Core BDR Evidence
Looproof / Automated B2B Lead Generation Workflow
Built a B2B lead-generation workflow that identified high-intent companies through hiring activity signals and enriched decision-maker contacts for outbound sales.
- Confirmed outcomes
- Generated 2,000+ B2B leads and reduced lead research time from two weeks to two hours.
- Commercial judgment
- Used pipeline evidence and buyer-risk analysis to evaluate demand quality before deeper product commitment.
- Anthropic BDR mapping
- Pipeline generation, strategic outbound, initial qualification, lead handoff quality, and process refinement.
| Source signal | Targeted companies actively hiring on Wanted and Groupby because hiring activity was the strongest visible signal of HR workflow pain and budget movement. |
|---|
| Data workflow | Reverse-engineered recruiting-site API structures, collected company lists, then chained Perplexity search with GPT-4o-mini extraction to classify official, recruiting, and manager-level contact paths. |
|---|
| Outbound quality | Built multi-touchpoint lead records so outreach was not dependent on one generic inbox. Used company context and short personalized openers instead of one bulk message. |
|---|
| Run output | Built the crawler in two days, created 2,000+ outbound-ready leads, improved the share of recruiting-relevant contacts, and helped open conversations including NICE Information Service. |
|---|
| Learning | Random cold email reads as spam; outreach to companies with visible hiring intent can read as a relevant proposal. The next improvement would be lead scoring across the 2,000-account list. |
|---|
Open workflow detail
AI Launch Ops Evidence
ArtistAgent / AMA
Built the launch and lead-routing layer for an AI-assisted workflow product across landing pages, UTM-tagged acquisition paths, payments, CRM-style routing, FAQ content, support entry points, and outreach logs.
- Confirmed outcomes
- Maintained outreach operations with 146 logged attempts and separated executed outreach from channel-level failures.
- Operating logic
- Connected acquisition paths, support entry points, and follow-up records so early interest could be routed and learned from.
- Anthropic BDR mapping
- Inbound handling, CRM-style routing, follow-up hygiene, launch ops, and AI product communication.
| Launch surface | Connected landing pages, UTM-tagged acquisition paths, payment flow, CRM-style routing, FAQ content, and support entry points into one launch layer. |
|---|
| Outreach ops | Logged 146 outreach attempts and separated executed outreach from failed or blocked attempts so follow-up quality and channel learning did not get mixed together. |
|---|
| BDR relevance | Shows the operating habit behind lead handoff: route interest, preserve context, keep follow-up records clean, and learn which channel actually produced usable demand. |
|---|
AI Product Communication Evidence
Performance Marketing Agency CS Chatbot
Improved an AI customer-support workflow for a performance marketing agency by defining routing logic, evaluation structure, and grounded-answer constraints.
- Confirmed outcomes
- Increased answer accuracy from 27% to 85% on the same internal test set.
- Operating logic
- Separated routing, grounding, and evaluation instead of treating model choice as the only variable.
- Anthropic BDR mapping
- Technical product translation, grounded AI expectations, responsible explanation, and customer-facing clarity.
| Problem | A single-path CS chatbot could not balance policy stability, grounded coverage, latency, and cost. Always-on RAG caused instability; always-off reduced coverage. |
|---|
| Owned work | Defined routing signals, RAG ON/OFF decision rules, router-only output schema, grounded-answer constraints, and a fixed evaluation protocol on roughly 200 queries. |
|---|
| Operating decision | Split policy/structured FAQ from document-grounded explanatory cases, then refined routing exceptions after reviewing failure modes. |
|---|
| BDR relevance | Gives concrete language for explaining AI systems without overpromising: what should be routed, what should be grounded, and when the system should constrain itself. |
|---|
Open AI workflow detail
GTM Experiment Evidence
Handy
Ran positioning and acquisition experiments for a fintech product, using paid tests to refine target segments and go-to-market direction.
- Confirmed outcomes
- Validated early message-market fit with CTR 11.48% and CPC $0.39.
- Operating logic
- Used acquisition data to refine segment focus instead of relying on broad positioning assumptions.
- Anthropic BDR mapping
- Data-driven insights, message testing, target refinement, and GTM learning loops.
| Market test | Ran paid acquisition tests for a fintech/tax product and used real user behavior, not only positioning opinion, to refine the target segment. |
|---|
| Iteration | Moved from a broader test with CTR 4.27% to a narrower Uber/freelance segment that reached CTR 11.48% and CPC $0.39. |
|---|
| Commercial signal | The project connected acquisition performance to investor-facing GTM validation, including Techstars Tokyo shortlist evidence. |
|---|
| BDR relevance | Useful as supporting evidence for message testing, target narrowing, and using data to decide which audience deserves more sales attention. |
|---|
Inbound Learning Evidence
PDF Grammar Checker
Turned an internal workflow tool into a live web product and used concierge-style email support to capture and learn from early inbound leads.
- Confirmed outcomes
- Moved from internal tool use to live beta operation with early inbound learning.
- Operating logic
- Handled edge cases manually instead of dropping early users, using support conversations as product and demand signal.
- Anthropic BDR mapping
- Inbound handling, customer signal capture, support-led qualification, and early process improvement.
| Internal problem | Reduced repeated PDF proofreading work by moving from manual extraction to a web product workflow. |
|---|
| Product workflow | Packaged the tool as a browser-based Flask/Railway beta and used a hybrid parsing approach with PyPDF2 and Google Vision for files without text layers. |
|---|
| Inbound handling | When files failed, directed users to email files manually and returned results concierge-style, turning failure cases into support conversations and QA data. |
|---|
| BDR relevance | Shows early inbound behavior: capture the lead, keep the user from dropping, learn why the workflow failed, and convert support logs into product/process improvements. |
|---|
GTM Strategy Evidence
Landbase GTM Strategy Proposal
Prepared a GTM strategy document for an agentic AI sales platform, covering ICP selection, trigger-based outbound, PQL handoff, paid pilot design, activation metrics, risks, and stop rules.
- Confirmed scope
- Mapped ICP prioritization, trigger-based outbound, PQL definition, 90-day paid pilot, 14-day Aha flow, Pilot-to-Annual conversion, and stop rules.
- Operating logic
- Focused on observable buyer signals, handoff timing, proof generation, and clear criteria for when the motion should change.
- Anthropic BDR mapping
- Strategic outbound thinking, qualification criteria, sales handoff design, and API/AI GTM literacy.
| ICP logic | Prioritized accounts with visible pressure to improve sales productivity and enough urgency to prove value quickly. |
|---|
| Outbound motion | Used triggers such as hiring, growth, sales motion changes, and process pressure rather than generic account lists. |
|---|
| Handoff rule | Defined PQL around campaign execution, proof-report completion, and teammate invitation before deeper sales motion. |
|---|
| Risk control | Included stop rules so GTM activity would change when signal quality was weak instead of simply increasing volume. |
|---|
Open GTM strategy detail