חזרה לדשבורד
    Case Study

    How One Designer Saved Her Client Meeting With a Materials Library She Didn't Know She Needed

    Ben from Boardspace.ai
    November 7, 2025
    7 min read

    Look, I'm not going to pretend this was some revolutionary moment in design history. It wasn't. But it was one of those Tuesday afternoons where everything that could go sideways, did.

    A designer—let's call her J.—had a 2pm staircase presentation inside a Brooklyn brownstone renovation. The client changed their mind daily: oak treads, then painted risers, then brass railings someone spotted on Pinterest.

    By 11:30am she had zero usable options. The only reason this story matters is that she stumbled into the materials library + orchestrator flow in Boardspace and fixed everything in under an hour.

    The Part Where Most Designers Lose Three Hours

    J.'s default scramble mirrored every chaotic pre-meeting sprint. The usual loop looked like this:

    Best case? She'd produce one mockup by 1:45pm. Worst case? Walk in empty-handed.

    • 1.Pinterest scroll for treads (15 minutes)
    • 2.Back to Pinterest for railing photos (10 minutes)
    • 3.Some random platform to recall a paint swatch
    • 4.Manual Photoshop mashup while cross-checking whether those materials actually exist

    What She Did Instead (And Why It Matters)

    Her recovery workflow was boringly efficient—and that was the entire advantage:

    Total time: 90 seconds. Four decisive clicks beat 40 minutes of open-tab chaos.

    • 1.Snapped a quick phone photo of the existing stairs
    • 2.Opened the Inspiration Gallery → filter: “stairs” + “material_research”
    • 3.Selected four micro references: raw oak tread grain, brushed brass baluster, semi-gloss white riser, LED strip concept

    The Part That Would've Eaten Her Afternoon

    She added the staircase photo + four references to chat and pressed “Send.” The orchestrator took over.

    Prompt: “Show me these oak treads and brass rails on my staircase—give me three variations with different layouts.”

    • 1.Auto-detected this as a material placement request
    • 2.Routed to Flux-based architectural models for precision
    • 3.Nano Banana handled finishing passes—no model selector, no drop-down guessing

    What Came Back in 45 Seconds

    Three variations landed in under a minute, each one honoring the original perspective:

    Client thought they were contractor mockups because every material looked intentional, not AI-random.

    • 1.Oak grain oriented three ways across the treads
    • 2.Brass railings rendered as full balusters vs. minimal cable systems
    • 3.White risers with and without LED reveals
    • 4.Original geometry preserved, zero hallucinated décor

    The Client Meeting: 22 Minutes

    She presented three layouts of the client's actual staircase—same perspective, different material placements.

    • 1.Meeting length: 22 minutes
    • 2.Design + prep time: < 60 minutes
    • 3.Immediate approval on option two (brass balusters + LED strips), no extra iterations

    Why This Actually Works (The Technical Bit)

    Under the hood, the orchestrator did the heavy lifting:

    Micro references eliminate ambiguity. “Oak on treads, brass on rails” becomes enforceable, not wishful wording.

    • 1.Five inputs (one base photo + four micro references) were classified as a material placement request
    • 2.Flux architecture blended each reference onto matching geometry using conditional inputs
    • 3.Variations tweaked placement parameters instead of re-rolling the entire scene, delivering cohesive outputs

    The Part That Actually Saves More Time Than the Generation

    Skipping Pinterest loops matters as much as fast renders:

    The same surgical references that guide the AI also give you spec-ready sourcing.

    • 1.Every gallery asset carries manufacturer metadata
    • 2.Approved option already links to real suppliers
    • 3.No post-meeting “let me find out if this exists” follow-up

    Why This Matters For Your Next Project

    Same pattern, different space. Reuse the flow:

    Applies to:

    • 1.Start with one photo of the real space
    • 2.Grab 3–4 micro references (materials, not mood boards)
    • 3.Describe placements in plain language
    • 4.Let the orchestrator route to the correct agents

    What's coming to Boardspace:

    • • Kitchen backsplashes with tile chips
    • • Exterior cladding swaps
    • • Flooring transitions or stair runners
    • • Any detail where the material matters more than a whole-scene vibe

    A Quick Note on Token Costs (Because Everyone Asks)

    Reality check on the budget:

    Smart routing prevents the “try every model” tax.

    • 1.Inspiration Gallery access + selecting references: free
    • 2.Three orchestrated variations: 2 tokens total (~$0.20)
    • 3.Manual guessing with random AI tools usually burns 3+ tokens before you find the right model

    One Last Thing

    This workflow scales beyond staircases:

    Pattern stays the same: real photo + micro references + natural language = production-ready visuals while the orchestrator handles tooling.

    • 1.Kitchen photo + tile samples → instant backsplash concepts
    • 2.Facade photo + cladding swatches → accurate exterior studies
    • 3.Floor plan sketch + materials → cohesive client-ready deck

    Conclusion

    The bottleneck isn't render speed—it's the indecision and tab-juggling before you even choose a tool. Gathering precise materials first lets the orchestrator auto-select the right agent, so you review accurate options in minutes.

    J.'s staircase story is repeatable: one base photo, focused references, clear placement prompt, three client-ready variations under an hour, token spend under a latte.

    J. is a Brooklyn-based designer specializing in historic renovations. She still uses Photoshop—but not for staircase mockups anymore.

    Open the Inspiration Gallery