In the latest episode of From the Ground Up, Matt Jezyk (VP, Industry Strategy and Solutions) sat down in person at Hart Howerton's New York office with Noushin Radnia from their design technology team, joined remotely by Tatjana Dzambazova, Motif's Head of AI. Noushin was candid about what the evaluation process actually looks like, how adoption spreads without a top-down mandate, and where AI rendering fits into a workflow that still includes trace paper.
TLDR: The AEC firms moving fastest with new technology are the ones with a rigorous evaluation process, a small group of curious people, and a platform that makes results visible across the whole team.
What a rigorous tool evaluation actually looks like
Hart Howerton is a multidisciplinary design firm with nearly 300 people across five offices. Noushin's design technology team supports all of them which means every tool decision carries real weight. Before anything reaches the studio floor, it gets vetted for licensing, data privacy, and training burden.
The question guiding that evaluation isn't "is this impressive?" It's: who built this, how long will they be around, and what happens to our data? For AI tools specifically, contractual clarity on model training has become a baseline requirement in enterprise AEC procurement. Firms that can't get a clear answer from a vendor are moving on.
Interface familiarity is the other gate. A platform that requires its own training program is a hard sell when designers are already context-switching between Revit, Rhino, Bluebeam, and more. The tools that spread organically — without a mandated rollout — tend to be the ones that feel like something designers already know.
How adoption actually spreads through a firm
Hart Howerton now has around 60 active Motif projects across the firm. Noushin's explanation for how it got there was simple: "Our most valuable asset is curious people."
The turning point she described was a competition. The team was working under pressure, producing high-quality output, and hit every internal and external deadline. That surprised people. The day after submission, three or four colleagues walked in asking how to get started. "They associated that quality with Motif," she said. From there, adoption spread the way it tends to in design firms — not through training programs, but through trust. Give people a playground, let them see what they can produce, and curiosity does the rest.
The fact that Motif runs in a browser helped too: no IT installation, no desktop dependency. Designers could pick it up on an iPad or phone without a ticket to the help desk which, as Matt noted, is still a relatively new thing in AEC and one that quietly lowers the barrier to getting started.
How to make AI rendering actually useful for design development
The shift in AI rendering that Noushin described came down to geometric precision. Earlier AI rendering tools could produce compelling images, but with no reliable connection to the model. Proportions shifted. Perspectives drifted. For senior designers trying to use AI rendering as a communication tool rather than a finishing step, that gap was disqualifying.
What changed was the connection between live 3D model capture and the rendering engine. When a render is directly informed by the geometry in Revit or Rhino, the feedback loop runs both ways. Noushin shared a telling example: a landscape principal discovered that two palm trees sketched at the wrong scale came back at exactly the wrong scale in the render. His reaction wasn't frustration — it was: next time I need to sketch in scale. That's the difference between a tool that generates outputs and one that understands architectural intent.
Setting realistic expectations for where these tools fit
When asked about balancing platform depth with simplicity, Noushin was direct: "For us, Motif is still an extension of how we are currently working." Hart Howerton still uses Photoshop for final image processing, still prints renderings, still has senior designers on trace paper. A new platform doesn't displace that — it connects the authoring tools to the people who need to see the work, across disciplines and offices.
For design technology teams building a case internally, that framing is useful. The tools most likely to get adopted aren't the ones promising to replace existing workflows. They're the ones that slot in between what people already do and work the way they do.
What's new in Motif: March 2026
Tatjana walked through three features shipping this month:
Resolution selection. Designers can now choose 1K, 2K, or 4K at generation time. Lower resolutions are significantly faster for early-phase exploration, with quality that holds up for internal iteration. Upscale when a direction is worth keeping.
Image upscaling. Images generated at 1K or 2K can be upscaled after the fact, decoupling the exploration workflow from the final deliverable.
Text-to-image generation. Generate any images from text prompts directly in the board — and create contextual mood boards rather than assembling references manually.
From the Ground Up is Motif's webinar series exploring how technology is reshaping AEC practice. Watch other webinar episodes on YouTube or explore Motif's AI visualization capabilities.
