01 · How to use this page
Pick a row. Make one ad. Learn one thing.
Use this whenever you need a new ad or social post. Pick one customer problem, make one version of the ad, and learn one useful thing from it.
Your 10-minute workflow
Start at a row in the table below. Follow these four cards in order.
- 1
Choose a customer moment
Pick one row where the work situation feels familiar. Start with the awkward, frustrating moment—not the feature.
- 2
Make one ad from it
Use the row’s opening line, image or meme idea, and customer benefit. Together, they tell you what to make.
- 3
Decide what you are testing
Write one question: “Will this audience respond better to this hook or that hook?” Keep everything else the same.
- 4
Write down what happened
After it runs, note the result and update the row. The next person should be able to see what we already learned.
What one completed row looks like
Example: the customer has to explain the same project again
Customer moment
They are typing the same background into a new tool for the third time this week.
Make this ad
A familiar “here we go again” work meme, then: “Your assistant should not make you start from zero.”
Test this
Keep the message. Try one meme visual against one simple product/demo visual.
02 · What customers are telling us
Start with the customer’s problem—not the product feature.
Use these comments to choose a real problem for the ad. They can guide the message, but they do not prove a ProductiveBot claim on their own.
| Customer problem | What they say | How it feels | How to use it in an ad | Source |
|---|---|---|---|---|
| Re-explaining context | “I have to explain the same thing again.” | Drained, dismissed, behind | Show the repeated explanation. Then show an assistant that already has the needed context. | X Research, 2026-08-14 |
| Technical setup is a second job | “This is absurdly complicated.” | Overwhelmed, intimidated, stuck | Show an owner who needs work done, not more setup work. | X Research plus ProductiveBot paid-media audit |
| AI output does not equal completed work | “It gave me a draft, now I still have to make it happen.” | Skeptical, let down, cautious | Show a conversation turning into one clear next step for the person to review. | X Research, 2026-08-14 |
| Business data feels exposed | “I do not know what happens to my client information.” | Protective, uneasy, risk-aware | Lead with ownership and control. Explain what happens with cloud models when it matters. | Web research, 2026-08-14 |
| Admin is scattered everywhere | “It is in my inbox, messages, notes, and head.” | Drowning, guilty, worried about missed opportunities | Show the messy work first. Then show one conversation creating an organized next step. | X Research plus ProductiveBot content voice playbook |
03 · Ad ideas to test
Choose a Start here row. Make one ad. Use what happens to choose the next test.
Each row tells you who the ad is for, what work problem to show, what you can say, and what to test next. Keep the ad about one real customer problem—not AI in general.
Customer type
How familiar the customer is with the problem
Proof state
| Row | Product fact | Customer situation | What the ad says | Visual idea | Good formats | Proof and limits | Next test | Ready? |
|---|---|---|---|---|---|---|---|---|
| CM-01Start here | Guided setup wizardVerified | Solopreneur with a StoreWhat is happeningThey have watched AI demos, but every tutorial turns into a technical project.How it feelsIntimidated and annoyed that getting help requires becoming technical.Problem familiaritySolution aware | Skip the setup taxOpening line“You wanted help with the business, not another system to set up.”Customer benefitA working AI assistant starts from a guided setup instead of a blank technical project. | Split screen: 17 tutorial tabs and a terminal on one side; a calm first conversation on the other. Meme option: the 'I just need it to work' reaction image over the tutorial-tab chaos. | Founder video; before/after static; reaction meme | Setup Wizard; product truth confirms under-five-minute setup claim.Keep it accurateUse the under-five-minute claim only with the approved product-truth wording. Do not imply every custom workflow is configured in that time. | Test pain-led setup hook against outcome-led first-workflow hook. | Brief ready |
| CM-02Start here | TotalRecall memoryVerified | Former Executive Going IndependentWhat is happeningThey keep pasting brand details, client context, and decisions into a new chat before useful work can begin.How it feelsDrained by repetition and irritated that the tool feels forgetful.Problem familiarityProblem aware | The re-explaining taxOpening line“If you have typed 'as I said before' into AI, you are doing unpaid onboarding.”Customer benefitPersistent business context reduces the need to start from zero each time. | Recurring calendar event titled 'Explain my business to AI again' versus a message that begins with relevant context. Meme option: a groundhog-day loop of the same prompt pasted again. | Screen-led demo; carousel; meme with explainer caption | TotalRecall persistent memory system; feature is verified.Keep it accurateSay it reduces repeated context, not that it remembers everything. Never expose a real customer's data in a demo. | Compare frustration language ('again?') with time-reclaimed language ('start at the useful part'). | Brief ready |
| CM-03Start here | Scheduled / recurring workVerified | Performance MarketerWhat is happeningThe weekly report, research pass, and follow-up queue are all important but keep slipping behind urgent work.How it feelsGuilty, reactive, and tired of being the reminder system.Problem familiarityProblem aware | The work that waits until you rememberOpening line“Your calendar is full. The important repeatable work is still waiting for you to remember it.”Customer benefitA defined recurring task can run on a schedule and surface work for review. | A weekly task moving from a buried sticky note into an automated morning summary. Meme option: a person holding 14 reminder strings while a calendar laughs. | Faceless workflow demo; founder story; short static sequence | Scheduled and recurring work is verified.Keep it accurateShow the exact scheduled task and the review point. Do not call ProductiveBot an employee or imply actions happen without configured permissions. | Test missed-revenue framing against calm-monday-morning framing. | Brief ready |
| CM-04Start here | Multi-channel messagingVerified | Trades / Service OperatorWhat is happeningThey are on the road between jobs when a customer question, invoice issue, or follow-up needs attention.How it feelsPulled in every direction and worried something important will be missed.Problem familiarityProblem aware | The business does not wait for desk timeOpening line“The customer called while you were driving to the next job. The follow-up should not have to wait for tonight.”Customer benefitYou can work through the channels where you already communicate instead of waiting to get back to a desk. | Truck cab voice note or phone message becomes a clear, reviewable follow-up plan. Meme option: office productivity advice shown next to a muddy work boot and a ringing phone. | POV video; customer-story edit; static meme | Multi-channel messaging is verified; use only supported channels in the demonstrated configuration.Keep it accurateDo not imply automatic sending. Show that the operator chooses what goes out. | Test truck scenario against after-hours customer-response scenario. | Brief ready |
| CM-05Start here | Approval gates / human in the loopVerified | Privacy-Bound ProfessionalWhat is happeningThey want help drafting sensitive work, but the thought of an AI sending something on its own makes them hesitate.How it feelsProtective, cautious, unwilling to risk reputation.Problem familiaritySolution aware | Help without handing over the wheelOpening line“Useful AI should make the next move easier, not make the final call for you.”Customer benefitProductiveBot can prepare work for review while the person stays in control of consequential actions. | A draft with a clear review checkpoint before send. Meme option: a car with an AI passenger-seat driver trying to grab the wheel, then the owner calmly holding it. | Screen demo; privacy carousel; founder explanation | Approval gates and human-in-the-loop controls are verified.Keep it accurateBe specific about the workflow and approval point. Do not promise universal protection or automatic policy enforcement. | Test control-forward message against risk-avoidance message. | Brief ready |
| CM-06Start here | Mac mini (owned hardware)Verified | Privacy-Bound ProfessionalWhat is happeningThey want AI help with real business information but do not want their working system to be a mystery in someone else's cloud.How it feelsUneasy, protective, unwilling to trade control for convenience.Problem familiaritySolution aware | Own the machine that does the workOpening line“Your business deserves more than a browser tab you rent by the month.”Customer benefitThe customer owns a physical AI workstation that remains in their possession. | An owned Mac mini on a desk with a simple 'your hardware, your workflow' treatment. Meme option: 'Where is your AI actually living?' map pin joke contrasting a mystery cloud with the owner's desk. | Product film; static visual; privacy explainer | Mac mini owned hardware is verified.Keep it accurateExplain cloud-model boundaries when relevant. Do not say data never leaves the device if a cloud provider is used for a prompt. | Test ownership framing against privacy-and-control framing. | Brief ready |
| CM-07Start here | Scout (support + guidance)Verified | Solopreneur with a StoreWhat is happeningThey like the idea of AI, but fear being stranded the first time something technical stops working.How it feelsNervous, unsupported, afraid of wasting the purchase.Problem familiarityProduct aware | The fear behind 'I am not technical'Opening line“The real reason most people do not try AI agents: they do not want to be alone when it breaks.”Customer benefitScout provides built-in AI support and step-by-step guidance for common issues. | A real support conversation resolving one clearly bounded issue, with the human still driving. Meme option: a tiny business owner facing a giant error dialog, then a calm guide appears beside them. | Customer testimonial clip; screen-led case study; reaction meme | Scout is verified. Internal evidence includes a documented customer troubleshooting experience.Keep it accurateDo not promise human pickup, a response SLA, or universal resolution. Use customer testimony only with approved permission and exact wording. | Test fear-of-breakage opener against proof-led customer-resolution opener. | Needs approved customer asset |
| CM-08Later | Skill StoreVerified | Technical BuilderWhat is happeningThey worry a system will be obsolete as models and useful workflows change.How it feelsSkeptical, future-conscious, tired of re-buying tools.Problem familiarityProduct aware | Do not buy a frozen toolOpening line“The worst time to buy an AI system is right before the useful part changes.”Customer benefitNew capabilities can be added after purchase through the Skill Store. | A simple then/now capability shelf that grows over time. Meme option: someone buying a new phone every time an app updates versus installing a new skill. | Founder explanation; product walkthrough; static comparison | Skill Store is verified.Keep it accurateDo not promise a specific future skill, free access forever, or that every skill works with every setup. | Test future-proofing against buy-once ownership language. | Brief ready |
| CM-09Start here | Content MachinePartial | Performance MarketerWhat is happeningThey have ideas and recordings but lose momentum when every channel needs a different asset and someone has to coordinate it all.How it feelsBehind, creatively depleted, frustrated by unfinished drafts.Problem familiarityProblem aware | One idea should not die in a notes appOpening line“The hard part is not having an idea. It is getting it from the voice note to something your team can actually use.”Customer benefitA rough idea can become a production-ready short-form brief for human review. | Voice-note waveform becomes a clean brief, then a human marks it up. Meme option: an idea trapped in a notes app behind yellow police tape. | Screen demo; creator POV; carousel | Content Machine is partial and not fully verified end to end.Keep it accurateInternal concept only until end-to-end status changes. Do not run public paid creative that claims availability or outcomes. | Hold for private creative development and verification. | Internal only |
| CM-10Start here | Real workflow completionVerified | Former Executive Going IndependentWhat is happeningThey use ChatGPT for drafts but the follow-through still lives in scattered tabs, reminders, and their own head.How it feelsLet down by surface-level productivity, hungry for momentum.Problem familiaritySolution aware | A draft is not a doneOpening line“If AI gives you a draft and you still have to remember every next step, it did not take the work off your plate.”Customer benefitProductiveBot helps turn conversations into completed work across the tools and workflows people already use. | A conversation moving through a simple three-step chain: decision, ready-to-review action, outcome update. Meme option: a finished-looking cake that is still raw in the middle. | Workflow demo; founder monologue; before/after carousel | Core positioning. Pair each execution claim with a demonstrable, specific workflow.Keep it accurateNever promise autonomous completion. Name the workflow, available integration, and review boundary. | Test anti-chatbot hook against business-momentum hook. | Brief ready |
| CM-11Later | Model and engine neutralityVerified | Technical BuilderWhat is happeningThey have invested in tools before, only to be forced into a new platform or pricing change when the market shifts.How it feelsCynical, locked-in, protective of past investment.Problem familiarityProduct aware | Keep the system, choose the engineOpening line“AI changes every month. Your entire business setup should not have to.”Customer benefitThe system can work across supported model and engine options rather than relying on one vendor's roadmap. | A stable workstation in the center with model names orbiting around it. Meme option: a person rebuilding a house every time a lightbulb is upgraded. | Technical explainer; static comparison; founder POV | Model and engine neutrality is verified.Keep it accurateAvoid exhaustive compatibility claims. State that supported options evolve and verify the current setup. | Test lock-in pain against buyer-control message. | Brief ready |
| CM-12Start here | TotalRecall memory + approval gatesVerified | High-volume sales professionalWhat is happeningThey leave calls with promises, context, and next steps, then lose track of who needs an answer first.How it feelsAnxious about missed opportunities and embarrassed by late follow-up.Problem familiarityProblem aware | The follow-up should not depend on your memoryOpening line“You did not lose the deal on the call. You lost it three days later when the follow-up disappeared.”Customer benefitA defined workflow can turn call context into prioritized, ready-to-review follow-ups while the rep retains control. | Call notes, inbox, and CRM fragments collapse into an ordered follow-up queue. Meme option: an unopened calendar reminder wearing a tiny sales trophy. | Workflow demo; sales pain static; testimonial clip | Adjacent ICP signal in product truth; TotalRecall and approval controls are verified.Keep it accurateTreat as a targeted workflow concept. Do not claim CRM automation or customer results without a verified implementation and evidence. | Test missed-deal tension against calmer-pipeline-control message. | Needs workflow proof |
Nothing matches those filters yet. Add a row only when you have a different customer situation to test.
04 · Record what you learned
An ad only helps if the result tells you what to try next.
After it runs, record the basics in the row, ad brief, or campaign log. Views alone do not tell you what worked.
Date launched
Channel / placement
Audience / awareness
Control ID
Creative asset URL
Spend / impressions
Thumb-stop or 3-second hold
CTR / landing-page view
Qualified action
CPA / revenue signal
Winner, loser, or inconclusive
What changed next
Make a clear call
Mark the result winner, loser, or not enough evidence yet. Then choose one next step: keep the customer type and try a new image, keep the image and try a new opening line, or stop using that message. This turns one ad into useful knowledge for the next one.
Research and claim ledger
External research gives us customer language and possible problems to explore. It does not prove ProductiveBot claims. For paid ads, only product truth and the Value Map tell us what we can promise.
data/positioning-map.json: ProductiveBot feature, stakeholder, and proof status sourcedocs/product-truth.md: canonical product and claim sourcememory/paid-media-foundation.md: paid-media audit and proven content signalsdocs/productivebot-content-creation-guide-2026-07-27.md: content approval and proof requirementsdocs/productivebot-content-voice-playbook-2026-07-21.md: voice and customer-moment guidanceX Research, 2026-08-14: recurring SMB AI friction themes and languageWeb research, 2026-08-14: SMB AI skills, privacy, trust, and ROI barriers