AI discovery and process mapping — sample deliverable from a maritime crewing engagement

    The DiscoveryMaritime Crewing Agency · 22,000-profile database
    Shipped to production

    Chapter 01

    The Discovery

    Two weeks inside the operation. We shadowed the people doing the work, not just the org chart — every workflow documented, every exception categorized, every hour accounted for.

    4

    team members shadowed

    6

    systems accessed

    14

    workflows documented

    9

    automation opportunities

    2-week engagement timeline

    Week 1

    Day 1

    Kickoff + scope alignment

    Owner · Operations lead

    Day 2

    CV intake — full process shadow

    Crewing manager

    GmailExcel

    Day 3

    Candidate search shadow

    Crewing manager

    ExcelEmail archive

    Day 4

    Certificate tracking review

    Operations lead

    ExcelPaper files

    Week 2

    Day 6

    Outreach process mapping

    Crewing manager

    WhatsAppPhone

    Day 7

    Data quality assessment

    ExcelPortals

    Day 8

    Opportunity identification + ROI model

    Day 10

    Roadmap design + final review

    Owner

    Chapter 02

    The Findings

    Where the hours and the deals were actually going. Each finding is quantified and traced to the workflow it lives in.

    01

    CV processing eats the team alive

    Hundreds of seafarer CVs arrive by email every month. Each one is retyped into spreadsheets by hand — 15–30 minutes per CV. That alone is ~20 hours of manager time a week.

    GmailExcel
    HIGH
    02

    The database can't answer questions

    "Find me a Chief Engineer with tanker experience and valid documents" means digging through old emails and memory. A single vacancy search takes hours and misses candidates.

    ExcelEmail archive
    HIGH
    03

    Certificate expiry is invisible

    No tracking of document expiry dates. Expired certificates get discovered at contract time — deals slip, vessels wait.

    ExcelPaper files
    HIGH
    04

    Outreach is one-by-one, with no memory

    Managers message seafarers manually on WhatsApp, one at a time. No follow-up cadence, no record of who was contacted about which vacancy.

    WhatsAppPhone
    MEDIUM
    05

    Data is duplicated and incomplete

    The same seafarer exists in multiple sheets with different data. A large share of profiles miss rank, contact, or experience fields.

    Excel
    MEDIUM

    Chapter 03

    The Plan

    Shipped in waves, not a big-bang launch. The first automation pays for itself before the second one starts.

    Wave 1 · Weeks 1–2

    Email → CV parser

    CVs parse themselves from the inbox into structured, deduplicated profiles. Manual entry ends here.

    ≈20 hrs/week of manager time back

    Wave 2 · Weeks 3–6

    AI search + certificate monitoring

    Natural-language candidate search over the whole database. Every document tracked, expiry alerts on autopilot.

    Vacancy search: hours → seconds

    Wave 3 · Weeks 7–12

    Dashboard, reports + outreach

    One dashboard for the whole operation, Telegram reports, and email + WhatsApp outreach with templates and history.

    Every candidate touch recorded and repeatable

    Chapter 04

    The Results

    This is where our sample differs from most demos: the plan above actually shipped. Numbers from production, three months in.

    21,802+

    seafarer profiles structured

    97%

    extraction accuracy on fresh CVs

    20+ hrs

    manager time back, every week

    10 sec

    to find a matching candidate

    Six modules live: CV parsing from email, AI candidate search, certificate monitoring, dashboard, Telegram reports, and email + WhatsApp outreach. The system keeps learning on every CV it sees.

    The workflow map is yours to keep either way.