Artificial intelligence is no longer an abstract future topic for the connected home and design-build channel. During our recent Connected Design AI Roundtable, that point came through clearly from both the research that opened the conversation and the panelists already putting AI to work inside their companies.
The discussion built on our recent Industry Pulse Survey, which found that 79 percent of respondents are either actively using AI tools in daily workflows or at least experimenting with them. The most common use cases remain practical: proposals, client communications, product research, marketing, documentation, and support. In other words, the industry is looking for ways to reduce the drag around the mundane work that pulls us away from higher-impact opportunities.
Kyle Steele, founder and CEO of Global Wave Integration and FutureCare Solutions Group, framed the opportunity around bottlenecks. His team looked at AI by first asking where the business was actually feeling pain. One example was proposal generation.
“We figured out it took him on average about four hours to get a bid out from start to finish,” Steele said of one key team member’s proposal process. “But it’s much longer than four hours because he’s always being interrupted.” After implementing AI-assisted workflow tools, he said, “Johnny’s down to 10 minutes to get a bid out.”
That kind of result is why the panel kept returning to AI as a workflow layer, not simply a writing or research assistant. Steele noted that a similar poll inside Azione Unlimited found broad AI usage, but far fewer companies deploying it across the business. The next step, he suggested, is understanding how AI can do things on behalf of a team, rather than simply answer a question before a human goes back to another platform.
For Ari Supran, CEO of Sonance, the real shift happens when professionals stop treating AI as a search box and start treating it as something closer to a coworker. “The goal of every interaction should be to get something done,” he said. That mindset, he argued, is what turns AI uncertainty into momentum. Once someone has that first productive moment, “they start to see the pattern,” and the snowball effect starts to take over.
Supran also encouraged attendees to think carefully about the products they use. AI, he said, is “not one monolithic thing.” A simple chat interface may be useful, but more agentic tools can be pointed at folders, spreadsheets, PDFs, dashboards, project files, and internal workflows. At Sonance, this has led to deeper investment in enterprise AI access, internal skills, and model-agnostic systems that are impacting nearly every facet of their day-to-day processes.
Nick Donahue, founder and CEO of Drafted, made a similar point from the design side. Drafted uses AI to help people generate home design options quickly, shortening the early ideation phase that can otherwise take weeks. “Now you can go in and see iterations spawn in front of you within five, 10, 20 seconds,” Donahue said. The goal is to help users move faster from imagination to a design they can actually advance.
Alex Capecelatro, founder and CEO of Josh.ai, pushed the conversation toward business impact. He challenged attendees to think beyond experimentation and ask a harder question: “How do we double the business without doubling the headcount?” For Josh.ai, he said, the light bulb moment came when AI became something teams could use to build, configure, support, and improve real workflows.
Capecelatro also raised an important caution around data and platform trust. His advice was not simply to ask whether AI can do the job, but to understand the company behind the tool. “The parent company’s motives will always influence whatever the product is doing,” he said. For firms handling client information, project files, home data, and business records, that question will only become more important.
Still, the panel did not present AI as a magic switch. Capecelatro warned that hallucinations and confidently wrong answers remain a real business risk. “It doesn’t mean don’t use it,” he said, “but it means check the answer as though this were an employee presenting you a business plan.” The key is to avoid blindly trusting these platforms. They can make mistakes just like we humans.
That note of caution sat alongside a clear sense of urgency. Donahue encouraged professionals to spend real time testing the limits of the models to “see how autonomous this thing can run and work.” Steele urged attendees to be ambitious while reminding them that AI is only as good as the context and instructions it receives. And Capecelatro was even more direct: “If you don’t embrace this today, if you’re not doing this now, know that your competitors are.”
That balance may be the industry’s clearest next step. AI is already here, but the winners will not be the companies that simply try the most tools. They will be the ones who identify real workflow pain, give AI enough context to be useful, protect their data, verify the output, and keep their people in the loop. As Steele put it in his closing advice: “Every interaction should start with a human and end with a human.”
Bottom line: Start exploring these tools now and don’t be afraid to learn and adjust along the way.
