EnergyAI Enablement
The AI habit
that stayed after we left.
The team built shared tools, a stronger engineering environment, and a community to keep learning together.
Engineers highly confident with AI
- Daily AI use
- 50%to100%
- AI rated very effective
- 17%to33%
- Would recommend AI adoption
- 83%to100%
Shared tools · AI-ready environment · peer learning
AI was already in the room.
A rapidly scaling energy technology company runs a large ERP platform its customers depend on every day. AI use was already happening, but experience varied across teams: some were fluent with it, while others were still getting set up.
Leadership wanted a durable, measurable capability without rewriting the ERP. The work had to make the existing environment and delivery practices work with AI tools.
The working environment had to catch up.
The pulse survey showed uneven confidence and use. Some engineers used AI daily, while others turned to it only occasionally. Practices and confidence were concentrated in pockets of the organization.
The codebase added friction: inconsistent style, tests tied to live databases, oversized files that strained model context, and components that were hard to work with. These conditions made AI assistance slower and harder to use safely.
Prompts, workflows, and conventions lived in individual heads. Without a shared foundation, useful learning stayed with the person who found it.
Build a shared operating practice.
The team connected MCP integrations for Azure DevOps, GitHub, and New Relic, configured Claude Code for active engineers, and built shared slash commands for QA-field generation, issue logging, and recording useful patterns. AI-assisted pull-request review joined the standard DevOps flow. A context-priming command brought work-item details, attachments, and dependencies into the model.
The team aligned the environment around the ERP. Linting moved into CI, tests gained isolation through mocks and service abstractions, and dependency injection and smaller files made the code easier to work with.
The practice showed up across roles.
Individual examples showed different uses across the team. One engineer became productive with Claude Code after an access blocker cleared. A product manager connected Copilot to Azure DevOps for backlog and story work. QA automated repetitive test scaffolding, and an engineer turned raw work items into structured development plans from schema context.
Before-and-after pulse surveys recorded a rise in highly confident engineers from 17% to 67%. Daily AI use moved from 50% to 100%, the share rating AI very effective rose from 17% to 33%, and recommendation rose from 83% to 100%.
Ownership stays with the team.
The company's strongest adopters lead an Agentic AI Community of Practice. A shared prompt and slash-command library is in place, and the team is shaping an internal Agentic AI Playbook to keep standards and learning moving after the engagement.