LivePlan · AI-Powered Business Validation
I designed an AI-powered idea validation tool that helps early-stage entrepreneurs test and refine their business assumptions before committing to a full business plan — transforming LivePlan from an execution tool into a discovery platform.
93.4%
Finished the canvas
Nearly everyone who started completed the core sections — the simplified design kept people from dropping off.
70%
Acted on a pivot suggestion
Most users took the AI up on its suggestions, so the guidance felt like help, not orders.
90.7%
Rated the AI helpful
Users found the AI's evidence and feedback genuinely useful — not generic filler.
The Challenge
LivePlan’s existing users were already past the idea stage — they had a business concept and needed help with funding and execution. But marketing research revealed massive untapped demand higher in the funnel. Keyword search volume and Reddit sentiment analysis showed thousands of early-stage entrepreneurs asking variations of the same question: “Is this a good idea for a business?”
The existing Business Plan tool could technically be used for idea validation, but it was like using a sledgehammer to hang a picture. The world of entrepreneurship already had several tools to help with this: most notably the Business Model Canvas and Lean Canvas. However, those canvasses are intimidating and difficult to execute on for normal business owners.
The opportunity was clear: build a dedicated tool for idea validation that could capture users earlier in their journey and naturally funnel them into LivePlan’s core business planning experience.
A walkthrough of the Idea Canvas validation experience
Discovery
The discovery team — myself as designer and researcher, an engineer, and a PM — worked closely with marketing to define the market need and funnel fit. We approached validation from three angles simultaneously.
Marketing identified the opportunity through keyword search volume for business validation tools and social sentiment analysis of early-stage entrepreneur communities, particularly Reddit.
I conducted interviews with existing early-stage customers using the Business Plan as a makeshift validation tool, and with entrepreneurship educators who teach business validation to first-time founders.
We reviewed tools like VenturePlanner and Leanstack to understand the landscape and identify gaps that a LivePlan-native validation tool could uniquely fill.
Key Insight
Business planning feels like work. It should feel like discovery.
Most idea validation tools require you to identify the assumptions, and then go out and figure out if they’re true — which is a lot of work. After hearing from users that this process is overwhelming and difficult, we wanted this new experience to feel like users were checking something off their to-do list, rather than adding items to it.
This framing shaped every design decision that followed — from the section-by-section AI feedback to the confidence score that tells you whether your idea has legs.
Design Decisions
The traditional single-page canvas is hard to read on screen and its information density is intimidating to people without MBAs. We used the space afforded by a web app — not a piece of paper — to make the task more approachable.
Educators had strong (almost cult-level) opinions about Lean Canvas vs. BMC. The BMC assumes you've already validated your problem and solution — exactly what our users hadn't done. We went with Lean Canvas, but engineered the tool to be framework-agnostic for education partners.
Traditional canvases are great collection tools, but they don't validate anything. We leveraged our existing Market Research deep-research capabilities to provide instant evidence analysis on each assumption — rating them strong, moderate, or weak.
Individual assumptions get evidence ratings, but the Idea Check evaluates the entire business holistically — examining market size, revenue model, regulatory risks, competition, and fatal flaws — producing a confidence score with specific pivot suggestions.
The Product
The star of the product
When an assumption has weak evidence, the AI doesn’t just flag it — it analyzes adjacent possibilities and suggests concrete alternatives. If your solution is “Dog food made of dog poop” (illegal and unsafe), the system suggests pivoting to “Nutritional supplements that help dogs stop eating poop.”
70% of users replace their original assumption with the suggested alternative — proof that AI works best as a collaborative partner, not just a critic.
A weak rating can read as a dead end, but it's really the most engaging moment in the flow. Instead of leaving users stuck, the AI treats that signal as a starting point — surfacing stronger, evidence-backed assumptions in adjacent areas. If a more profitable market shares the same problem, for instance, it points users toward a concrete pivot rather than just flagging the weakness.
Design Approach
Every assumption gets instant AI analysis. No batch processing, no submit-and-wait. The design prioritizes the feeling of having a knowledgeable co-founder looking over your shoulder.
Start with a simple problem statement, layer on solution, market, channels, and revenue. The Idea Check only appears after sufficient sections are complete, preventing premature judgment.
Because the canvas framework maps to LivePlan's business plan template, validated assumptions flow directly into a full business plan — users who've done the hard thinking don't have to start over.
Outcomes
The numbers confirmed the bet: the tool was approachable, the AI guidance landed, and we proved it fast. Idea Canvas also gave LivePlan a way to help people while they’re still exploring an idea — before they’re ready for a full business plan.
93.4%
Nearly everyone who started completed the core sections — the simplified design kept people from dropping off.
70%
Most users took the AI up on its suggestions, so the guidance felt like help, not orders.
90.7%
Users found the AI's evidence and feedback genuinely useful — not generic filler.
3weeks
Shipped Fast
Start to finish — fast enough to prove the bet before over-investing. The whole thing went from a blank canvas to a working product in a single sprint cycle.
Future Vision
I designed a comprehensive traditional validation toolkit staged for future release — customer discovery interview guides, landing page smoke tests, pricing and revenue model surveys, and online community sentiment analysis panels.
Each instrument generates ready-to-use research tools with share links, turning LivePlan into the user’s repository for both AI and human-gathered validation data. Staged design ensures the feature can grow without requiring a redesign.
Prototype screens from the traditional validation toolkit — customer discovery interview guide with auto-generated shareable link and sentiment analysis panel