This AI website design case study shows how Wizardly built a full company site for Lithios, an advanced lithium extraction startup. Typically, a project like this would cost $40K and take 10 to 12 weeks. Instead, Wizardly delivered it for $10K in 3 weeks by using AI tools to cut production time without cutting quality.
An AI website design project: $40K scope for $10K in 3 weeks.
Lithios is a lithium extraction startup with serious technology. Specifically, their electrochemical systems pull lithium from brine at up to 40% lower cost than conventional methods. In addition, they had been operating in the field with a key U.S. lithium producer, testing across 20+ brine profiles on 4 continents. However, their website still looked pre-seed.
As a result, they needed a site that communicated enterprise credibility to resource owners, project developers, and strategic partners. In total, the scope included a homepage, technology explainer, product page with scaling architecture, team page, and contact. In other words, not a landing page, but a real company site for a real company.
Traditionally, this approach would have meant a 10 to 12 week build at $35 to $40K. Instead, Wizardly used AI-assisted workflows at every stage, from concept through production, to deliver the same scope in 3 weeks at 75% less cost.
The math that changes the conversation.
Enterprise technology. Pre-seed website.
Lithios had already deployed their ALE packs in the field. Furthermore, they had partnerships with lithium producers. In fact, their technology had been validated across 20+ brine chemistries on 4 continents, and they were logging thousands of hours of operational uptime.
However, when a potential partner or investor visited their website, none of that showed. Specifically, the site did not communicate the sophistication of the product, the rigor of the science, or the scale of the opportunity. As a result, it looked like a company still in the lab, not one already operating in the field.
For a company selling to resource owners, project developers, and strategic operators, brand credibility is not cosmetic. Rather, it is commercial. Above all, the site needed to match the product.
How AI tools shaped this website design case study.
Importantly, this was not about replacing the design process. Instead, it was about compressing the timeline at specific stages where AI tools could accelerate production without sacrificing craft. Consequently, every tool had a clear role, and every output was finished by hand.
Same scope. Different economics.
What $10K and 3 weeks actually produced.
To be clear, this was not a stripped-down MVP. In fact, the site Wizardly delivered for Lithios includes a full technology explainer with process diagrams, a product page with animated pack-to-plant scaling architecture, tabbed brine compatibility sections, a process flowsheet comparison (ALE vs. conventional DLE), embedded product videos, field deployment photography, and a team page.
Overall, the design language communicates what Lithios needed: this is a serious infrastructure company, not a lab experiment. Specifically, the visual approach is dark, technical, and precise, with enough visual energy to hold attention through complex electrochemistry concepts.
What worked and what to watch.
We do not cherry-pick. Here is the full picture.
What Worked
Importantly, this was not a one-off discount. The AI-assisted workflow (Claude Code for prototyping, Weavy for video, AI photo editing for environments) created genuine production efficiencies that apply to similar builds. In particular, the savings came from compressed timelines, not reduced scope.
Because Lithios already had existing CAD drawings as the starting point, Weavy animations meant we skipped the usual 3D modeling phase entirely. As a result, we turned their engineering assets into marketing assets at a fraction of the traditional cost.
By spinning up layout ideas through Claude Code connected to Figma's MCP server, we could explore more concepts faster. Then, we committed to the best direction with full manual design refinement. In essence, the tool expanded the creative aperture without compressing the craft.
For example, the Lithios site has tabbed product sections, animated scaling diagrams, process flowsheet comparisons, and embedded video. These are features that typically live in a $35K+ build. Clearly, the AI tools did not lower the ceiling. Rather, they lowered the cost of reaching it.
What to Watch
Notably, these tools did not automate the build. They accelerated it in the hands of designers who already knew what good output looked like. Consequently, the cost savings are only possible with a team that can direct, evaluate, and finish AI-generated work. In short, the tools amplified skill; they did not replace it.
Since the site launched recently, we can show what was built and what it cost. However, we do not yet have GA4 data on traffic, engagement, or conversion performance. Therefore, a separate data-driven case study may follow once there is enough runway to measure.
The CAD-to-video pipeline worked specifically because Lithios had detailed engineering drawings ready. In contrast, clients without existing technical assets would need more upfront production, which changes the cost equation. Accordingly, the 75% figure assumes reusable source material exists.
Although the landscape and deployment environment imagery reads as realistic, there is a line between "edited for visual impact" and "misleading," especially for a deeptech company selling to technical buyers. For this reason, we were careful to stay on the right side of it.
What would you build if it cost 75% less?
With AI-assisted workflows, startup-stage budgets can now produce enterprise-grade websites. If your site does not match your technology, let us fix that, faster and cheaper than you think.
Start a ProjectEngagement: 3-week AI website design and development project including brand positioning, site architecture, product animation, and environment photography.
Cost comparison based on Wizardly's standard pricing for equivalent scope without AI-assisted workflows.
Video production timeline comparison based on traditional motion graphics production for similar product explainer content.
AI tools referenced: Claude Code (Anthropic) with Figma MCP integration, Weavy for animation, AI-assisted photo editing for environment imagery.
No GA4 performance data available at time of publication. Site recently launched.
Prepared by Wizardly - wizardly.co
See the work behind the numbers.
The results don’t happen without great design. Take a look at the visuals we created for Ketryx during this project.
