against the brief, kill eight, develop two. The cost of exploring ideas just collapsed. Use that to raise the bar on what ships, not to ship more noise.
AI raises the floor. Your editors raise the ceiling.
The teams winning with AI content aren't publishing raw output. They're using AI to get to a strong first draft in minutes, then spending the saved hours on the parts buyers actually feel: sharper hooks, better proof, real stories. Volume is not the strategy. Quality at volume is.
Chapter 08
Measuring marketing ROI with AI.
If you can't show the return, the budget conversation gets harder every quarter. Measure in three layers, in this order. Each layer is harder to prove and more valuable to show.
01
Efficiency: hours reclaimed
Baseline how long key workflows take today, then measure after. Hours saved times loaded cost per hour is a number finance respects, and you can report it within the first month.
02
Output: throughput and quality
Assets shipped per person, cycle time from brief to publish, and revision rounds per asset. Track quality too: engagement per piece, on-brand pass rate. More volume at lower quality is a loss dressed as a win.
03
Impact: pipeline and revenue
Attribute the way you already attribute: pipeline influenced by AI-assisted campaigns versus the prior baseline. Don't invent a new attribution model for AI. Use your existing one and mark which work the system touched.
04
Report it like an operator
One page, monthly: hours reclaimed, throughput change, quality signals, pipeline touched, and next month's automation target. Have AI assemble it from your exports. The report about AI should be produced by the system it describes.
Reclaimed-hours calculator
Rough out layer one of your ROI case. Move the sliders to match your team.
0
Hours reclaimed per month across the team
$0
Annual value of reclaimed time at loaded cost
Your inputs, your estimate. Assumes 4.33 weeks per month. Reclaimed time only counts if you redeploy it, so pair this with the throughput and pipeline layers above.
Chapter 09
The mistakes to avoid in 2026.
We see the same failures across companies of every size. All seven are avoidable, and all seven are cheaper to avoid than to fix.
01
Tool sprawl without a system
Fifteen AI subscriptions, no shared context, no standards, no owner. Activity everywhere, results nowhere.
Fix: one primary platform, shared Projects, a named owner per workflow.
02
No context, then blaming the model
Generic prompts produce generic output. The model didn't fail. The setup did.
Fix: positioning, voice guide, and ICP research in project knowledge before anyone writes a prompt.
03
Publishing raw output
Unedited AI content reads like unedited AI content. Your audience notices, and your brand pays for it.
Fix: human approval on everything that leaves the building. No exceptions.
04
Automating judgment instead of production
Strategy calls, pricing, crisis response, and brand bets stay human. Automate the load, not the decisions.
Fix: a written line between what the system drafts and what people decide.
05
No baseline, no proof
If you didn't measure before, you can't prove after. Then the program dies in a budget review it should have won.
Fix: baseline your top five workflows this week, before you automate anything.
06
Treating it as a side project
An intern experimenting on Fridays is not a transformation. Systems get built when leadership treats them as infrastructure.
Fix: an executive owner, a roadmap, and time on the leadership agenda.
07
Skipping team training
The gap between teams isn't access to AI. Everyone has access. The gap is skill, and skill is trained, not assumed.
Fix: structured training on real workflows, not a lunch-and-learn about prompts.
Two-minute readiness check
Seven questions, one per mistake. Answer honestly. Nobody sees this but you.
We run AI through one primary platform with a named owner, not scattered subscriptions.
Our positioning, voice guide, and customer research live in shared AI workspaces the whole team uses.
Nothing AI-drafted publishes without a named human approving it.
We have a written line between what AI drafts and what people decide.
We baselined our key workflows before automating, so we can prove the gain.
AI has an executive owner, a roadmap, and time on the leadership agenda.
The team has been trained on real workflows, not just shown a prompt demo.
The systems in this guide run on one thing: a story worth telling.
AI can produce a thousand assets a week. If the message underneath them doesn’t answer why buyers should choose you, you’ve automated confusion. When you aren’t different from your competition, you lose to lowest price.
Why Buyers Say No is the other half of this playbook: how to find the superpower buyers actually pay for, and how to sell it in a market flooded with AI-made noise.
Six prompts, one per system in this guide. Paste them into your AI assistant, replace the bracketed parts with your reality, and start. They work best inside a workspace loaded with your context. In Claude, that’s a Project with your positioning and voice guide.
Personalize the prompts
Fill these in once. Your details drop into the prompts below live, highlighted so you can see exactly what changed before you copy.
Your name, what you sell, and who buys it. Example: Acme, project software for construction teams.
The three names you lose deals to most often, separated by commas.
Your one-sentence answer to why buyers choose you. No written one yet? Use this pattern: we help [who] get [outcome] without [pain].
Prompts updated with your details.
Strategy · Ch 04
Positioning pressure test
Here is our positioning: [paste positioning statement]. We are [company, one line]. Our top three competitors are [names]. Argue against our positioning three times: once from each competitor's point of view, as if they were pitching against us in a deal. Then argue against it once more as our most skeptical buyer. End with the three weakest points in our positioning and one way to strengthen each.
Research · Ch 04
Voice-of-customer synthesis
For context, we are [company, one line]. I'm pasting notes and transcripts from recent sales calls and lost deals: [paste]. Extract every question, objection, and hesitation buyers raised. Group them into themes, rank the themes by frequency, and flag which ones our current messaging fails to answer. Format as a table: theme, example quote, frequency, gap in our messaging.
Briefs · Ch 05
Brief builder
Context: we are [company, one line]. Build a creative brief for: [two-line description of the asset]. Use our message hierarchy in this project. Include: the one-line job of this asset, audience and buying stage, the message pillar it carries with proof points, two reference examples of what good looks like, what to avoid, and format constraints [channel, length, deadline]. Keep it to one page. Flag anything you had to assume.
Workflows · Ch 06
Workflow SOP writer
I'm going to describe a workflow my team repeats every week: [describe steps, inputs, and the final output]. Turn this into a standard operating procedure you can execute: numbered steps, required inputs, output format, and a quality checklist. Then tell me which steps you can do fully, which need my review, and which should stay human. I'll save this SOP to this project so we can run it weekly.
Content · Ch 07
Pillar-to-spread engine
Context: we are [company, one line]. Here is a pillar asset: [paste transcript, report, or link summary]. Using the voice guide in this project, create: 5 short social posts with distinct hooks, 1 email to our list, 1 sixty-second video script, and 3 ad headline variants. Each piece must stand alone, be native to its channel, and carry the same core message. Do not reuse sentences across pieces.
ROI · Ch 08
Monthly AI ROI report
Build a one-page monthly report from this data: [paste hours baselines, hours after, assets shipped, cycle times, and pipeline data]. Structure: hours reclaimed and cost equivalent, throughput change versus baseline, quality signals, pipeline influenced by AI-assisted work, and one recommendation for next month's automation target. Write it for a CFO: numbers first, no hype.
The First 30 Days
Week 1: Baseline and load
Time your five most repeated workflows. Build your first project with positioning, voice guide, and ICP research. No automation yet, just the foundation.
Week 2: First workflow
Pick the workflow that costs the most hours. Write the SOP, run it with AI, keep a human at the gate. Measure the time.
Week 3: Content engine
Run the pillar-to-spread prompt on one existing asset. Compare edit time against your old process.
Week 4: Report and pick the next one
Build the one-page ROI report and put it in front of leadership. Then choose workflow number two. Systems compound from here.
Systems compound. Start with one workflow this week.
Don't try to transform everything at once. Pick the workflow that costs your team the most hours, build it properly with the steps in chapter 06, measure it, and report it. The second workflow is easier. By the fifth, you have a marketing function your competitors can't quickly copy.
Before You Go
New systems, straight to your inbox.
We publish playbooks like this as we build them for clients. Be first in line for the next one.
You’re in. The next system lands in your inbox.
Why Listen To Us
The playbook behind this toolkit has receipts.
We renovate sales and marketing: fix the story, build the assets, run the machine. The systems in this guide come from that client work. A few verified results:
1,000%
Motorola
Traffic growth, with 600% revenue growth behind it.
$2M
RedFox Metaverse
In 40 days, on 1.7M views.
$1M
Netlinks
Closed from the first presentation built on the new story.
$250M+ in client revenue across 1,000+ engagements.
Featured in Forbes, Entrepreneur, Inc., HuffPost, and Success Magazine.
Put It To Work
Want your team trained on Claude?
We train marketing teams to build these systems on their own work, their own voice, and their own workflows. Or start with a free diagnosis and we'll show you where the hours are hiding.
