Analyst Report
IDC Spotlight: Why Synthetic Data Is the Next Agentic Development Priority
Data Management,
DevOps,
AI
The AI Validation Gap: Why Agentic Development Is Waiting on Test Data
AI-assisted coding tools and agentic frameworks now draft, refactor, and extend code in seconds. Testing hasn’t kept pace. Every feature generated with AI must be tested, validated, and provisioned before it can ship. The data those tests need is often what holds everything up.
This IDC Spotlight paper looks at Delphix Synthetic Data, an AI-first synthetic data solution. It examines how Delphix masking and on-demand synthetic data generation, governed as a single pipeline, can help close the AI validation gap.
Fill out the form to get your copy and explore:
- Why legacy approaches fall short and how you can overcome the test data bottleneck in agentic development.
- Benefits of adopting a modern synthetic data solution and why synthetic data is key in AI software development.
- Consideration of Perforce Delphix Synthetic Data, including key technical capabilities and use cases.
What IDC’s Research Tells Us About Agentic Development
- 72% of organizations have AI agents in production in at least one business area. (IDC FERS Survey 2026, Wave 1, March 2026)
- Delphix Synthetic Data is built to support those agents, with an API and MCP-accessible design that lets data requests be triggered from automated pipelines.
- 65% are actively deploying agentic AI for test data generation and management, the most-used of nine agentic software testing use cases surveyed. (IDC Agentic Application Development and DevOps Survey, December 2025)
- Delphix Synthetic Data is the first AI‑native solution designed to generate realistic, scenario‑specific test data.
- 55% cite sensitive data exposure as their top agentic AI security concern. (IDC Agentic Application Development and DevOps Survey, December 2025)
- Delphix keeps sensitive data safe by never allowing it to enter non-production and never sending production data to an LLM.
- 68% expect spending and coverage for agentic AI in software testing to increase over the next 12 months. (IDC Agentic Application Development and DevOps Survey, December 2025)
- This reinforces the importance of investing in the right data management platform to support AI software testing while keeping data safe.
Five Questions the IDC Spotlight Answers
Why is test data the bottleneck when AI writes the code?
Coding and prototyping have accelerated, but testing and environment provisioning haven’t. Conventional ways of finding, masking, or subsetting production data were never built to keep pace with agents.
Delphix Synthetic Data helps close the gap. Unlike legacy generators and point solutions, it is part of one platform that combines synthetic data, masking, data delivery, and central governance. Teams and agents get data on-demand, delivered to their own environments, and it scales as teams and agents grow.
Masking or synthetic data?
The IDC study’s answer is both. Masking protects real production data and suits tests that must reflect real-world behavior at scale. Synthetic generation creates data that doesn’t exist yet for greenfield development, new features, edge cases, and negative testing. Delphix does both, and together, they run through one governed pipeline.
What makes a synthetic test data solution trustworthy?
Realism and referential integrity. Synthetic data that doesn’t reflect production distributions and business rules creates false confidence, and integration tests can’t run when keys don’t align across systems.
Delphix Synthetic Data profiles production data so generated data matches its real-world shape and distribution. It also keeps keys aligned within and across databases.
Can a general-purpose LLM generate your test data?
The IDC study finds that prompting a general-purpose model isn’t the same as operating a purpose-built synthetic data solution, which must also deliver governance, scale, and repeatability. Delphix Synthetic Data uses AI too, but it does what LLMs can’t: referential integrity across systems, business-aware data, governance, and automated delivery.
What if even masked production data is off-limits?
A small but growing number of regulated organizations bar any production data, even masked data, from non-production environments. This makes fully synthetic generation the only compliant option. Delphix Synthetic Data supports synthetic-only environments that meet these policies.
Who Should Read “IDC Spotlight: Why Synthetic Data Is the Next Agentic Development Priority”?
This Spotlight paper is written for the leaders responsible for releasing software safely at AI speed:
- Application development leaders whose teams are generating code faster than they can test it.
- QA and testing leaders who need realistic, referentially intact data for every scenario, including edge cases.
- Test data management leaders modernizing how data is delivered, masked, and generated.
- DevOps and IT Operations leaders working to remove provisioning waits from the pipeline.
- InfoSec and compliance leaders responsible for keeping sensitive data out of non-production environments.
CIOs, CTOs, and CISOs sponsoring agentic AI initiatives will find a clear case for putting test data on the agenda.
A Preview from the Analyst Paper
“IDC recommends that organizations investing in agentic development assess the need for a modern validation strategy that directly closes the test data gap, rather than assuming that faster code generation will automatically translate into faster overall software delivery.”
Read IDC’s Analysis on Synthetic Data for Agentic Development
See how enterprises can close the AI validation gap with governed, on-demand test data. Complete the form to access IDC’s full analysis, covering:
- The trends shaping adoption.
- The challenges to plan for.
- How the Delphix platform solves the gap by bringing together synthetic data generation, masking, data virtualization and delivery, and centralized governance.