Google enhances debugging, performance in Opal, its low-code AI-based app builder

Google has rolled out new upgrades to Opal, its low-code AI-based application builder, aimed at significantly improving debugging capabilities and its overall performance, as part of its push to streamline AI app development for non-experts.

Opal, which was first introduced by Google’s Labs division in July in public beta, was billed as an experimental low-code tool for non-technical users that would help them build AI-based mini applications by chaining together prompts, tools, and models, using natural language and visual editing.

“Opal targets citizen developers, product owners, and focused business owners seeking faster time-to-market and measurable productivity gains in application development… Potentially, it may also be used by IT service providers to accelerate prototyping and delivery productivity for focused use cases,” said AS Yamohiadeen, practice director at Everest Group.

In order to improve debugging features inside Opal, Google said that it would allow users to run their workflows step-by-step in the visual editor or iterate on a specific step in the console panel.

This step-by-step testing and analysis will enable users to identify errors in real-time, thereby reducing the time required to complete debugging the application, Megan Li, senior product manager of Google Labs, wrote in a blog post.

Li said that “significant” changes were made to Opal to improve its performance, especially when it comes to using it to create an application.

One of these changes includes allowing parallel runs when a user is running complex workflows with multiple steps, Li wrote.

However, analysts warned that Opal might not be suited for complex enterprise use cases despite the upgrades to Opal and their potential in automating repetitive workflows.

“It’s well-suited for marketing automation, lead management, and reporting tied to Google services. However, it lacks the depth of enterprise-grade process orchestration, integration, and governance found in other, more sophisticated platforms designed for mission-critical process automation,” said Charlie Dai, principal analyst at Forrester.

Seconding Dai, Yamohiadeen pointed out that enterprise business application suites with their AI building tools will continue to hold an advantage over Opal due to entrenched data models, process ownership, native connectors, and governance frameworks.

“They also benefit from embedded distribution, identity and security integration, and compliance certifications, which reduce integration risk, accelerate adoption, and support predictable operations at scale,” Yamohiadeen said.

Google is also expanding the availability of Opal, and users will now have the choice to access the low-code AI-based application builder in 15 countries, including Canada, India, Japan, South Korea, Vietnam, Indonesia, Brazil, Singapore, Colombia, El Salvador, Costa Rica, Panamá, Honduras, Argentina, and Pakistan.

This expansion, according to Yamohiadeen, is Google’s strategic maneuver to extend its stack from the platform to the application layer, expanding adoption and aligning with a vision where AI composition becomes a primary interface for business solutions.

“If the roadmap lands maturity features and vertical packs, Opal can evolve from exploratory mini applications to durable enterprise workflows with sustained productivity impact,” Yamohiadeen said.

Conceptually, Opal’s closest rival is AWS’ PartyRock, which the hyperscaler had launched in November 2023. Other offerings that Opal can be compared to include the likes of Microsoft Copilot Studio, Salesforce Einstein 1 Studio, ServiceNow Creator Workflows and AI Agent Studio, Oracle’s toolchain, and adjacent builders such as Bubble, Make, and n8n, Yamohiadeen said.