2025 will be an exciting year for AI adoption across companies of every sort. From Salesforce promising that AI Agents will help your restaurant move your reservation inside when it’s going to rain, to predictions that AI can replace all of SaaS, we expect that the reality on the ground will be a bit more pragmatic. Companies are going to finally move out of the AI proof-of-concept phase into a period of practical applications and implementations with a focus on ROI.

Contextual’s fully managed AI Agent and Workflow Orchestration platform is ideally suited to help organizations design, develop, and deploy production-ready AI solutions across a range of functions, tasks, and processes. Our focus and commitment is to the following:

1. Companies should ‘Own Their AI’ - By this we mean organizations need to avoid creating a significant dependency on a single AI provider or model. The AI Foundation is evolving rapidly, solutions are often multi-model (meaning they require multiple steps calling discrete AI functions), and different AI functions are better at different things. 
2. Not Everything is an AI Agent - AI Agents are all the rage, but in many cases simple chat completion and ‘traditional software’ get you 100% of the AI value you seek with far less complexity and risk of autonomous systems ‘going rogue’.
3. Human Tasks First - If you have a task within the organization that requires significant human oversight or input, this is an ideal place to start exploring AI solutions. This can begin with enrichment first and then expand to automation over time.

With these truths in mind, we continue to innovate on the Contextual platform to deliver the most critical scope of features and functionality required to dramatically accelerate your successful adoption of AI technologies. We encourage you to hop in and get going to explore these critical new features.

Custom domains

Contextual’s HTTP Agents provide a powerful ‘front-end’ that can deliver deeply functional API endpoints or complete web-based applications for your human-AI interface. To enable personalization and customer or user-facing branding, Contextual now supports Custom Domains on an HTTP Agent end point. So instead of your agent serving on app on ‘service-name.tenant-name.my.contextual.io’ you can serve every endpoint from your API or your website from ‘yourdomain.com’. With this addition, Contextual is delivering everything you need to build complete, end-to-end AI and SaaS applications on the platform.

Paid tenant subscription plans will find support for unlimited Custom Domains for HTTP Agents on the Agent configuration screen.

Check out the full documentation for custom domains. 

Agent live logging

Depending on the scope and complexity of your Contextual solution, our powerful logging infrastructure can generate a tremendous amount of insight about the performance, session characteristics, and potential errors within your system. From simple messages to complex session data, logging is crucial throughout the lifecycle of development for your system. Depending on your service tier, you have access to different log retention periods, which aid in ‘going back in time’ to determine how your system performed.

Sometimes, though, you want to see logs as they are happening, either to observe system performance in real-time, trap an error or a bug, or determine how to operate more efficiently. Contextual’s Live View for logging delivers on this goal. You can choose one or more Agents to view, determine if you want to observe sessions or messages, and then watch your logs stream in real time with the ability to pause at any moment you need to dig in deeper.

New agent logging levels

Depending on your development pattern or the complexity of the AI Solution you’ve built on Contextual, you may need to log a little or log a lot. During early development, more verbose logging can be helpful. You can observe how your system is performing, check data structures and identify where errors are cropping up. Later on, you may just want to capture warnings for unexpected events or errors if something problematic occurs.

Logging in Contextual is performed in your Flow via a ‘Log’ node. These nodes can be set to a level of debug, info, warn, and error and when triggered in the course of a Flow, log the information you choose. Log information shows up at the Agent (that is running the Flow) and in an aggregate Tenant Logs view. These logs can be searched, reviewed and viewed live in the Contextual dashboard.

During standard operation, Agents running Flows which include Log nodes will only output warn and error levels to your tenant logs. However, we have now made it easy to temporarily change the Log Level of your running Agents to also include debug and/or info messages with zero interruption to Agent operation. The Agent will revert to default Log Level after 60 minutes.

You can adjust the log level of an Agent that is currently running in one of a few ways:

• From the Agent Index Data Grid including an action to change the level for all Agents at once
• From the Agent Detail view, in the header that is shared across all tabs

‍Simply click the overflow menu (3 dots) item for running Agents in the Agents list view, and choose Change Log Level to change the Log Level for your agent for the next 60 minutes.

Remember, changes to the Agent Log Level will be in effect for 60 minutes from the change, then will revert to the default level of Warn, Error.

While there are many more features being added to the Contextual platform all the time, these three were were sharing more explicitly to make sure you had all of the power of our low-code, highly visual AI platform at your fingertips.

Want a demo of everything Contextual can do for you? Give us a shout.‍