Prospecting and account research
Turn lead lists, ICP rules, websites, and public sources into scored account briefs, first-line personalization, and follow-up prep.
Agent workflows for companies
I help teams implement practical AI workflows for research, monitoring, reporting, prospecting, content, and internal operations. The work is built around real files, real tools, human review, and portable skills instead of another black-box platform.
Messy browser research that used to become a manual weekly report.
Human approval remains required before anything is sent or published.
Built for teams that need more capacity without losing quality control.
The offer
If your team repeatedly moves between browser tools, CRMs, spreadsheets, dashboards, docs, tickets, client channels, and internal standards, part of that work can become a controlled agent workflow.
I map one repeated workflow, define review boundaries, build the skill or workflow instructions, test against example cases, and hand off a pilot your team can inspect and improve.
Map the repeated work and approval boundaries.
Build portable instructions, examples, and tool rules.
Test against realistic cases and score outputs.
Run with human review before expanding.
Not another AI platform
The starting point is not a generic prompt library. It is one actual process your team repeats and can judge.
Skills, instructions, examples, rules, outputs, and rubrics are inspectable. Your team should understand what the agent is doing.
Agents can draft, research, summarize, and prepare. Sensitive sending, publishing, mutation, or production actions stay gated.
Workflow menu
These are examples for the first pilot. The strongest offer is tailored after seeing how your team actually works.
Turn lead lists, ICP rules, websites, and public sources into scored account briefs, first-line personalization, and follow-up prep.
Track competitors, search changes, launches, news, campaigns, and market signals, then produce a reviewable client-ready brief.
Convert research, product updates, datasets, and founder notes into posts, briefs, landing-page sections, and reusable source summaries.
Use agent-assisted browser checks, repository context, screenshots, and review rubrics to catch UI, copy, and workflow issues before release.
Proof
The commercial offer is new, but the underlying work is not: startup research systems, agent skills, source-backed writing, browser workflows, and experimental agent UX.
A startup research product and proof engine for competitor discovery, Product Hunt intelligence, market maps, source review, and founder-facing reports.
Portable skills, rules, examples, QA rubrics, and implementation patterns for teams adopting Claude Code, Codex, Cursor, and agent runners.
A local reading layer that turns passages and voice notes into Second Brain synthesis, project insight, and durable source-linked knowledge.
A portable psychological and continuity layer for agents, exploring memory, personality, inner process, and relationship-bearing software.
First step
Send one repeated workflow your team handles every week. I will review it and show where an agent workflow could help, what should stay human, and what a pilot would include.