AI Agent Action: Choose Anthropic or Google Models

TL;DR

The AI agent action within HighLevel workflows now supports Anthropic Claude and Google models, alongside a redesigned model picker that includes effort level settings. This is a foundational update for AI usage in workflows. No immediate action is required unless you are already leveraging the AI agent action.

Who this affects

HighLevel has not confirmed who this specifically affects. Agencies actively using or planning to use the AI agent action in workflows are the primary audience. Those who do not use AI in their automations can safely disregard this update for now.

What changed

Several key changes have been implemented:

• The AI agent action now supports Anthropic Claude models.
• The AI agent action now supports Google models.
• The model picker in the AI agent action has been redesigned.
• This redesigned model picker now allows choosing a specific model provider (Anthropic, Google).
• It also allows setting an effort level: low, medium, or high.
• A higher effort level means deeper reasoning for tasks.
• A lower effort level is intended to keep things fast for simpler tasks. HighLevel has not confirmed any limits, plan gating, or rollout status for these changes.

Where to find it

This update is found within the AI agent action. To access it, navigate to your HighLevel account, then go to 'Automations' and create or edit a workflow. When adding an action, select the 'AI agent' action. Within the settings for this action, you will find the redesigned model picker that allows you to choose between model providers and set the effort level.

How agencies will use this

Agencies can leverage this to fine-tune their AI-powered automations. For instance, an agency running a client's lead nurturing workflow might use the AI agent action to personalize follow-up emails based on lead behavior. With this update, they could choose a specific Google model for quick, transactional responses where speed is critical, setting the effort level to 'low'. Conversely, for more complex tasks like drafting nuanced social media posts or summarizing detailed customer service interactions within a workflow, the agency might opt for an Anthropic Claude model with a 'high' effort level to ensure deeper reasoning and more sophisticated output. This allows for greater control over the balance between speed, cost (implied by effort level and model choice), and quality of AI output directly within their HighLevel workflows, without needing external integrations for different AI providers.

Our take

This is a standard update. It provides more options for agencies already using or considering the AI agent action in HighLevel workflows. The ability to select between Anthropic and Google models, along with adjustable effort levels, offers greater flexibility in tailoring AI behavior to specific tasks. It's a foundational improvement to HighLevel's native AI capabilities, but it doesn't represent a seismic shift for most agencies this week. For those already integrating AI into their processes, it's a valuable refinement that warrants a closer look.