42Prompts
Get the free guide

Business / D2C Founder OS / Edition 2026.1

42 Prompts D2C Founders Must Know

Turn scattered AI chats into a reusable operating system for understanding customers, building offers, creating campaigns and learning from results.

Buy in India — ₹299Buy internationally — $9Preview six complete workflows

One-time digital purchase: ₹299 in India or $9 internationally. International payment activates after Razorpay approval. Seven-day digital return window.

101digital-first A4 pages
7 × 6systems × workflows
6full public previews
2026.1versioned release

The operating journey

Seven systems. Six decisions each.

01

Foundation

  • 01Define the business objective
  • 02Map AI opportunities
  • 03Build the Business Context File
  • 04Codify brand truth and voice
  • 05Create the approved-claims register
  • 06Define QA and safety rules
02

Customer intelligence

  • 07Define the ideal customer profile
  • 08Map jobs to be done
  • 09Mine customer reviews
  • 10Mine support tickets
  • 11Map competitors and alternatives
  • 12Synthesize customer insight
03

Product and offer

  • 13Develop the value proposition
  • 14Choose positioning
  • 15Design offer architecture
  • 16Form a pricing hypothesis
  • 17Design a product bundle
  • 18Write the product-page brief
04

Campaign and creative

  • 19Define the campaign objective
  • 20Build an angle bank
  • 21Create a hook matrix
  • 22Draft ad-copy variants
  • 23Plan an organic content series
  • 24Brief an AI video creative
05

Conversion and retention

  • 25Design the landing-page flow
  • 26Create a lead-capture exchange
  • 27Build approved WhatsApp replies
  • 28Design cart-recovery messages
  • 29Create post-purchase onboarding
  • 30Plan retention and win-back
06

Operations and decisions

  • 31Turn work into an SOP
  • 32Build a support-response library
  • 33Review inventory and assortment
  • 34Prepare a vendor handoff
  • 35Write the weekly metrics narrative
  • 36Create a decision memo
07

Learning and scale

  • 37Design an experiment
  • 38Diagnose creative performance
  • 39Run a postmortem
  • 40Update the knowledge base
  • 41Audit an automation or agent
  • 42Build a 90-day AI operating roadmap

No teaser fragments

Six complete workflows to judge the product.

Open a workflow. The paid guide applies this same contract to all forty-two jobs.

03Build the Business Context FileA compact, reusable source of truth that improves every later workflowFull preview

Inputs

  • Verified brand and product facts
  • Current customer and channel knowledge
  • Claims, policies and operating constraints

Example / Nivasa Home

Input: Nivasa sells one ₹1,299 organiser. Washable claim is tested. Recycled-cotton percentage awaits certificate.

Output: Approved: washable per care test dated 12 Aug. Prohibited pending evidence: '70% recycled cotton'. Unknown: repeat-purchase window.

Master prompt

You are the operating partner for a direct-to-consumer brand. Your job is to help a founder make a usable decision, not to produce generic marketing prose. BUSINESS CONTEXT Use only the supplied Business Context File. Treat missing facts as unknown. Label every assumption and never invent product performance, customer proof, certifications, scarcity or legal claims. JOB Transform the supplied facts into a structured Business Context File. Preserve source wording for approved claims, identify conflicts and leave unknown fields blank. INPUTS - Verified brand and product facts - Current customer and channel knowledge - Claims, policies and operating constraints METHOD 1. Restate the decision in one sentence. 2. Separate evidence, interpretation and assumptions. 3. Complete the analysis required for this job. 4. Produce the deliverable in the exact requested structure. 5. End with the next action, owner and earliest useful measurement. REQUIRED OUTPUT Brand facts; audience; product table; positioning; voice do/don't examples; approved and prohibited claims; channels; constraints; unknowns; last-reviewed date. QUALITY GATE - Facts, hypotheses and unknowns are visibly separated - Every sensitive claim has an evidence reference - File is short enough to paste into a new AI session If information is missing, ask no more than three high-leverage questions, then provide a clearly labelled provisional version.

Quality checks

  • Facts, hypotheses and unknowns are visibly separated
  • Every sensitive claim has an evidence reference
  • File is short enough to paste into a new AI session

Next action

Assign one owner and review the file whenever product, policy or positioning changes.

Common failure: Filling gaps with plausible copy and quietly turning guesses into brand facts.

09Mine customer reviewsA traceable language bank of pains, outcomes, objections and proofFull preview

Inputs

  • Anonymised first-party reviews
  • Public competitor reviews with URLs
  • Date range and product scope

Example / Nivasa Home

Input: 27 organiser reviews; 11 mention clearing the dining table and 5 mention soft sides collapsing.

Output: Strong job language: 'clear the table in one trip.' Product risk: structure. Do not claim 'holds its shape' until validated.

Master prompt

You are the operating partner for a direct-to-consumer brand. Your job is to help a founder make a usable decision, not to produce generic marketing prose. BUSINESS CONTEXT Use only the supplied Business Context File. Treat missing facts as unknown. Label every assumption and never invent product performance, customer proof, certifications, scarcity or legal claims. JOB Code review evidence without blending sources. Extract exact short phrases, recurring themes, counter-evidence and confidence. INPUTS - Anonymised first-party reviews - Public competitor reviews with URLs - Date range and product scope METHOD 1. Restate the decision in one sentence. 2. Separate evidence, interpretation and assumptions. 3. Complete the analysis required for this job. 4. Produce the deliverable in the exact requested structure. 5. End with the next action, owner and earliest useful measurement. REQUIRED OUTPUT Theme table; frequency; short verbatim phrase; source; stage of journey; implication; counter-signal; five copy inputs—not finished claims. QUALITY GATE - Every quote is traceable - Frequency and sample size are visible - Negative and contradictory evidence is retained If information is missing, ask no more than three high-leverage questions, then provide a clearly labelled provisional version.

Quality checks

  • Every quote is traceable
  • Frequency and sample size are visible
  • Negative and contradictory evidence is retained

Next action

Add validated language to the insight bank and route product risks to the owner.

Common failure: Using a handful of vivid comments as if they represent the entire market.

17Design a product bundleA customer-journey bundle with clear incremental value and margin logicFull preview

Inputs

  • Products and attach-rate history
  • Jobs customers complete together
  • Inventory, shipping and margin constraints

Example / Nivasa Home

Input: Organiser plus cable pouch and washable desk mat.

Output: Remote-work reset bundle: carry, connect, clear. Test ₹1,899 vs individual ₹2,147; verify shipping weight and whether mat increases the same job's completion.

Master prompt

You are the operating partner for a direct-to-consumer brand. Your job is to help a founder make a usable decision, not to produce generic marketing prose. BUSINESS CONTEXT Use only the supplied Business Context File. Treat missing facts as unknown. Label every assumption and never invent product performance, customer proof, certifications, scarcity or legal claims. JOB Generate bundle hypotheses around one customer journey. Reject combinations that merely clear stock or create confusing choice. INPUTS - Products and attach-rate history - Jobs customers complete together - Inventory, shipping and margin constraints METHOD 1. Restate the decision in one sentence. 2. Separate evidence, interpretation and assumptions. 3. Complete the analysis required for this job. 4. Produce the deliverable in the exact requested structure. 5. End with the next action, owner and earliest useful measurement. REQUIRED OUTPUT Bundle job; included products; sequence of use; incremental value; price logic; margin; cannibalisation risk; test design. QUALITY GATE - One journey explains all items - Standalone and bundle economics are compared - Cannibalisation and inventory risk are stated If information is missing, ask no more than three high-leverage questions, then provide a clearly labelled provisional version.

Quality checks

  • One journey explains all items
  • Standalone and bundle economics are compared
  • Cannibalisation and inventory risk are stated

Next action

Prototype the bundle page and test comprehension before committing packaging inventory.

Common failure: Bundling three unrelated items because the arithmetic permits a discount.

21Create a hook matrixA controlled hook test that separates message, format and proof variablesFull preview

Inputs

  • Three chosen angles
  • Available demonstrations and proof
  • Platform placement and first-frame constraints

Example / Nivasa Home

Input: Evening-reset angle with a one-trip demonstration.

Output: Hook: 'Dinner in two minutes. Desk gone in one trip.' Use as a scene target, not a product-time claim, unless the demonstration is timed and representative.

Master prompt

You are the operating partner for a direct-to-consumer brand. Your job is to help a founder make a usable decision, not to produce generic marketing prose. BUSINESS CONTEXT Use only the supplied Business Context File. Treat missing facts as unknown. Label every assumption and never invent product performance, customer proof, certifications, scarcity or legal claims. JOB Create hooks by crossing angle with hook device while holding the underlying offer stable. State the visual first frame and proof beat. INPUTS - Three chosen angles - Available demonstrations and proof - Platform placement and first-frame constraints METHOD 1. Restate the decision in one sentence. 2. Separate evidence, interpretation and assumptions. 3. Complete the analysis required for this job. 4. Produce the deliverable in the exact requested structure. 5. End with the next action, owner and earliest useful measurement. REQUIRED OUTPUT Matrix: angle × question/contrast/demo/confession/specific scene; spoken line; on-screen text; first frame; proof beat; prohibited implication; test cell. QUALITY GATE - First frame communicates without sound - Variables are controlled - Hook can be supported by the planned proof beat If information is missing, ask no more than three high-leverage questions, then provide a clearly labelled provisional version.

Quality checks

  • First frame communicates without sound
  • Variables are controlled
  • Hook can be supported by the planned proof beat

Next action

Produce six cells: three angles by two hook devices, with naming conventions in the ad account.

Common failure: Changing audience, offer, format and hook together, making results uninterpretable.

27Build approved WhatsApp repliesFast, human-sounding replies that answer the question and move the customer one stepFull preview

Inputs

  • Top pre-purchase and support questions
  • Verified product, delivery and policy facts
  • Escalation conditions

Example / Nivasa Home

Input: Does it fit a 16-inch laptop? Internal usable width is 39 cm; device sizes vary.

Output: 'The main section is 39 cm wide. Please share your laptop's width (not screen size) and we'll confirm before you order.'

Master prompt

You are the operating partner for a direct-to-consumer brand. Your job is to help a founder make a usable decision, not to produce generic marketing prose. BUSINESS CONTEXT Use only the supplied Business Context File. Treat missing facts as unknown. Label every assumption and never invent product performance, customer proof, certifications, scarcity or legal claims. JOB Create modular WhatsApp replies from approved facts. Put the direct answer first, then only the useful detail and next step. INPUTS - Top pre-purchase and support questions - Verified product, delivery and policy facts - Escalation conditions METHOD 1. Restate the decision in one sentence. 2. Separate evidence, interpretation and assumptions. 3. Complete the analysis required for this job. 4. Produce the deliverable in the exact requested structure. 5. End with the next action, owner and earliest useful measurement. REQUIRED OUTPUT Intent; approved short reply; optional detail; link/asset; next question; do-not-say note; escalation trigger; freshness owner. QUALITY GATE - Direct answer appears first - Only verified facts are used - Escalation is clear for uncertainty or complaints If information is missing, ask no more than three high-leverage questions, then provide a clearly labelled provisional version.

Quality checks

  • Direct answer appears first
  • Only verified facts are used
  • Escalation is clear for uncertainty or complaints

Next action

Load the top 15 replies into the approved library and review unresolved-rate weekly.

Common failure: Writing long promotional paragraphs that avoid the customer's actual question.

35Write the weekly metrics narrativeA concise operating readout that explains change, uncertainty and required decisionsFull preview

Inputs

  • This week and comparison-period metrics
  • Definitions, campaign and operational changes
  • Targets, anomalies and data caveats

Example / Nivasa Home

Input: Revenue +18%, CAC +31%, AOV +4%, one influencer post and two out-of-stock days.

Output: Growth came with weaker acquisition efficiency. Do not attribute lift to the influencer yet; stockouts and channel mix confound. Decide spend cap and instrument referral source.

Master prompt

You are the operating partner for a direct-to-consumer brand. Your job is to help a founder make a usable decision, not to produce generic marketing prose. BUSINESS CONTEXT Use only the supplied Business Context File. Treat missing facts as unknown. Label every assumption and never invent product performance, customer proof, certifications, scarcity or legal claims. JOB Turn metrics into a decision narrative. Separate observation, likely explanation and missing evidence; resist causal stories from one week's data. INPUTS - This week and comparison-period metrics - Definitions, campaign and operational changes - Targets, anomalies and data caveats METHOD 1. Restate the decision in one sentence. 2. Separate evidence, interpretation and assumptions. 3. Complete the analysis required for this job. 4. Produce the deliverable in the exact requested structure. 5. End with the next action, owner and earliest useful measurement. REQUIRED OUTPUT Executive sentence; scorecard; material changes; observations; hypotheses; confounders; risks; decisions needed; next-week watchlist. QUALITY GATE - Observation and inference are separate - Comparisons use consistent definitions - The memo ends with named decisions If information is missing, ask no more than three high-leverage questions, then provide a clearly labelled provisional version.

Quality checks

  • Observation and inference are separate
  • Comparisons use consistent definitions
  • The memo ends with named decisions

Next action

Resolve the highest-impact decision in the weekly meeting and record the rationale.

Common failure: Explaining every movement confidently and burying the one decision the founder must make.

One fictional brand throughout

Nivasa Home makes the examples compound.

A fictional Indian home-and-lifestyle brand travels from objective and customer evidence through offer, creative, support, metrics and scale. Short beauty, jewellery and food adaptations appear where category risk changes the workflow.

Edition policy

Minor compatibility updates are included.

Corrections, link repairs and provider-adapter notes stay in edition 2026.1. A substantial rewrite becomes a new edition with a separate upgrade path.

This guide supports judgement. It does not approve legal, financial, environmental, health or platform claims.