case study 03 / 05 — manufacturing

years of manual process automated in one quarter

order intake, quality paperwork, and supplier matching — moved from inboxes and spreadsheets to agents with human sign-off.

client
an industrial components manufacturer, 4 plants
sector
manufacturing ↗
duration
13 weeks
team
5 engineers, 1 ml engineer, 1 process lead
the result, in numbers
71%of order documents processed without touch
9 fteof capacity returned to higher-value work
13 wksfrom first workshop to full rollout
the brief

every customer order arrived as a pdf, an email, or an edi message — and each was retyped into the erp by hand. quality certificates and supplier confirmations followed the same path, adding days to every order.

previous automation attempts relied on templates that broke whenever a customer changed a form.

the build

13 weeks, 4 phases, one team.

  1. 01diagnose2 weeks

    shadowed the order desk, measured volume, variance, and error rates.

  2. 02prove4 weeks

    extraction and matching agents evaluated on six months of history.

  3. 03integrate5 weeks

    erp write-back with approval queues and full audit trail.

  4. 04hand over2 weeks

    evals, monitoring, and retraining owned by the plant it team.

before
  • every order retyped by hand
  • templates that broke weekly
  • two-day order acknowledgement
  • errors found at invoicing
after
  • 71% touchless processing
  • layout-agnostic extraction
  • same-day acknowledgement
  • validation at intake
the team stopped being typists and started being the people customers call first.
— head of operations, an industrial components manufacturer
the result

incoming orders, certificates, and confirmations are read, validated against the erp, and queued for one-click approval. anything uncertain is routed to a person with the evidence highlighted.

the order desk moved from data entry to exception handling and customer contact.

practices
stack
  • python
  • azure openai
  • langgraph
  • sap
  • postgres
  • mlflow
next case study04 / 05real-time operational insight across plants
get started

ready to build something great?