From a product brief and a corpus of customer reviews, build a complete mini-campaign: positioning one-pager, landing-page copy with A/B variants, a two-week social calendar, a launch email, and a metrics sheet — then run a brand-voice and fact-check pass across every asset, because the reviews and the brief do not tell the same story.
You are the one-person marketing team at Verdant Home Pvt Ltd, a fictional D2C brand launching AquaPure Mini, a countertop water purifier for renters. The founder hands you a brief and a folder of beta-customer reviews, and wants the full campaign kit by evening. This capstone chains together everything the track has taught: drafting at speed, grounding claims in sources, catching contradictions between sources, and running a verification pass across a set of assets where one wrong claim replicates into five. The brief says what the founder believes; the reviews say what customers experienced. Where they disagree, your campaign must side with the evidence — a claim that survives the brief but not the reviews is a refund queue waiting to happen.
AQUAPURE MINI - INTERNAL PRODUCT BRIEF (founder's draft)
Product: countertop RO+UV purifier, no plumbing, no installation.
Fill the 6L tank, purified water in minutes. For renters and
small kitchens.
Price: Rs 8,999 launch price (MRP Rs 11,999). Filters: Rs 1,499
per set, replace every 6 months.
Claims we can make: removes 99.9% of dissolved impurities; zero
installation; tank purifies a full 6L in under 8 minutes; runs
whisper-quiet; 1-year warranty, extendable.
Target: urban renters 25-40, tier-1 cities, currently using
20L delivery cans or boiling.
Positioning idea: "RO-grade water without drilling a single hole."
Launch: email to 4,200-person waitlist + Instagram/LinkedIn organic
over 2 weeks. No paid budget this quarter.
Success: 300 units in 30 days; email CTR above 4%; less than 2%
return rate.
R01 5* "Set up in literally 10 minutes out of the box. As a
tenant this is the product I waited years for."
R02 4* "Water tastes great. Tank takes about 15-20 min for a full
fill-and-purify though, not the 'minutes' the box implies."
R03 5* "No drilling! My landlord is happy, I am happy."
R04 3* "Good water but louder than expected at night in a studio
flat. Like a low fridge hum."
R05 5* "Replaced our 20L cans. Paying for itself already."
R06 4* "Purification is solid per my TDS meter. Full tank takes
closer to 18 minutes in my experience."
R07 2* "Filter light came on at 4 months, not 6. Support says
heavy usage shortens filter life. Wish the box said that."
R08 5* "Perfect for my 1BHK. Fits under the cabinet."
R09 4* "Great product. Slight hum at night but fine in the day."
R10 3* "Third filter month and flow slowed a lot. Cleaning the
pre-filter helped. Manual does not mention this."
R11 5* "Moved flats twice this year - just packed it like a
kettle. Genuinely zero installation."
R12 4* "Tastes better than the can water. Takes a while to purify
a full tank but we just fill it at night."
Mine the reviews before reading the brief twice. Prompt: "Cluster these 12 reviews into themes: what customers love, what they complain about, and any claims about product performance (times, durations, noise, filter life) with the review IDs. Quantify each theme — how many reviews support it?" Verify the counts yourself against the review IDs (Lab 3 discipline — AI counts are drafts). You now hold the evidence base every asset must survive.
Reconcile brief vs reviews — the claims audit. Build a claims table: each brief claim, supporting evidence (review IDs or "brief only"), contradicting evidence, and your ruling — keep, soften, or drop. Work through all five "claims we can make". At least two rulings will not be "keep". For each softened claim, write the honest version the evidence does support ("purifies while you sleep — fill at night" is stronger marketing than a disprovable "under 8 minutes" anyway). This table is the spine of the capstone; every later asset gets checked against it.
Write the positioning one-pager. Prompt with the brief, the review themes, and your claims table: "Draft a positioning one-pager: audience, core insight, positioning statement, three proof-backed benefit pillars (cite review IDs), one honest limitation reframed as guidance, and the launch offer. Use only claims marked KEEP or the softened versions from my table." Edit it: the founder's "no drilling" line is genuinely strong and review-validated (R03, R11) — keep it central. Check no dropped claim leaked back in.
Draft the landing-page copy with true A/B variants. Ask for hero section (headline, subhead, CTA), three benefit blocks, a specs strip, FAQ (five entries — include filter life and purification time, answered honestly per your table), and social proof pulled verbatim from reviews (quote real review text, IDs noted). Then the A/B pair: two hypothesis-driven hero variants — A leads with the renter/no-installation angle, B leads with the replace-your-cans economics angle (R05, R12). A/B means one variable and a reason, not two random headlines; write the hypothesis line above each.
Build the two-week social calendar. Prompt: "14-day organic calendar, Instagram + LinkedIn, for this launch: for each post — day, platform, format, hook line, one-line brief, CTA, and which claims-table entries it uses. Mix: 4 product, 3 social-proof, 2 founder-story, 2 educational (water quality), 2 objection-handling (noise at night, filter life), 1 launch-day. No claim outside the table." Move it into your spreadsheet. Sanity-check the two objection-handling posts especially — marketing that pre-empts the known complaints (noise, filter life) converts the 3-star reality into trust instead of returns.
Write the launch email. To the 4,200-person waitlist: subject line (plus one alternate for testing), 150-word body, one CTA, launch-price urgency with the MRP anchor, and a P.S. with the strongest verbatim review quote. Constraint to give the AI: "every factual claim must match my claims table; state the filter-replacement cost honestly near the offer." Hiding the Rs 1,499/6-month consumable is short-term thinking the 2%-returns target cannot afford.
Build the metrics sheet. Columns: asset, metric, target, source of truth, check-in date. Rows from the brief's success criteria (300 units/30 days, email CTR >4%, returns <2%) plus per-asset metrics you add: A/B winner criterion (CTR difference over N sessions), social engagement per theme, waitlist conversion. Add a "claims watch" row: track filter-life complaints in the first 60 days — your claims audit made a prediction, and the sheet should test it.
The cross-asset verification pass — the capstone's exam. Now audit the full kit as one set. Three sweeps: facts — every number (price, MRP, tank size, times, filter interval, warranty) identical across all five assets and consistent with the claims table; voice — read all assets aloud back to back; flag any sentence that sounds like a different company wrote it, and fix to one register (warm, direct, honest-technical); claims — grep yourself for the dropped/softened claims ("99.9%", "under 8 minutes", "whisper-quiet", "6 months") and confirm no asset carries a banned version. Log every inconsistency found and fixed — finding zero means sweep again, because first-pass multi-asset kits are never internally consistent.
In a multi-asset campaign, one unverified claim doesn't fail once — it replicates into every asset and fails everywhere at launch scale.
A campaign is only as honest as its claims table. Mine the evidence first, rule on every claim before drafting anything, generate the five assets fast against that single source of truth — then verify the set, because in campaigns, errors don't just ship, they multiply.