Insights from 80+ practitionersAI Usage Survey 2025

EmeSoft AI Center

A manifesto, radar, and metrics hub grounded in how our developers, quality analysts, data, and cloud teams really use AI each sprint-translating raw survey data into a forward-looking playbook.

Book an Immersion

“

AI isn’t a cost - it’s the engine behind operational excellence and future growth.

”

AI Applications Manifesto

8 principles that keep our AI practices grounded

Built from our AI Usage Survey, these shared themes show a clear message from Dev, QC, Data, and Cloud teams: use AI to automate, but keep accountability at the core.

secure-transparency

Secure and Transparent by Default

AI tools are used through approved accounts only, with activity logs always on. Every AI-generated output clearly shows its source so teams can understand how it was created.

human-oversight

AI Assists, Humans Approve

AI helps us work faster, but people make the final decisions, especially for important code, system changes, and data work. Human oversight stays at the center.

prompt-context

Prompt Craft Needs Rich Context

Good results require good context. Every AI request includes basic business goals, technical details, and limits so the answers are accurate and dependable.

tool-mix

Reliable Multi-Tool Stack

We combine tools that work well together instead of relying on only one. A balanced toolset gives stronger results and steady productivity across projects.

selective-automation

Automate the Simple Work

We let AI handle routine and repetitive tasks so engineers can focus on deeper thinking, problem-solving, and work that requires judgment.

evidence

Track the Numbers That Matter

AI usage is measured with simple metrics like time saved, quality of output, and number of issues found. Decisions are based on data, not guesses.

sustainable-scale

Be Smart About Costs

Teams watch AI usage and use shared prompts or templates to avoid waste. Smart spending ensures we move fast without overspending.

feedback-loop

Feedback Helps AI Improve

Everyone shares regular feedback on what works and what doesn’t. These insights help us refine tools, prompts, and workflows so AI becomes more useful over time.

Technology Radar

Where EmeSoft stands today

Quadrants reflect sentiment on the tools/techniques that move our work. Rings show maturity - from Adopt (standard) to Hold (watch only).

Platforms

Conversational interfaces and AI suites with governance

Languages & Frameworks

AI-first IDEs, prompt packs, and workspace patterns

Tools

Embedded assistants living inside workflows

Techniques

Reusable prompting strategies and delivery rituals

Ring legend

Adopt

Default choice across teams

Trial

Being scaled with controls

Assess

On the watchlist

Hold

Pause until clarity

ChatGPT

Platforms · Adopt

Primary conversational workspace with audit logging for all roles.

AI Metrics

Survey-backed KPIs

These stats come from the 2025 AI Usage Survey. We track how much teams use AI, how much faster they work, and which use cases actually became part of our daily workflow.

Org-wide AI Adoption

84%

+9 pts YoY

Teams with weekly AI usage

Avg Productivity Lift

27%

+4 pts QoQ

Self-reported 4-sprint rolling

AI-Assisted Pull Requests

58%

+11 pts QoQ

PRs merged with LLM suggestions

Decision Confidence

91%

Stakeholders satisfied with AI reviews

Adoption vs Productivity

Role-level data from sprint retros

% of role population

Productivity trend

Quarterly observed lift

Tool share

Share of weekly mentions

Top use cases

Developers

Code generation38%
Refactoring24%
Unit tests22%
Debug support16%

Quality

Test cases40%
Bug triage28%
Regression plans19%
Exploratory ideas13%

Data

SQL/ETL36%
Insights27%
Validation21%
Dashboard copy16%

Cloud

IaC scripts33%
CI/CD26%
Troubleshooting23%
Cost tuning18%

Developers

Quality

Data

Cloud

Role-based AI

How every role applies AI daily

Bullet points and KPIs come from team retros plus the dedicated Dev, QC, Data, and Cloud survey deep dives.

Developers

Copilot is embedded in VS Code for 92% of product squads, shortening boilerplate delivery and unit-test creation.

Adopt

Generate CRUD services, DTOs, and GraphQL schemas
Refactor legacy modules with context-aware prompts
Draft unit tests and integration test scaffolds

Productivity lift

+28%

Hours saved / sprint

~210

Quality Engineers

QC teams rely on ChatGPT + Copilot to ideate negative paths, with 64% fewer missed regression scenarios.

Adopt

Generate test cases mapped to user stories
Cluster defects to highlight repeat offenders
Draft exploratory charters and acceptance notes

Automation coverage

78%

Productivity lift

+24%

Open the QC AI playbook

Data Engineers

Data squads prompt AI workspaces for SQL, dbt macros, and anomaly explanations to accelerate analytics sprints.

Trial

Generate SQL/ETL skeletons with validation steps
Explain anomalies in dashboards for stakeholders
Draft data-quality rules and schema evolution plans

Cycle time reduction

-32%

Adoption

76% weekly

Dive into the Data Engineer playbook

Cloud Engineers

Cloud enablement pairs Copilot with Terraform + GitHub Actions templates to keep infra stories on schedule.

Adopt

Generate Terraform and Bicep scaffolds
Draft CI/CD workflows with security gates
Troubleshoot CLI errors and tuning guidance

Productivity lift

+31%

IaC automation

68%

Review the Cloud Engineer playbook

Partner with us

Bring the EmeSoft AI Perspective to your org

We run immersion workshops sharing the manifesto, radar, and playbooks. Get a tailored plan using the best practices Dev, QC, Data, and Cloud teams validated in the AI Usage Survey.

ai@emesoft.net

Avg response time: under 24 hours