← Work/Case study
How an interior design studio compressed repetitive pre-pitch research with practical AI.
Before every project, the same research had to happen again: collect reference photos, sanity-check listing context, and understand the neighborhood before the client conversation. The answer was not a big custom app; it was better AI-assisted research behavior plus a few lightweight tools.
The problem
Designers spent real time researching the homes they were about to walk into. The work was not hard once; it was hard because it repeated. Someone had to pull photos, compare listing context, and build enough market understanding to walk into the first conversation prepared. The real question was not just “what should we build?” It was “how can the team use existing AI tools and smarter habits to do this better?”
What shipped first
Botworks shaped the work into a practical AI-assisted research workflow, then added small tools around the repeated steps that still needed structure. The studio did not need a full platform. It needed a better way for existing employees to prepare faster and more consistently.
Why off-the-shelf software did not solve it
The work crossed public listing sites, market context, visual reference gathering, and the studio's own client-prep habits. Generic design software did not own the research step, and real-estate tools were not built for a designer walking into a budget conversation. The gap was small but persistent, which made lightweight custom tools a better fit than a new platform.
AI habits plus lightweight tools
Tool 1
Listing photo scraper
Pulls every photo from a real-estate listing across the major US sites in one shot — for reference boards, before-and-afters, or briefing files.
Designer use: assemble a visual brief on a property before the client meeting without manually saving 40 photos.
Tool 2
Estimate accuracy checker (active listings)
For a home currently for sale, walks the AVM's own estimate history and recent comps to score how trustworthy that estimate is.
Designer use: before pitching scope, know whether the price the client is anchored to is defensible.
Tool 3
Estimate checker (sold homes)
For a home that already sold, back-calculates what the AVM thought the day it went under contract and compares it to the actual sale.
Designer use: sanity-check what comparable homes in the neighborhood actually closed at — useful when the client is benchmarking budget.
What became possible next
The tools are small, but the operating pattern is the same as the HVAC case study: find the recurring manual step, decide whether AI behavior or software should help, and let the business decide whether the next bottleneck is worth building around. Different industry, same motion.
What the client owns
The workflow pattern, the AI-use guidance, the tool instructions, and the outputs created for client prep. This was an early lightweight engagement, but it follows the same principle as larger Botworks systems: the business should be able to use the work without being trapped inside an agency black box.
Want better AI workflows for your team?
Tell me what your team researches before every project — the patterns, the time sinks, and the places where AI might help. I'll help sort what to teach, automate, or build.