Skip to impact
ABAbhirup Banerjee

AI-assisted delivery, measured.

Agile Coach who runs AI-assisted delivery in insurance, measured and human-governed.

PMP · PSM · SAFe 6 Scrum Master. Bengaluru, open to Bengaluru or remote.

abhirup@abhirupbanerjee.comLinkedIn

Impact

  • ~114 hours a month returned to five delivery teams (~0.7 FTE). 50+ AI-coauthored automations on Azure DevOps and Teams, running ~1,400 times a month unattended with human-in-the-loop gates.
  • 955 verified automated test cases, up from 402. Built through a Requirements to Test to Automation traceability engine; caught 100+ mis-tagged tests before they skewed coverage.
  • Governed, not autopiloted. Planner, executor and verifier agents on Anthropic Claude, with dry-run gates, human approvals and separation of duties. 74 of 149 automation commits carry an AI co-author trailer, tracked from the repo.
  • 53 legacy boards consolidated into 30+ live dashboards (DORA flow, quality, gate-pass, risk, commitment vs carryover) that feed Sprint Planning for every POD and leadership.

Now

Project Manager & Scrum Master, Insurity EAIS (AI-Driven Lifecycle) Program

Artizent · Oct 2025 to present · Bengaluru (Hybrid)

Hands-on Scrum Master for two PODs; leads the Scrum Master Community of Practice on a 6-pillar AI-enabled operating model.

  • Coach 30+ engineers and lead the Scrum Master community of practice. Hands-on Scrum Master for two of the five PODs.
  • Azure DevOps process for that lifecycle: 17 work item types and 23 rules. Feature advancement waits on AI-readiness and traceability. Sixteen wrongly regressed milestones put back.
  • The weekly executive deck is generated from the board, and a data check can refuse to publish it.

Before and after AI-assisted delivery

  • Retrospectives

    Before

    Retros were routine and rarely produced real actions.

  • Weekly exec reporting

    Before

    The deck was put together by hand every week.

  • Agent work

    Before

    AI agents ran with no structure and no check on token use.

Stackwich is my personal open-source side project, not an Insurity product. It has helped every Scrum Master on the program. It is not used in developers' code generation.

Path

  1. Artizent / Insurity EAIS

    Oct 2025 to present

    Project Manager & Scrum Master · Bengaluru, hybrid

    AI-SAFe program modernizing a legacy policy-administration system. Four product PODs and a platform POD, under gate-controlled governance.

  2. SS&C Health

    Nov 2023 to Jun 2025

    Senior Scrum Master · Remote, US

    DomaniRx claims adjudication. Processing 20% faster, and post-adjudication downtime cut 15%.

  3. Navy Federal Credit Union

    Feb 2023 to Nov 2023

    Senior Scrum Master · Remote, US

    Mission Data ART. Financial-data accuracy up 25%, system defects down 40%.

  4. Gap Inc.

    Nov 2021 to Feb 2023

    Senior Scrum Master · Remote, US

    Identity transformation on SailPoint and Azure AD. Full security compliance, half the manual provisioning time.

  5. AT&T

    Feb 2019 to Nov 2021

    Agile delivery lead · Richardson, TX

    HALO C backend inside an ART. 1,000+ backlog items, zero regression in hardening, serving HBO Max.

  6. Apple

    Mar 2018 to Mar 2019

    Senior Scrum Master · Sunnyvale, CA

    IS&T Retail engineering. Security vulnerabilities down 90% in four months.

  7. Earlier

    2005 to 2018

    CaptureRX, Texas HHSC, AT&T (Tech Mahindra), Oracle (UBS), TCS, Infosys.

Side projects

Stackwich is my own project, not an Insurity or Artizent deliverable. Some of my program work now runs faster using it.

Side projects, built in my own time on personal subscriptions. No work accounts or work subscriptions used. Vibe-coded with generative and agentic AI, mostly for the fun of building.

How I'd start in your team

  • What would the new hire need to have achieved in a year to exceed expectations?
  • What has the team already tried on this problem, and why didn't it work?
  • Where does delivery lose time today?

Contact