Feature Management & Experimentation
Explore planned Feature Management and Experimentation capabilities and enhancements.
This page tracks planned capabilities and enhancements for Feature Management and Experimentation.
SaaS release status: GA, SMP release status: Not yet planned
Now
Q3 2026, Aug-Oct 2026
Config Management
Enterprise-grade configuration management with approvals, audit logging, impression tracking, schema validations, and real-time delivery.
Measurement for Config Management
Integration with Cloud Experimentation to track Config Management as assignment sources, enabling impact measurement and alerting capabilities.
Governance for Config Management
Granular control over Config Management through RBAC, OPA policy enforcement, and enhanced audit trails.
Pipeline-gated flag changes
Enforce strict release workflows by gating feature flag updates behind your pipelines, third-party integrations like ServiceNow, cross-system validations and more.
AI Config Management
Extended Config Management to include out-of-the-box templates for AI product configurations.
Config Management, continued
Continued expansion of Config Management across both standard and AI configs, including progressing AI Configs from alpha toward general availability.
Harness dashboards for FME
A unified set of dashboards that combine production activity and usage insights to keep teams aligned on system health and impact.
Dart SDK support
Extend cross-platform coverage with the introduction of Dart SDK, enabling direct use on native Dart applications.
Expanded guardrail metric insights
Understand the impact of feature releases on guardrail metrics with causal insights for percentage-based rollouts and correlational insights for all other releases.
Programmatically manage experiments and metrics
Enable CRUD capabilities for experiments and metrics using APIs and MCP, allowing for automated configuration and external integration.
Next
Q4 2026, Nov-Jan 2027
Feature Flags as Code
Manage the full lifecycle of flags directly from your codebase using GitX, enabling version-controlled configurations and code-reviewed releases.
Additional data warehouse support
Extend Warehouse Native Experimentation to support Databricks.
CUPED for Cloud and WHN Experimentation
Introduce CUPED to leverage pre-experiment data and reduce variance across cloud and warehouse-native experiments.
Flag prerequisites
Top level flag dependencies enforced (new) as well as within rules (existing).
Additional data warehouse support
Extend Warehouse Native Experimentation to support Trino.
Policy as code for experiments
Leverage Harness Policy as Code to enforce governance policies on experiment changes.
Unified billing and setting management
Unified control plane for billing and subscription management with streamlined FME settings.
Experiment pipeline support
Treat experiments as pipeline objects in Harness pipelines.
Expanded language support for SDKs
Extend cross-platform coverage, including support for additional languages.
Pipeline steps for Config Management
Manage Config Management changes directly from Harness pipelines with dedicated pipeline steps.
Expanded thin client SDK support
Expand remote evaluation (secure mode) support to additional thin client SDKs, including React.
Later
Q1 2027+, Jan 2027 & beyond
Dimensional Analysis for WHN Experimentation
Introduce dimensional analysis for Warehouse Native Experimentation (WHN) to uncover deeper trends and segment-level impact.
Experimentation Agents
A suite of intelligent agents to automate the full experiment lifecycle, from hypothesis design and prioritization to results analysis and value realization.
Additional OpenFeature Provider support
Expand multi-language support for vendor-agnostic feature management, including Ruby.
Project movement support
Allow FME projects to be moved from one Harness organization to another.
Audit log unification with Harness Audit Trail
Deliver FME audit logs and admin audit log events in Harness Audit Trail to support a single source of truth.
Released
What has been released
Feature Flag Cleanup Agent
Automate technical debt reduction with an agent that identifies and helps remove stale feature flags directly within the new UI experience.
AI chat for flag operations
Perform CRUD operations, modify flag definitions, and check rollout status through a natural language interface powered by the Harness Model Context Protocol (MCP) Server.
Harness multi-environment support
Support for additional Harness environments, including EU and Single Tenancy.
Metric checks in automated release pipelines
Integrate built-in data checks in your pipelines to automatically validate release health based on performance metrics.
Remote evaluation client-side SDKs
The power of choice: Ensure rule privacy exactly where you need it with a cloud SDK engine that evaluates flags remotely for thin SDKs clients, eliminating rule exposure.
Advanced configuration support for pipelines
Enhanced pipeline support for segments, flag sets, impression toggles, and metadata for better release tracking.
Additional data warehouse support
Extend Warehouse Native Experimentation to support Google BigQuery.
Warehouse Native Experimentation
Run experiments directly in your data warehouse with Warehouse Native Experimentation, now generally available.
Split integration into Harness
Support for additional Harness environments, including Prod0, Prod3, and Prod4.
OpenFeature provider updates
Ongoing support for OpenFeature providers in .NET, Python, React, and Angular.
Granular permissions in RBAC
Migrate permission management for FME object and environment-level permissions to Harness RBAC. Previously titled, Split integration into Harness, Part 2.
Alert webhook
Automate downstream processes based on FME data with alert webhooks.
Flag impressions properties bag
Decorate impression records with properties to use in downstream processing. Support for the following SDKs and Split Proxy / Synchronizer is coming this quarter: Go, PHP Thin Client (via SplitD), .NET, and Flutter.
OpenFeature provider updates
Ongoing support for OpenFeature providers in Android, iOS, Web, Angular, React, Java, Node.js, Python, .NET, and Go SDKs.
Harness Forward proxy for FME
Centralize traffic going outside of a customer's cloud. Unlike the current Split Proxy, it does not require environment-specific configuration.
Experiment entry event filter
Define an entry event at the experiment level to filter exposures, ensuring sample sizes only reflect users who were exposed to your experiment.
Fallback treatments
A configuration option that lets you define a default treatment and optional configuration to be returned instead of the standard control.
Rule-based segments
Assign feature availability for user groups based on different conditions, with all the power of FME targeting.
Client side SDKs cache expiration
Control when local cache on device expires.
Flag impressions toggle
Disable the flow of impressions for individual flags.
Reimagined experimentation design
New workflow for designing experiments, decoupling experimentation analysis from flag monitoring use cases.
Reimagined experimentation dashboard
Tabular experiment results dashboard + new features like comparison of multiple treatments & sample size visualization.
Experiment sample population chart
See accumulation of sample population over time. Identify unexpected assignment or traffic level changes.
Flag impressions properties bag
Decorate impression records with properties to use in downstream processing. Supported in Browser, Android, iOS, JavaScript, React, React Native, Redux, Node.js, and Java SDKs.
Elixir SDK
First of new SDKs to be added after joining Harness.
AI results interpretation conversation
AI-generated metric results summary can be asked follow-up questions.
Large segments
New segment type enabling large-scale audience targeting up to 1M keys. Even higher limits available by request.
Access Split from within Harness app
Allow Harness customers to authenticate and access Split from the Harness application.
Last updated
Was this helpful?