Ep 56: AI & Product Procurement with Daniel House Club Founders
AI & Product Procurement with Daniel House Club Founders
Two brothers who built a design firm, hated the back end of it, and turned that problem into a membership-based procurement platform for designers everywhere — now powered by AI from vendor data extraction to image formatting.
- Daniel House Club started as an internal tool to solve what Peter and Alexander hated about running a design firm. That same problem — messy procurement, unpredictable margins, fragmented vendor relationships — is shared by designers at every experience level.
- AI cut product request time from 45 minutes to 5 — with a human still in the loop for quality control. The goal is not to eliminate the team; it is to reallocate them toward work that actually requires judgment.
- Vendor data is messy and AI is getting better at cleaning it up. From PDF price lists with images at the top and specs buried in footnotes, to full website scraping for catalog data — the technology is narrowing the gap between what vendors provide and what designers can actually use.
- AI brand voice training requires persistence and specificity. "Don't use luxury or elegant" is not enough. You need a documented style guide, a list of overused AI phrases to avoid, and a custom GPT that references both every time it writes.
- The future of sourcing is a Google Lens that only searches trade vendors. That feature does not fully exist yet — but the pieces are assembling and designers should know what to ask for when it does.
Peter and Alexander Spalding are brothers who ran a residential design firm for several years before building Daniel House Club — a membership-based procurement and back-of-house platform for interior designers. What started as an internal tool to solve their own sourcing pain points has grown into a full logistics platform serving designers nationally, with a concierge team, curated vendor catalog, and an increasingly AI-powered product request and catalog management system.
The Origin Story: Building the Tool You Wish Existed
Peter and Alexander started as a design firm in 2015. Peter describes loving the design work. Alexander describes hating nearly every day of it — specifically because the back end of the business was painful. Procurement was unpredictable. Maintaining trade minimums with enough vendors to keep projects feeling unique was exhausting. And the most profitable part of the business — furniture procurement — was also the most administratively demanding.
Alexander started building Daniel House Club as an internal tool: a way for Peter to go to a job site and quickly know what products they had access to, what margin they could count on, and what would actually fit the client's direction. It worked. And when they started showing up at IDS events and meeting designers who had been practicing for 30 years with the same exact problems, they realized the platform had a much larger audience than just their own firm.
"We want to partner with designers and do all the stuff you think is boring and sucks and you hate — and we'll happily do that for you. Because the thing Peter's bad at is what we built this company around."
— Alexander SpaldingThe platform today functions as a full back-of-house operation: a curated vendor catalog, a concierge team that handles order tracking and vendor follow-up, a product request system, and an increasingly AI-powered data and image pipeline. For designers who do not want to hire a procurement coordinator but need the function covered, Daniel House Club fills that role at a fraction of the staffing cost.
AI in the Catalog Pipeline: From Messy PDFs to Usable Data
The biggest operational challenge for any trade platform is vendor data. Vendors provide it in every format imaginable — some send beautifully structured spreadsheets, others send PDFs with the product image at the top, the price buried somewhere in the middle, and the specs in a footnote. Getting all of that into a consistent, searchable format used to require significant manual labor. AI is changing the ratio.
The tool Daniel House Club uses (built by a company on their advisory board) can ingest vendor PDFs and extract structured data — product names, prices, dimensions, materials — even when that data is scattered throughout an inconsistently formatted document. It can also be trained on a vendor's website to pull catalog data automatically without anyone on the team having to manually copy it.
Alexander is careful to note: the speed improvement did not reduce headcount. The catalog team is still one of their largest teams — because there is more work than ever. What changed is how much of their time goes to copying and pasting data versus reviewing, correcting, and improving it. That shift matters for both efficiency and job satisfaction.
Consistent Product Photography — Without Paying for Reshoots
Vendor product images are as inconsistent as vendor data. One vendor shoots their chandeliers at a 5×7 aspect ratio with a warm grey background. Another ships 16×9 lifestyle images. A third sends 1×1 sillos with the product offset to the left. When these all end up on the same catalog page — or in a designer's project board — the result looks like a mess.
Daniel House Club is rolling out AI-powered image normalization across new product additions. The process: take any incoming product image, identify the main product, center it in a square frame, set the background to white, and add a 15% gradient drop shadow. Every product on the platform will eventually display with that consistent treatment as the primary thumbnail, while the original vendor images are preserved and accessible on the individual product page.
"There are some products we just don't market because the images are so weird. Now we can showcase more of the catalog — because the images look consistent and professional."
— Alexander SpaldingThe immediate benefit for designers: product images pulled from Daniel House Club for use in design boards and client presentations will have consistent aspect ratios and clean backgrounds, which means less time in Canva or Photoshop cleaning things up before they look presentable. This is exactly the kind of behind-the-scenes infrastructure investment that most vendors are not making — and that makes a real difference in the designer's daily workflow.
Alexander's Personal AI Workflow
Alexander is the tech-oriented half of the partnership. His day-to-day AI use spans writing, research, data analysis, and meeting notes — and he has specific opinions about which tools do which things well.
On brand voice: Alexander notes that getting any LLM to stop writing like an AI is an ongoing process. The word "luxury" keeps appearing no matter how many times he has removed it from the training. His workaround is a dedicated copywriter custom GPT trained on a brand guide — that GPT gets summoned into other conversations using the @ function in ChatGPT rather than rebuilding context every time.
Jenna's tip from the conversation: Build a running list of AI overused words and phrases — she has one that is 150 items long — and paste it into every new GPT you train. Telling the model "use these sparingly" dramatically reduces the robotic filler language and flowery filler phrases that leak into AI-generated copy.
Deep Research and What the Future of Sourcing Looks Like
One of the more technically interesting parts of this conversation is the discussion of deep research — the ability to point an AI at your own cloud storage, email, calendar, and documents and ask it to synthesize across all of it simultaneously. In Gemini and ChatGPT, this requires connecting those accounts through the tool's settings. Once connected, a deep research prompt can reach into your Google Drive, search past emails, cross-reference calendar notes, and surface information you forgot you had.
Alexander used it to run PhD-level statistical analysis on cohorts of Daniel House Club members — examining behavior patterns at specific points in time, segmented by customer type. Work that would cost hundreds of thousands of dollars with a data analyst, done in minutes. The limitation is data volume: massive spreadsheets can time out. The workaround is starting with a subset of the data and scaling up as you get a feel for what the tool can handle.
Looking ahead, the feature both Jenna and Alexander are most excited about: image-based product search that works exclusively for trade vendors. The equivalent of Google Lens, but instead of returning Wayfair and Amazon results, it searches only the catalogs of actual trade suppliers. The pieces are being assembled — Daniel House Club is actively working toward real-time inventory checks across their vendor catalog, and the reverse image search component is in development at the companies they are watching.
"Google Lens already exists for consumers — but the results are junk. Amazon, Wayfair, mass retail. We need that same technology pointed specifically at trade vendors. That is the future of sourcing."
— Jenna GaidusekDisclaimer: This blog was written using AI as a recap from the recording then edited by the author for accuracy and details.
