ADSelected work

AI workflow · Product Operations

Internal knowledge.Instant answers.

A company-wide AI assistant that turned scattered documentation into source-linked answers inside Slack.

SlackMakeGoogle DriveChatGPT

Workflow, knowledge, prompts and rollout by me. Technical implementation with Engineering.

Live systemSLACK / AI / 01
  1. 01
    INPUTFree-text Slack question
  2. 02
    ORCHESTRATEMake + HTTP
  3. 03
    RETRIEVEApproved Google Drive content
  4. 04
    RESPONDAnswer + source link
✓ Source-linked✓ Same language✓ Safe fallback
50employees
6languages
−50%back-and-forth
−30%inaccurate answers

Built together

Product thinking.
Technical partnership.

Clear ownership from design to handover.
MY SCOPE01

I designed the system people would actually use.

Use caseWorkflow designKnowledge modelPromptingValidationRolloutTrainingHandover
ENGINEERING02

Engineering made it run.

Make executionHTTP + APICredentialsTechnical maintenance
01Search firstAnswers grounded in approved content
02Show the sourceDrive link returned with every answer
03Fail safelyFallback plus automatic owner alert

Adoption that lasted

1+year liveStill in use after my handover and departure.

Pilot. Learn. Scale.

Project groupCustomer SupportAll departmentsHR onboarding

About 60 FAQ topics across six languages, plus Product and company documentation.

Anonymised retrospective

Impact figures are approximate, based on Slack message volume, meeting duration and reported answer corrections over six months.

More build evidence

Beyond one workflow.

Automation, integration and AI adoption in practice.
01Zapier

Production automation

Built operational workflows, including routing releases to the right reviewer.

02JSON + Smartcat

Systems integration

Connected translation workflows to internal systems through JSON-based integrations.

03AI rollout

Cross-market adoption

Led requirements, testing, governance and adoption for a customer-facing AI assistant.