Type
AI Product Design, Lead Generation, Interactive Experience, UX Strategy
Industry
Staffing, Executive Assistant Placement, AI-Enabled Recruiting
Client
DOXA Talent
Status
Delivered to DOXA Talent, not yet publicly deployed
Built With
Vercel, Supabase, Resend, Claude API
Project Overview
Many business owners know they’re overwhelmed but aren’t sure whether they’re ready to hire an Executive Assistant, or what they’d actually hand off first. DOXA needed a consultative lead-generation experience that educated prospects while qualifying them for a discovery call, without feeling like a staffing-vendor pitch.
Designed and built the entire assessment, concept, three-step user flow, interface, scoring logic and AI-generated personalized recommendations, end to end in Claude.
The Challenge
A generic contact form can’t tell DOXA whether a visitor is genuinely ready to hire, or convince that visitor they’re ready either. The assessment needed to do both jobs at once: educate someone who has never worked with an offshore EA on what that relationship actually looks like, and generate a specific enough answer that a vague “maybe someday” turns into a booked call.
- Replace a generic lead form with a consultative, educational experience.
- Generate a genuinely personalized result, not a static outcome from a lookup table.
- Translate someone’s specific workload into the specific tasks a DOXA EA would take off their plate.
- Make the DOXA cost comparison concrete and credible, not just a marketing claim.
- Qualify leads through the interaction itself, then hand them straight to a discovery-call booking flow.
The Approach
Designed and built the entire assessment in Claude, from concept and user flow through interface, scoring logic and personalized recommendations. The result is a three-step interactive experience, About You, Your Situation, Your Result, that evaluates a prospect’s delegation readiness and generates a customized readiness score identifying the specific tasks a DOXA Executive Assistant could take off their plate from day one.
Generated live, not canned
The eight-question quiz feeds directly into the Claude API, which writes the visitor’s result in real time: an overall EA Ready score out of 100, a three-part breakdown across time pressure, delegation readiness and operational load, and a personalized paragraph addressed to them by name and role. Nothing about the result is a static template with variables swapped in. It’s written fresh for every visitor. Built and deployed on Vercel, with Supabase storing each lead and result behind a simple internal log so the team could see who had completed the assessment, and Resend delivering the personalized assessment straight to the visitor’s inbox.
A concrete “what changes”
The result page names the specific, AI-trained work a DOXA EA would own from day one: inbox triage, calendar management, meeting summaries, client follow-up, travel logistics, research, alongside a direct cost comparison between a fully loaded US-based EA and a DOXA placement. Abstract value becomes a specific, defensible number.
Built to cross-sell and convert
Every result routes to a “Book a free discovery call” CTA, with a secondary path into DOXA DreamTeam AI for visitors who want to see what a full AI-enabled team could look like, connecting this assessment to the wider DOXA AI-lead-gen ecosystem rather than treating it as an island.
The assessment
A personalized, AI-written result
What Makes It Work
Consultative, not transactional
Eight honest questions replace a five-field contact form. By the time someone reaches their result, they’ve already thought through their own workload. The assessment does the consulting before a human ever joins the call.
Genuinely personalized
The Claude API writes each result from that visitor’s actual answers, name, role and industry, not a lookup table of five preset outcomes. It reads like a considered response, because it is one.
Specific numbers, not vague claims
A named score, a three-part breakdown and a direct cost comparison replace generic language like “save time and money.” Specificity is what makes the pitch credible.
Built, not just designed
Concept, interface, scoring logic and the live AI integration were all built directly in Claude, with no separate design-to-development handoff between the idea and the working tool.
Our Role
We owned this project end to end for DOXA Talent: concept, user flow, interface design, scoring logic and the working AI integration, with no separate design or development partner. Every screen shown on this page came out of that single build.
The Outcome
Delivered an AI-powered lead magnet that educates prospects, personalizes the buying experience and generates highly qualified leads. The assessment helps shift the sales conversation from “do I need an EA?” to “when can I get started?”, while reinforcing DOXA’s consultative approach to outsourcing rather than a transactional staffing pitch.
For us, the project is another working example of AI design engineering end to end: strategy, UX, interface and a live AI integration, conceived and built in Claude from a blank page to a finished, working tool, without a separate development handoff.