Real shops. Real numbers.

Before CloseFast

Emma ran three Etsy stores selling print-on-demand wall art and apparel. By her own measurement she spent 4 to 5 hours a day researching competitors, drafting titles and tags, generating mockups in Photoshop, and replying to reviews. Revenue was steady at ~$1,500/mo combined across 3 shops but she was burning out — she'd taken on listings as a part-time job and it had crept into a full-time one without the income to match.

What changed

She subscribed to Pro on April 8, 2026. Onboarding took 14 minutes (Etsy + Printify connections). Within the first week she:

• Generated 87 new listings with the AI Listing Generator (~15 seconds per draft, edited an average of 18 seconds before publishing)
• Activated the Repricing Agent on all 3 stores with floor at COGS+60%, ceiling at niche 90th percentile
• Turned on the Review Responder for auto-drafted replies (she still approves each one before sending)

The numbers, week by week

• Week 1: Revenue $420 — slow ramp, mostly the new listings still in Etsy's "fresh" boost window
• Week 2: Revenue $810 — repricer caught a $4 underpricing on her bestseller
• Week 3–4: Revenue $1,720 / $2,180 — competitor sweep flagged a tag opportunity she immediately incorporated
• Week 5–6: Revenue $3,940 → +162% over baseline

"I was spending 4–5 hours/day on listings and competitor research. CloseFast's Inventory Engine writes SEO titles in 15 seconds and my Etsy revenue jumped 162% in the first 6 weeks."

What she still does manually

• Reviews each AI-drafted listing before publish (avg 18s edit time)
• Approves repricing decisions in the Decisions queue
• Final sign-off on review replies
• Mockup style choices in Printify

Total time spent on the operation: ~12 minutes/day, down from 4 hours. The agents do the work; she does the judgment.

Before CloseFast

Daniel ran a multi-platform POD business — same products listed on Etsy, Printify, and Shopify. He used a competing AI tool that bundled "unlimited AI" into a $99/month plan but in practice throttled at 50 listings/month. He suspected he was paying 10–20× the raw cost of the AI tokens but couldn't prove it because the vendor never published unit pricing.

What changed

Two things, in order:

• Switched to Starter ($19/mo) — same listing capability, paid $80 less the first month
• Added his own Anthropic API key (BYOK) at /keys — total AI bill became Starter $19 + Anthropic ~$8 = $27 vs $99 before

The arbitrage moment

Within the first 48 hours of running the Arbitrage Scout agent, the dashboard surfaced 16 SKUs with price spreads ≥ $5 between Etsy and Shopify. The biggest spread was $14 on a single t-shirt SKU (Etsy $19, Shopify $33). Daniel raised the Etsy price to $26 — split the difference — and saw conversion drop only marginally while revenue per sale climbed 37%.

"BYOK was the seller. I bring my own Claude key — pay ~$8/mo in tokens instead of $99 to other tools with worse output. The arbitrage agent caught a $14 spread between Etsy and Shopify on day 2."

3-month cost summary

• Old vendor (3 months): $297
• CloseFast Starter (3 months): $57
• Anthropic API usage (3 months): $24
• Total: $81 vs $297 → saved $216 + better feature set

The honest tradeoffs

• BYOK billing is variable — heavy listing months cost more (max month: $14)
• Anthropic separately requires a $5 minimum top-up
• If Anthropic has an outage, the listing generator falls back to Groq automatically (free chain)

Before CloseFast

Sarah is a working mother running an Etsy wall art shop solo. She had ~80 listings making about $2,400/mo but couldn't grow because she had no time — listings stayed static for months because she'd come home from her day job too tired to optimize them. She wanted automation she could trust, not "insights" she'd have to act on later.

The 12-minute setup

Sarah signed up at 9:42 PM on a Tuesday. By 9:54 PM she had:

• Connected her Etsy shop via the Sentinel page
• Run the Inventory Pulse agent — got 14 specific findings about underperforming listings
• Activated the Pricing Advisor agent with auto-approve on safe actions only (raise price, lower price, never delist)
• Set the Repricing Agent's cooldown to 24 hours (so prices don't ping-pong)

The first overnight

Sarah went to bed at 10:30 PM. By the time she woke up at 6:15 AM:

• 128 tags had been re-scored against current niche competitors
• The Pricing Advisor had bumped 9 listings up by $1–$3 based on competitor 50th-percentile gaps
• 7 orders had come in for re-tagged listings — including 2 for a print that hadn't sold in 3 weeks

"Autopilot is a game-changer. I sleep, agents post, repricer adjusts. Woke up to 7 new orders my pricing-advisor agent re-tagged overnight. Setup took 12 minutes. Site Audit alone was worth the subscription."

Two months in

• Revenue: $2,400/mo → $4,100/mo (+71%)
• Listings actively optimized per week: 0 → ~20 (the agents do it, she reviews on Sunday)
• Time spent: 20 min/day reviewing the Decisions queue, approving or rejecting

What still requires her attention

• Adding new product designs (creative work — agents don't replace this)
• Customer service for non-standard questions (agents handle the routine)
• Sunday review of the week's decisions (~30 min)

The platform doesn't replace the seller. It removes the busywork between the seller and the customer.