Forward launches AI tool to test network changes first

Forward has made Forward Predict generally available. The software is designed to test proposed network changes before they are applied to live systems.

The launch marks a broader push into tools for network teams and AI systems that assess change risk before deployment.

Forward Predict runs proposed changes against a digital twin of a production network, allowing organisations to check the likely effect on connectivity, security and compliance before any command is executed in production.

The product has been in beta since May, with customers in financial services, media and technology using it to assess routing, firewall and policy changes, identify unnecessary work and reduce manual checking.

One named user is IG Group. According to Forward, the financial services company expanded its use of the product from validating BGP and routing changes to firewall rule workflows.

"We've been using Forward Predict for a few months now and continue to find new use cases," said Steve Bamford, Senior Network Engineer at IG Group.

"We extended it from validating our BGP and routing changes to our firewall rule workflow, where automated path checks now catch requests for rules that aren't even necessary before they're implemented. That's cut our firewall change delivery time by roughly 95 percent and is saving our engineers more than 30 hours of manual work every month," Bamford said.

Forward also said one unnamed global company expects the product to generate more than USD $3 million in annual benefits through lower manual effort and better change execution.

Change testing

Network changes can create several operational risks, including outages, unexpected security exposure and compliance failures. Forward Predict is intended to identify those issues before deployment by modelling proposed changes to routing, access control lists, network address translation and firewall rules.

The software can also regression-test changes against required network behaviour and vendor-specific configuration logic. This is intended to give network engineers and automated systems evidence on whether a proposed change should proceed or be revised.

Nikhil Handigol, Chief AI Officer and Co-Founder of Forward, framed the launch around the use of AI in network operations, where software agents may propose changes faster than human teams can assess them.

"Know before you act is more than a tagline, it's the operating model Agentic NetOps requires," said Nikhil Handigol, Chief AI Officer and Co-Founder of Forward.

"AI agents can propose change at machine speed, but speed without proof of correctness just accelerates risk. Forward Predict verifies the impact of every proposed change with Forward's mathematically accurate digital twin before execution. That gives agents the evidence they need to operate safely, and gives engineers the confidence to move faster," Handigol said.

Wider platform

Alongside the product launch, Forward outlined updates across its wider Forward Enterprise platform. These include general availability of Zscaler ZPA and IBM Cloud coverage within the same network model, as well as support for additional devices from vendors including Cisco, Check Point and F5.

Forward also added new CIS-certified benchmarks, updated DISA STIG support and vulnerability intelligence with KEV and CWE classifications. It expanded access to network information through Forward AI, Network Maps and enhancements to its Network Query Engine.

Forward argues that AI-led network operations require a reliable model of the live environment rather than assumptions or outdated data. In this approach, proposed changes are tested against the actual state of the network, and failures return evidence showing why a change should not be applied.

The process is intended to sit between AI-generated intent and production execution, reducing the likelihood that an automated or human-proposed change causes disruption once deployed.

During the beta programme, customers across several sectors used the product to reduce change risk and shorten delivery times, with IG Group reporting that automated path checks identify unnecessary firewall rule requests before implementation.