Start with one business priority and five ideas.
Choose whether the priority is getting more customers, selling more to existing customers, or selling to them more often. List five tasks and score how much each would help that priority.
SMART MARKETER LIVE 2026 · DAY 1 TAKEAWAYS
Russ Henneberry
Choose an AI project that supports your business priority, then turn it into a workflow you can build and test. Use Russ’s scorecard to compare ideas, decide where people should stay involved, and identify the next practical step.
2026-10-20 · AI project selection · Growth × Readiness · Human judgment · Workflow planning
Choose an AI project that supports your business priority, then turn it into a workflow you can build and test. Use Russ’s scorecard to compare ideas, decide where people should stay involved, and identify the next practical step.
PUT THIS SESSION TO WORK
Use Growth × Readiness to compare your ideas, decide what to automate or keep human, and map a practical first test. Choose the activity that fits where you are now.
Choose whether the priority is getting more customers, selling more to existing customers, or selling to them more often. List five tasks and score how much each would help that priority.
Score Growth from 1–5. Score five readiness checks from 1–5: clear steps, low-stakes errors, ready data, connected tools and manageable volume. Readiness is the lowest check, not the average; multiply Growth × Readiness to get GR.
A high score does not automatically mean a task should be automated. Keep people involved where trust, relationships, judgment or service create value. If AI will assist a person, score that combined workflow.
Choose the highest GR among suitable Automate or Assist projects. Check the weakest readiness factor, break ties with the business priority, and name a success metric before moving ahead.
Walk one real example through the workflow with the people who own it. Identify inputs, tools, human decisions and feedback loops. Use an AI interview to fill gaps in the plan, then test and improve the workflow.
Open a choice to see the first question and what you’ll create. Continue the work in your connected Coach.
Compare five ideas against your business priority and choose a suitable first project.
If inputs are missing: Help brainstorm when there are fewer than five ideas. Keep uncertain scores blank and ask for evidence; do not rank incomplete rows as if verified. Do not copy the Northstar example scores or metric into the user’s business.
Complete when: The user has confirmed the scoring, human involvement, chosen project and success metric.
Find what is holding one promising project back and decide how AI and a person should share the work.
If inputs are missing: Use an interview to uncover missing steps or dependencies. Unknown data access, tool capability or error consequences stay unresolved until confirmed.
Complete when: The user has agreed on human involvement and the next step to resolve the limiting readiness factor.
Turn your chosen project into clear steps, tool handoffs and human approval points.
If inputs are missing: A rough description or partial diagram is enough. Interview for missing details; do not assume the user uses Northstar’s ad tools or has granted launch access.
Complete when: The user has checked the workflow, human approvals, unresolved dependencies and first test.
Your Coach will reuse your approved Offer Profile and Persona Profile. If those are missing, it will ask for your business name and website, help draft the profiles, and ask you to approve them before continuing your chosen activity. These activities apply Russ’s handouts and the reviewed closing excerpt to your business. Coaching paths and output templates are Coach-designed applications. This guide does not submit or store answers.
Read all takeaways ↓ · Open session resources ↓ · You can also ask the Coach a question about the session.
No. Readiness is the lowest of the five checks. Multiply that number by Growth. For example, G4 and checks 5, 4, 3, 5, 4 give R3 and GR12.
No. First decide Automate, Assist or Keep human. Rank suitable Automate and Assist projects; a Keep human task is not made eligible by a high score. The filled example gives both refill reminders and VIP thank-you notes a GR of 20, but keeps the thank-you notes human.
Score the AI + human version of the task. The person’s role is part of the design being assessed, rather than scoring a fully autonomous version and adding review afterward.
Use the business priority to break the tie and inspect the weakest readiness factor. Record the chosen project and its success metric.
No. Russ describes using a partial map or a picture of the process as the starting point for an AI interview to fill in the missing details.
The map is an ad-creation example showing tools by stage, ownership and approvals. The Coach can help map the tools and handoffs for your own task; the example does not establish that those integrations already exist in your business.
The transcript is a closing excerpt. The project-selection method and filled example come from the three prepared PDFs supplied separately by the owner. References identify which material supports each lesson.
Worksheet; Worksheet; Worksheet; Talk · 00:25; Talk · 04:40; Talk · 06:29
This index combines three owner-supplied, one-page prepared handouts with the reviewed closing transcript excerpt. Worksheet references point to the PDFs; transcript references open selected, paraphrased teaching with attendee exchanges omitted. It does not imply that the full presentation was recorded.
Choose one of three growth priorities and list five tasks to compare.
Growth anchors are 1 = nice to have, 3 = meaningful improvement, 5 = critical bottleneck; use 2 or 4 between anchors.
Score clear steps, low-stakes errors, ready data, connected tools and manageable volume. Readiness anchors are 1 = major barrier, 3 = work needed, 5 = ready. Use 2 or 4 between anchors.
R is the lowest readiness check, not the average. GR = Growth × R. G4 with checks 5,4,3,5,4 gives R3 and GR12.
Keep people where trust, relationships, judgment or service create value. Score the AI + human version for Assist.
Circle the highest GR among suitable Automate or Assist projects, examine the weakest factor, and break ties with the business priority. Write the project and success metric.
Prepared example: ad angles G3/R4/GR12/S; bundle offers G3/R4/GR12/S; refill reminders G5/R4/GR20/A; product FAQs G3/R3/GR9/S; VIP thank-you notes G4/R5/GR20/H. These are example judgments, not measured outcomes.
The example prioritizes selling more often to existing customers and selects refill reminders. It proposes 60-day repeat-purchase rate versus a holdout. Equally scored VIP thank-you notes remain Keep human.
Media buyer: Define the product, audience, offer, spend and goal.
AI + automation: Gather reviews, support tickets, orders and past ads.
AI + automation: Include customer insight, proof and a citation for each angle.
Creative strategist: Rank relevance and refine the positioning.
AI + automation: Prepare hooks, copy and creative briefs.
Creative strategist: Assemble the assets and check the landing-page fit.
Media buyer: Use the “OK to launch?” decision: yes advances to launch; no returns to step 6 for revision.
Media buyer: Set the spend cap and tracking, then publish.
AI + automation: Join spend and orders by angle and ad ID.
Media buyer: Keep, revise or stop; update the angle log and return to the brief for the next weekly test.
The media buyer owns budget, approvals and the AI/automation lane; the creative strategist owns positioning and assets. Steps 1–3 use Shopify, Gorgias and Meta Ads Manager; 4–8 use Google Docs, Higgsfield, Google Drive and Meta Ads Manager; 9–10 use Meta Ads Manager, Shopify and Google Sheets. These are the supplied example’s tools.
Map the workflow step by step before building it. Identify the people, AI steps and tools involved. A swim lane map is one way to show who does what; another clear process map can serve the same purpose.
In the ad-creation example, a media buyer and creative strategist first define what is being sold, the offer, spend and goal. They then collect customer language from reviews, support interactions and research to inform the advertising.
Drafting ad angles comes after establishing the brief and gathering customer language.
AI can help gather research and propose ad angles, but the creative person should review the proposals and decide which are worth advancing. The reviewer should explain what they like or dislike and why.
Use specific feedback about the proposed angles to revise the instructions or skill that produced them, such as explaining when ideas are repetitive and need more variety.
Break the workflow into small steps and identify which tools each step needs to access. Evaluation includes explaining why an output does or does not work, so the workflow's instructions can be improved.
Capture the process map and ask the AI to interview you for the missing details needed to write a specification. A partially completed map can be the starting point for that interview.
When you have described as much of the desired workflow as you can, ask the AI to interview you about the remaining details. Use the answers to prepare the build, then test the workflow and fix it as you go.
Improve the workflow through testing and specific feedback about what worked and what did not. An evaluation interface can provide a structured way to record the output and the reason it needs improvement.
Choose whether the priority is getting more customers, selling more to existing customers, or selling to them more often. List five tasks and score how much each would help that priority.
Score Growth from 1–5. Score five readiness checks from 1–5: clear steps, low-stakes errors, ready data, connected tools and manageable volume. Readiness is the lowest check, not the average; multiply Growth × Readiness to get GR.
A high score does not automatically mean a task should be automated. Keep people involved where trust, relationships, judgment or service create value. If AI will assist a person, score that combined workflow.
Choose the highest GR among suitable Automate or Assist projects. Check the weakest readiness factor, break ties with the business priority, and name a success metric before moving ahead.
Walk one real example through the workflow with the people who own it. Identify inputs, tools, human decisions and feedback loops. Use an AI interview to fill gaps in the plan, then test and improve the workflow.
SESSION HANDOUT
A one-page worksheet for comparing five potential AI projects.
Why use it: Choose a project that supports your business priority and is practical to implement.
Get started: List five tasks, score growth and the five readiness checks, choose A/S/H, then name your priority and success metric.
One-page PDF supplied by Russ for this session.
SESSION HANDOUT
Northstar Daily Supplements’ prepared comparison of five AI opportunities.
Why use it: See why a high-scoring task can still stay human and how a project earns priority.
Get started: Compare the refill-reminder and VIP thank-you-note rows. Use the reasoning as a guide; score your own business from its evidence.
One-page PDF supplied by Russ for this session.
SESSION HANDOUT
A ten-step example with people, AI tasks, tools, launch approval and a feedback loop.
Why use it: Turn a selected idea into a workflow with clear ownership and handoffs.
Get started: Walk one real example of your task with its owners, then map the steps, human decisions, tool access and next test.
One-page PDF supplied by Russ for this session.
The prepared Northstar Daily Supplements scorecard prioritizes selling more often to existing customers. Refill reminders score G5 × R4 = 20 and are selected for automation. VIP thank-you notes also score 20 but are marked Keep human, so the score alone does not decide what to automate. The proposed success metric is 60-day repeat-purchase rate versus a holdout; the handout does not report a measured result. Worksheet
The Northstar ad workflow uses AI to gather evidence and draft five angles, a creative strategist to choose three, and a media buyer to approve the launch. A rejected launch returns to asset revision. Results feed the next weekly test. This is an example workflow to adapt, not a requirement to use the same tools or run ads. Worksheet