Brief
Alila Fort Coorg is Direct Create’s flagship Culture Stack engagement of 2026. The work is for a luxury hospitality property in Karnataka’s Coorg (Kodagu) region operated under the Alila brand and held by Seabird Resorts, with the architectural design team led by Dar and Wagh. DC’s MULA Two Directions Culture Stack reads the property’s two directions of intent (rooted-cultural-landscape and Western-Ghats-naturalist) as a single coherent positioning, anchored in the Kodava cultural substrate, the Western Ghats forest ecology, and a coffee-and-honey agricultural thread. The engagement carries the largest active archive in the wiki and serves as the canonical exemplar referenced by five concept pages.
Scope and deliverables
DC’s scope is research, cultural strategy, and an art program advisory across the property. The deliverables run in two phases. Phase 1 (cultural strategy) covers the Culture Stack research document, art program proposal, site visit synthesis, and the Building Culture Stack scaffolding that translates research into design recommendations. Phase 2 (art program execution) covers curatorial framework, artist and artisan commissioning, and installation oversight across public areas, guest suites, landscape, and signage and wayfinding zones.
The scope ends at advisory and curatorial direction. The Dar and Wagh design team holds architectural and interior implementation responsibility; the Alila operations team holds guest experience and programming.
Timeline
2026-01-21. Alila Coorg Cultural Strategy Proposal issued; the engagement’s opening artifact. 2026-02-10. ALILA COORG Art Program Proposal issued. 2026-02-26. Art Program Proposal Update 1: revised scope and structure based on client feedback. 2026-03-05. Site visit to the Coorg property; field notes captured in the Alila Coorg Site Visit document. 2026-03-09. ALILA COORG Building Culture Stack scaffolding document published, translating field observations into the working culture stack. 2026-03-27. Dar and Wagh design team circulated four reference PDFs covering Public Areas, Guest Suite Schematic, Landscape Approach, and Signage and Wayfinding, anchoring the architectural side of the conversation. 2026-04-18. Action Tracker spreadsheet established as the operational backbone, alongside the external_note.md (client-facing narrative) and internal_note.md (internal team direction). The action tracker is the current state-of-engagement reference.
Cultural framing
The MULA Two Directions Culture Stack reads Coorg through two parallel lenses that the property can hold simultaneously. The rooted-cultural-landscape lens carries the Kodava community’s specific cultural memory: ancestor worship, the okka (clan) structure, the ainmane house typology, festival ecology around the Cauvery, and the dignified domestic interior tradition. The Western-Ghats-naturalist lens carries the biogeography of the property’s setting: shola forest, coffee shade canopy, wildflower ecology, monsoon hydrology, and the elephant corridor that runs adjacent to the site. The two lenses converge on a coffee-and-honey agricultural thread: smallholder coffee estates, traditional beekeeping, and the seasonal rhythms that pattern Kodava domestic life.
The MULA naming carries a specific intent. MULA reads as “root” in several Indian languages, and the Two Directions framing keeps the project from collapsing into a single visual or thematic mode. The brief recently barred a roots-and-underground metaphor for the deck and prompts as a deliberate hedge against the most obvious reading of MULA; the cultural argument is rooted, but the visual language is not.
Methodology in canonical form
This engagement is the cleanest demonstration of the Culture Stack methodology in active practice. The work moves through six segments in sequence: Nature Substrate (the Western Ghats geography and forest ecology), Living Culture (Kodava community life and ritual), Agricultural Thread (coffee, honey, monsoon agriculture), Guest Thesis (who the resort welcomes and the experience offered), Differentiation Strategy (positioning against other luxury hospitality in South India), and Design Language (translation into material, palette, and spatial recommendation). The 484-image archive classified across the 44-node IIP taxonomy is the visual substrate that grounds each segment.
The engagement also exemplifies the parallel team structure: DC holds cultural authority, Dar and Wagh hold architectural authority, and Alila operations hold experience authority. None reports to the others; the three braid together at deliverable handoffs.
Reading order
- Alila_Coorg_Cultural_Strategy_Proposal.docx: the January 21 opening cultural strategy proposal.
- 260226 Update1 ALILA COORG-ART PROGRAM PROPOSAL: the February 26 art program proposal update, the most developed version.
- Alila Coorg Site Visit: the March 5 site visit field notes.
- ALILA COORG building CULTURE STACK 09032026: the March 9 culture stack scaffolding document.
- Alila_Coorg_Culture_Stack_Action_Tracker: the April 18 action tracker.
Image catalog
The Alila Fort Coorg field photography archive was cataloged through the IIP pipeline on 19 April 2026, vision backend Gemini 2.5 Flash, source directory 01-alila-coorg/Image_Master/00_Field_Photography. The catalog sits at alila_fort_coorg/iip_output/catalog.json alongside this page and is referenced in the frontmatter as iip_taxonomy. 491 images carry classifications across filename, primary subject, scene type, mood, dominant colors, quality score, cultural markers, and a per-image key observation. The Gemini-generated schema is flatter than the Claude Sonnet schema used at Niyaara and Dadar, carrying scene type in place of image type and omitting orientation, lighting, secondary elements, region, and named-entity fields. A re-pass under the richer schema is a deferred item; the current catalog is the working artifact that the Culture Stack Automator pipeline reads against.
The catalog distributes across twelve scene types, dominated by close-range documentation. 175 images are craft details, 137 are architectural interiors, and 84 are architectural exteriors, together carrying the bulk of the heritage and material evidence. 41 material textures, 20 flora and fauna frames, and 13 ritual or ceremonial captures round out the heavier categories. 10 documentary records, 7 landscapes, 6 agricultural frames, 5 portraits, 3 waterscapes, and 1 abstract close the surface. Mood reads serene and contemplative across the heritage frames and documentary across the working records. 71 images carry a quality score of 9, anchoring the strongest visual material. Two patterns stand out for the cultural framing the page sets up. First, the archive’s named-marker signal skews toward temple and heritage architecture: traditional Indian wood carving, traditional Indian architecture, carved pillar capitals, and floral motifs dominate the cultural_markers field, with forest references on ten frames and temple references on eight. Second, the Kodava-specific markers the deck thesis depends on (ainmane, okka, coffee shade, honey, shola) read sparsely in this pass, with one coffee mention and zero ainmane or shola tags. The catalog therefore documents the heritage and forest substrate strongly while the living Kodava material remains under-cataloged, a gap the next field pass or a richer-schema re-pass should close.
Six images carry the highest quality scores and anchor any visual presentation of the project.
IMG20260310130603.jpg, architectural_exterior, quality 9: an expansive panoramic view of a heritage temple complex with a long tiled-roof building, a central stone Nandi mandapa, and manicured grounds. The reference image for the rooted-cultural-landscape direction the MULA framing opens with.IMG20260310130140.jpg, architectural_interior, quality 9: a long traditional corridor with a richly painted geometric and floral ceiling in gold, red, and brown, a single human figure providing scale against the ornate vault. The reference image for the dignified domestic interior tradition the Kodava chapter calls out.IMG20260310123410.jpg, craft_detail, quality 9: a deeply carved semicircular stone element with a layered scalloped fan pattern, likely a balustrade base, photographed at extreme close range. The reference image for the heritage stonecraft thread the design language chapter builds on.IMG20260311095727.jpg, ritual_ceremony, quality 9: a small dolmen-form stone shrine of rough-hewn slabs with smaller stacked rocks, framing a dark interior opening with traces of offerings. The reference image for the ancestor-worship and folk-ritual layer the Living Culture segment carries.IMG20260312101849.jpg, flora_fauna, quality 9: a macro frame of a branch carrying dried light-brown papery seed pods against an out-of-focus green canopy in bright daylight. The reference image for the Western Ghats forest substrate the naturalist direction reads against.IMG20260310152525.jpg, material_texture, quality 9: an extreme close-up of a carved granite base showing repetitive petal-form motifs in relief, with a plain reddish-brown course above. The reference image for the material-vocabulary brief the design recommendations chapter feeds.
The image files themselves live in the project’s Dropbox archive named in image_archive above. The four images at the top of this section are rendered inline against the absolute Dropbox path; the catalog stays relative per Schema Revision 8 so a future archive-root move is a one-line change in config/archive_root.json.
This Image catalog section follows the binding pattern recorded in operations/quality_benchmarks.md, extended with inline rendering per Schema Revision 8. The page body’s reference to a 44-node IIP taxonomy describes the conceptual node schema the Culture Stack Automator pipeline operates against; the present catalog was generated under the flatter Gemini scene_type schema and will be reconciled with the 44-node taxonomy in a subsequent IIP pass.
Automation postmortem
The Culture Stack Automator (CSA) pipeline (CD orchestrator, CSR researcher, IIP image intelligence, CW writer, GD graphic designer) was run end to end on this project across March and April 2026. The pipeline output, archived under MULA_Two_Directions_Build/, was not publishable. The research tools (CSR cultural research, IIP image catalog) produced usable input that fed the human-built deliverable; the visual production tools (GD slide rendering with Imagen, the slide compositor) did not.
What worked. CSR produced the Kodava substrate research, the Western Ghats biogeography reading, and the coffee-and-honey agricultural thread that anchor the cultural argument. IIP catalogued 491 field photographs across primary subject, scene type, mood, dominant colors, quality score, and cultural markers, and surfaced the 71 quality-9 frames that anchor the visual evidence. Both research outputs are referenced in the cultural framing section above and continue to feed the engagement.
What failed. GD with Imagen produced generic stock-grade imagery rather than the place-specific compositions the Culture Stack methodology requires. The PIL slide compositor carried structural defects that no amount of prompt iteration corrected: text was baked into PNG backgrounds at the rendering step (rather than placed as editable Keynote or PowerPoint layers), and blank slides were shipped as complete because the pipeline lacked a content-presence check before flagging a file done. The CW slide blueprints were coherent at the outline layer; the GD rendering step did not translate them into the visual register the brief required.
What was actually delivered. The client deliverable was built manually by Rajeev and Dway in Keynote across the same March to April window, drawing on the CSR research and the IIP catalog as input. The delivered deck is Presentations/MULA_Two_Directions_Art_Program.key, not any .pptx in MULA_Two_Directions_Build/. Any future reference to the Alila Coorg deck should point at the Keynote file as the canonical client artifact.
Cost note. The CSA run drew substantial Gemini credit across the CSR research expansion, the IIP classification pass, the CW blueprint drafting, and the GD image generation rounds. The spend produced two reusable artifacts (the CSR research, the IIP catalog) and one unpublishable visual production. The credit-versus-outcome ratio is recorded here so future engagement planning treats the GD stage as a research expense, not a production stage, until the visual production tooling is rebuilt.
Implications for the methodology. The split between research tools (worked) and visual production tools (failed) is the live constraint on CSA reuse. Subsequent engagements can lean on CSR and IIP without expecting GD to produce publishable slides; the manual Keynote build remains the production path until the visual production tools are rebuilt or replaced. The build lessons section in wiki/synthesis/dc_ke_build_lessons.md carries the wider framing.