Decisions → owners → what AI may do
AI is already shaping marketing decisions at most security companies, usually without anyone deciding that it should. This map lists the decisions that matter, scores which ones need governing first, and writes the rules for each one: who owns it, what AI may do, and where a person signs off.
To govern AI in a marketing team, organize around decisions, not tasks. List the decisions marketing makes, score each on business value, risk and how often it comes up, and govern the top three to five first. For each one, name a single owner, set how much AI may do (decide, recommend, or only assist), put an approval gate in the tools, and give people and AI the same context: ICP, approved claims, voice and funnel definitions. This tool starts you with 24 decisions security marketing teams make, scores them, writes a decision card for each of your top five, and builds a 90-day rollout. Free, no signup, and it exports to Excel and PDF.
Eight yes-or-no questions. Answer for how your team works today, not how the policy reads.
Twenty-four decisions security marketing teams make, with suggested scores from 1 to 5. Change any score that is wrong for your company, untick what does not apply, and add your own. Priority = value × risk + frequency, out of 30.
| Use | Decision | Value | Risk | Freq | Priority | Tier | AI role |
|---|
The five highest-priority decisions, and where every decision sits on value and risk. Circle size is frequency. Govern the top right first.
One card per decision. Each starts as a draft written for security marketing. Edit it until the owner, sales and finance would all sign it.
People and AI make consistent decisions when they work from the same material. Keep it in one place, such as a Claude Project or a shared drive your AI tools read from. Mark what you have.
Observe, Analyze, Execute, applied to your shortlist. The plan updates as you change the scores above.
Your map exports with everything you entered: scores, cards, context and the rollout. The playbook and the blank workbook are for sharing with your team before you start.
Every decision with live scoring formulas, your top five cards, the context checklist and the 90-day plan.
Readiness, shortlist, the five decision cards and the rollout, formatted to send to your CEO.
The full method in 16 pages: the decisions, how to score them, the AI roles, two worked cards and the rollout. See what is inside.
Download playbookThe same workbook with suggested scores, ready for a team working session.
Download blank workbookYour answers are kept in this browser so you can come back to them. Nothing is sent anywhere.
The thinking behind the map, and the parts worth challenging.
Most marketing teams plan around activities: campaigns, content, events. AI is very good at activities, so it slips into the decisions behind them without anyone choosing that. Organizing around decisions means listing the calls that actually move revenue and risk, such as which buyer to lead with or what outbound AI may send, and setting an owner and rules for each one. The tasks then follow from the decisions.
The ones that combine high business value with high risk, and especially those that come up often, because frequent decisions are where AI quietly takes over. In security marketing that usually includes threat research publication, competitive claims, customer references, what outbound AI may send, category positioning and what goes to the board. Score each decision as value times risk plus frequency, and govern everything at 22 or above first.
One named owner, who is consulted, the inputs the decision needs, what AI may do, what stays human, the approval gate and how often it is reviewed, the line nobody crosses, and how you will know the decision worked. A card is finished when the owner, sales and finance would all sign it.
It depends on the decision. This map uses three roles. Human decides, where AI only gathers and summarizes. AI recommends for approval, where nothing ships until the owner approves. AI acts within guardrails, where AI makes the call inside written limits and every decision is logged and sampled. Start conservative and move a decision up a level only when its error and override log is clean.
Shared context is the material people and AI both decide from: the ICP and buying committee, the positioning statement, approved and banned claims, voice samples, funnel definitions, customer reference rules and data access rules. When it lives in one place, a Claude Project for example, AI drafts from the same facts your team uses, and decisions stay consistent when people change roles.
On revenue outcomes, not only hours saved. Fewer than a third of CMOs say their AI investments are producing the returns they expected, and in Gartner’s 2026 CMO Spend Survey 70% said their processes are not yet mature enough to scale AI. Tie each AI use case to the decision it supports and measure that decision's outcome, such as meetings per 100 outbound sends or win rate in head-to-head deals.
Yes. No signup and no email gate. Your answers stay in your browser. You can export your map as an Excel workbook and a PDF, and download the playbook and a blank workbook for your team. Calibrating the scores with your team, getting the cards signed and setting up the approval gates is work I do with clients.
The map gives you the decisions and a first draft of the rules. Getting the owners to agree, building the shared context and setting the approval gates in the tools is the work that makes it stick.
I run the working session that scores your decisions with the people who make them, write and sign the top five cards, and set up the shared context and approval gates in the tools your team already uses. Then I report the results to your CEO.
Statistics credited to Gartner are from Gartner research for marketing leaders, 2026. The decision list, scoring method, AI roles and cards are Christine Castro's own.