Ryan's 3 Concrete Asks — How We Address Each + What We Show Thursday

Filed: 2026-05-26 by Mycelia Triggered by: Billy's text exchange with Ryan 2026-05-26 AM. Ryan named three black-and-white focus areas: image selection, photo retouching, development-related activities (internal). Billy committed to leading the Thursday/Friday call with those three + showing the broader extension after. Status: v0.1 plan — Billy reads first; agent network coordinates around it Source-of-truth: Ryan's EF_Tech_Stack_Image_Workflow.pdf (the actual document he sent us) maps directly onto his three asks. Every pain point he raised verbally is in his PDF in detail.

1. What Ryan said vs. what we built

Ryan's text, paraphrased:

"I keep coming back to areas where I'm having most success in black-and-white type things naturally. Repetitive tasks. The ability to create tools to help make discretionary decisions, workflows to help automate binary decisions, focusing on things of high repetition. Net net, I boil down the three main areas to: image selection, photo retouching, and development related activities for internal purposes."

What our deck currently leads with: brand DNA, marketing system, Remy agent, Pinterest-first media plan, 12-month channel plan, lookbook generation.

The honest read: Ryan asked for operational pain relief in his most-time-consuming workflows. We've been building a brand-marketing system. The platform substrate (Remy + agent architecture + brand DNA + GCP + Gemini multimodal) is the right foundation for solving his three asks — but we haven't built the specific surfaces yet. We've been pitching what we wanted to build, not what he asked for.

Billy's text correctly course-corrected: "I'll be sure it leads with the focus you have in mind but I do ask you be open to seeing something a little broader." Ryan agreed ("Sounds amazing"). The Thursday/Friday call needs to lead with his three + show the marketing system as the natural extension once you have an agent doing the operational work well.

2. Ryan's 3 asks mapped to his own PDF

His PDF documents the workflow + pain points in detail. The three asks map cleanly:

Ask #1: IMAGE SELECTION

What he does today (per PDF):

• ~40 model shoot days/year × 2,500 images per day = ~100,000 images per season
• Final selection target: ~2,600 images (12 collections × 200 + 200)
• Selection criteria (per PDF): gaze, styling, character, eye open/closed (First Cut); then specific product success + character detail (Final Cut). Target 10-12 final selects per product.
• 4 days of selection work for every 1 day of shooting — a significant time multiplier; the bottleneck that slows shoot-to-market materially.
• Tagging: ½ to 1 full day per shoot day, manual per-image in Lightroom (product identification, production name).
• Lightroom ratings are per-user — not shared across team members.
• No task queue. No per-user view of what needs action.

Annual cost of this work: ~160 days selection + ~20-40 days tagging = ~180-200 days/year of team time consumed by image selection alone.

Ask #2: PHOTO RETOUCHING

What he does today (per PDF):

• Workflow step 9: Retouching and final adjustments in Photoshop. All subsequent steps move through Dropbox.
• Done after Lightroom exports the selected DNGs.
• Manual, per-image. Quality + consistency depends on the retoucher.
• No batch processing infrastructure visible in the workflow.

Ask #3: DEVELOPMENT-RELATED ACTIVITIES (INTERNAL)

What he does today (per PDF — the "Development Use — Print Mockups" section, called out specifically):

"Internal-only practice: rather than creating a new sample in a new colorway, an existing product image is used as a base and the print is swapped out digitally. All other aspects of the image remain unchanged. Used to reduce the total number of sample requests. Results are mixed at present. Workflow is a direct image replacement — take an existing approved image, swap the print layer only."

This is the highest-leverage "internal dev" use case: a brand-aware print-swap pipeline that actually works. Current attempts produce mixed results. Tapt has Gemini-multimodal + Model Profile system + brand DNA loaded — this is exactly the kind of problem the platform is set up to solve well.

Plus likely also: Shopify theme work, Notion SOP automation, Celigo workflow assists — operational dev that today consumes team or freelancer time.

3. How Tapt addresses each — concrete capabilities + the math

Image Selection: from 4-days-per-1-shoot-day → 1 hour + collaborative review

What we build:

• AI-assisted first cut. Gemini multimodal scores each image on Ryan's stated criteria (gaze, styling, character, eye open/closed). Reduces 2,500 → ~250-400 candidates in under an hour per shoot day.
• Auto-tagging at delivery. LLM identifies model + product + setting + color + composition for every image automatically. Replaces ½ to 1 day of manual Lightroom tagging per shoot day with a near-zero-touch step.
• Brand-aware final cut. Score the working set on "specific product success" + "character detail" using EF's actual past final-selects as training reference (we already have access to lookbook + DTC asset archives). Surface top 10-12 per product for human approval.
• Shared web review interface. Team-distributed review with shared ratings + comments. Replaces Lightroom's per-user-ratings limitation.
• Task queue per team member. Per-user "what needs your review" view. Replaces the no-task-queue gap.
• Cross-product image discovery. Tag-based search across all folders. Eliminates the Lightroom-tag-query dependency.

The math (for Ryan to spreadsheet):

• Current: ~180-200 team days/year on selection + tagging
• With Tapt: ~30-50 team days/year on review-and-approve (the human-judgment layer stays human)
• Savings: ~130-150 days/year of team time recovered.
• At fashion-team labor rate of $300-500/day: $40K-75K/year in recovered ops cost OR 70-100 days of team capacity redirected to higher-value work (more shoots, better merchandising, brand work).
• Plus: shoot-to-market timeline compresses by 1-2 weeks per release = more launches per year × more revenue per launch.

Phase 1 deliverable (Day 30): First production-grade selection pass on an actual EF shoot. Demo the workflow on a real shoot batch during the call.

Demo for Thursday: show the platform analyzing a sample EF shoot batch (we use existing public EF photography). Display: auto-tags, first-cut scores, shared review UI, task-queue view. Concrete output Ryan can touch.

Photo Retouching: brand-aware batch + the print-swap pipeline

What we build:

• Brand-aware retouching presets. Trained on EF's editorial register — natural light only, soft warm tones, painterly composition, garment-foregrounded. Apply consistently across a shoot's selects.
• Common retouching tasks automated: skin smoothing, color correction, background cleanup, fabric drape adjustment. Where AI is good enough; falls back to human-Photoshop where it isn't (we don't pretend to replace the artist's eye for the gold-grade work).
• Print-swap pipeline (the "Development Use" pain). Purpose-built tool for "swap print on existing image, keep everything else": brand-aware texture matching, lighting preservation, fabric drape continuity. Current state per PDF: "Results are mixed." Tapt target: production-grade reliable. This is the highest-leverage internal capability we can ship for EF.
• Batch processing queue. Apply consistent treatment to a working set; route exceptions to human review.

The math:

• Current retouching is manual; rough estimate 15-30 min/image × 2,600 final images = 650-1,300 hours/year of post-production work.
• With Tapt's auto-treatment on ~70% of routine images: 400-900 hours/year recovered, retoucher focus shifts to gold-grade hero shots only.
• Print-swap: currently produces a few attempts/season at low success rate. Tapt-reliable print-swap = ~5-10x more print variations testable without sample requests. Direct sourcing cost savings.

Phase 1 deliverable (Day 30-45): retouching pipeline live on EF's actual shoot output. Print-swap tool tested on existing products.

Demo for Thursday: take an existing EF photo → show automated retouching variations (skin, light, palette) + a clean print-swap example side-by-side with the original. Concrete enough for Ryan to evaluate quality before signing.

Development Activities (Internal): the print-swap workflow + Shopify-Liquid + automation

What we build:

(a) Print-swap pipeline — overlaps with retouching. See above. The single most-asked-for internal dev capability per the PDF.

(b) Shopify-Liquid theme authoring via Remy. Was in our Y2+ capability roadmap; bring forward to Phase 1 if Ryan wants it. Remy writes theme code, you review + merge. Replaces freelance-developer cycles for routine site changes.

(c) Internal automation scripting. Celigo workflow assists (the existing integration middleware), Notion SOP generation, NetSuite ↔ Shopify data audits. Any high-repetition operational task gets a Remy-driven automation.

(d) Surgical website adds. The 5 page-additions we proposed (founder landing, artisan-country pages, Love Tòmas origin, sizing-honesty, customer celebration) executable by the platform with EF developer review.

(e) Cross-product image organization. Web UI that fixes the "products shot together can only live in one folder" pain — proper tag-based + permission-aware asset library across the team.

The math:

• Current internal dev consumes either team time or freelancer hours
• With Tapt-as-internal-dev-partner: most routine work goes to the agent + human-review; team time recovered for higher-value work
• Print-swap alone (per ask #2 retouching math) = significant sample-cost savings

Phase 1 deliverable (Day 30): print-swap pipeline live + first Shopify-Liquid theme change demonstrated.

Demo for Thursday: show the platform generating a Shopify-Liquid snippet for a real EF page change request + the print-swap demonstration from above.

4. The broader extension — why "scale of what we are bringing him"

Once Ryan has an agent doing his operational work well, the same agent already knows:

• His full brand DNA (voice rules, painter discipline, the three voice tiers, the canonical 12 blog posts loaded as RAG)
• His full product catalog (every product, color, supply chain origin, lookbook history)
• His full image archive (auto-tagged, brand-scored, searchable)
• His operational rhythms (shoot calendar, drop cycle, no-sale discipline)

Which means the same agent can also:

• Generate founder-voice emails in three tiers (Klaviyo-shippable, ready-to-edit)
• Produce paid social creative + Pinterest pins in EF's editorial register
• Draft PR pitches with EF-specific story hooks (artisan partnerships, founder origin, supply-chain depth)
• Write blog posts in any of EF's voice eras with appropriate restraint
• Maintain a 12-month marketing cadence in EF's natural rhythm
• Surface buried-gold moments (the Father's Day Ryan Shirt cross-merch that's been sitting on the shelf, the Spring Revival storytelling layer that exists but isn't told, the artisan-partnership content franchise across 5 countries)

The deck's competitive-positioning slide (Lumen's Slide 3b) shows where EF lives — the premium-but-not-luxury band, paying for longevity + proof-of-claims + curatorial taste. The marketing system protects + amplifies that position; the operational work is what gives the system enough context to do it well.

Pricing framing: the operational work justifies the engagement on Ryan's CFO math alone (the time/cost savings clear the spend). The marketing extension is upside that compounds over the 24-month bounded engagement.

5. Deck restructure — leading with Ryan's 3

Current deck structure (v0.26):

1. Cover
2. 17 years of brand
3. Painter at the center / two traditions 3b. Competitive positioning (Lumen)
4. "Every dollar spent should show its return"
5. Platform that thinks like your studio 6-16. Platform + plan + lookbook + model profiles + Y2+ + pricing + financials 17-20. Cost / scenarios / relationship / close

Proposed restructure (v0.27 — lead with Ryan's 3):

1. Cover (unchanged)
2. "What you asked for" — Ryan's three: image selection / photo retouching / dev activities. Acknowledged directly + framed as where we'll lead.
3. IMAGE SELECTION — the math (180-200 days/year recovered) + how it works + the demo
4. PHOTO RETOUCHING — brand-aware presets + the print-swap pipeline + batch processing
5. DEVELOPMENT (INTERNAL) — print-swap as the headline + Shopify-Liquid + automation surface
6. The agent that does the above ALSO does this — bridge slide to the broader work
7. Brand position (current Slide 3 / 3b — competitive tier table, three-thread thesis)
8. "Every dollar spent should show its return" (current Slide 4 — bridge from ops-math to broader math)
9. Platform that thinks like your studio (current Slide 5) 10-13. Platform live + station tour + Love Tòmas artisan-craft + supply chain (current Slides 6-8) 14-15. Voice + customer love (current 9-10)
10. Twelve months of compositions (current Slide 11)
11. Summer lookbook (current Slide 12 — proof the brand work is at parity with the ops work) 18-19. Channel plan + Pinterest-first (current 13a-13b)
12. Audience growth (current 14)
13. Model Profiles (current 15)
14. Y2+ benefit-first (current 16) 23-24. Cost shape + financials (current 17-18)
15. Twenty-four months → independence (current 19)
16. Close (current 20)

Net: 6 new slides up front (Ryan's 3 + the math + bridge); 20 existing slides re-sequenced. Doable Wed in tight Lumen + Mycelia coordination.

6. Demo restructure for Thursday call (45-60 min)

Phase 1 (15-20 min) — Ryan's three:

1. Image selection demo. Live on a sample shoot batch. Show: auto-tagging at delivery, first-cut scoring, brand-aware final cut, shared review UI, task queue.
2. Photo retouching demo. Take an existing EF photo. Show: brand-aware retouching presets applied + the clean print-swap (Marigold → Cerulean, e.g.) with lighting + drape preserved.
3. Development demo. Show: Remy writing a Shopify-Liquid snippet for a real-page change request. Plus: the print-swap revisited as the highest-impact internal dev capability.

Phase 2 (15-20 min) — the agent that does this also: 4. Walk through ef.taptstudio.com's tour stations — brand understanding, 12-month plan, gallery, model profiles. Connect: "the agent built the foundation for the ops work; same foundation runs your marketing." 5. The buried-gold moments (Love Tòmas artisan-craft depth, Father's Day Ryan Shirt cross-merch, Spring Revival storytelling layer). 6. The competitive-positioning slide + the three-thread premium thesis.

Phase 3 (10-15 min) — the engagement shape: 7. Bounded 24-month engagement, declining monthly, lift share, Y3-zero. 8. The financial scenarios (every scenario nets positive). 9. The Day 30 deliverable — Phase 1 ops capabilities live + first campaign shipped. 10. Q&A + decision.

Phase 4 (5 min) — close: 11. Three paths: sign / explore the platform / talk through specifics later.

7. Pre-Thursday work needed (rush integration)

Faber — capability scoping (urgent)

Quick assessment of what's buildable for the demo by Thursday evening:

• Image-selection demo: can we build a stub UI that runs Gemini scoring against a small batch of EF images + displays results? ~6-8 hours work?
• Retouching demo: can Gemini multimodal do meaningful retouching variants on a single image? Test with a few real examples + see if the output is demo-quality.
• Print-swap demo: the existing Model Profile system uses reference-based generation. Can we adapt to "keep this image's structure, swap only the print"? This is the trickiest of the three. If not buildable by Thu, we mock with before/after still images + commit to Phase 1 build.
• Shopify-Liquid demo: can Remy already generate a meaningful theme snippet? Probably yes with brand-DNA-aware prompting.

Faber call: what's shippable as live demo vs. what's mocked + Phase-1-committed. Pre-Thu time budget.

Mycelia — deck restructure

• Draft Slides 2-6 (Ryan's three asks + math + bridge). Lead-with-his-three storytelling.
• Re-sequence existing slides.
• Update speaker notes throughout to integrate the lead-with-ops framing.
• Refresh Slide 4 to bridge from ops-math to broader math.
• Update Slide 5 framing ("Platform that thinks like your studio") to include "and handles the operational work you asked us to help with."

Lumen — content review + opposition test

• Pressure-test the math claims (180-200 days/year recovery; ~$40-75K ops cost recovered). Is it defensible? Where is it weak?
• Audit the broader extension framing — does the bridge work or does it feel like upsell?
• Final deck read-through end-to-end with new sequencing.

Mercator — pricing-frame check

• Does the engagement structure still hold with this reframing? The pricing was built around marketing-system value; now anchored in ops-recovery math + marketing as extension. Adjust the talk-track for Slide 17-18 if needed.
• Confirm the math frames cleanly: ops-cost-recovered (CFO-compelling) + brand-premium-protected (strategic) + lift-share (upside).

Praetor — no immediate action

• The contract package stays as-shaped. The broader engagement covers all of this work regardless of the slide lead-in. Note: if Phase 1 explicitly delivers operational capabilities by Day 30, the SOW Schedule A needs that scope language.

8. What changes for Ryan when we do this right

He came to the call expecting a focused conversation on three operational pain points. He'll leave with:

1. A concrete, demo-able solution to his three asks — with math he can put in a spreadsheet.
2. A clear Phase 1 commitment — what's live within 30 days post-sign for his three areas.
3. The realization that the same agent doing his ops work IS the marketing system — not as a pitch but as a natural property of the architecture.
4. A bounded engagement shape that makes the financial math obviously work — the ops savings alone cover the engagement; the marketing extension is upside.
5. A platform already built — not a deck of recommendations.

The line Billy promised: "concrete strategy, brand, marketing and content with outputs you can use and test with right away, Variable but concrete." This plan delivers on that exactly.

9. Open questions for Billy

1. Reveal date — locked Thursday or shifted to Friday per Ryan's "Thursday evening or Friday any time"? Affects rush-integration pace.
2. Demo scope ambition — full live demo of all three (high lift; risky if any breaks) vs. live demo of one + mocked-with-Phase-1-commitment for two (lower risk; equally compelling if framed honestly)?
3. Pricing communication change — does the operational-math-justifies-engagement reframe affect the pricing structure conversation, or does the structure stay as-is + only the talk-track changes?
4. Any prior conversations with Ryan you remember that map to specific scoring criteria, print-swap test cases, or Shopify pain points we should anchor the demo on?
5. Approve to whisper-coordinate Faber + Lumen + Mercator immediately on rush-integration for Thursday?

— Mycelia, 2026-05-26 ~04:30 ET