Meta launches standalone AI app as Muse Spark 1.1 powers task automation

Meta's new standalone AI application, powered by Muse Spark 1.1, marks the company's most direct challenge yet to OpenAI and Google in the consumer AI assistant market.

Meta Platforms on July 24 began rolling out a standalone Meta AI application and enhanced Meta.ai features in select markets, using the new Muse Spark 1.1 model to handle planning, email integration and presentation creation.

"This is a significant step toward making AI assistants truly autonomous — handling multi-step tasks rather than just answering questions," Ahmad Al-Dahle, Meta's vice president of generative AI, said in a statement.

The new capabilities include connecting to email and calendar applications, creating slide decks and executing automated workflows on behalf of users. The rollout will expand to additional countries and platforms including WhatsApp in the coming weeks, the company said. Meta's total 2026 capital expenditure is projected at $125 billion to $145 billion, with AI infrastructure representing a growing share.

The launch positions Meta against OpenAI's ChatGPT, which already offers task automation through its GPT-4o model, and Google's Gemini, which integrates across Workspace. Meta's advantage lies in its 3.2 billion daily active users across Facebook, Instagram and WhatsApp — a distribution network no competitor can match.

Muse Spark 1.1, the foundation model powering the new features, represents Meta's latest push to close the capability gap with frontier labs. The model can break down complex requests into sequential steps, access external data sources through API connections and generate formatted outputs such as presentations and documents. Meta did not disclose benchmark scores for Muse Spark 1.1 against comparable models from OpenAI or Google.

The standalone app strategy marks a departure from Meta's previous approach of embedding AI features directly into its existing social platforms. By offering a dedicated application, Meta can compete for users who want a general-purpose AI assistant without engaging with Facebook or Instagram. The company plans to monetize the service through a subscription tier with higher usage limits, according to people familiar with the matter, though pricing has not yet been disclosed.

For investors, the question is whether Meta can convert its massive user base into paying AI customers. The company's Reality Labs division, which includes its AI research, posted $4.5 billion in operating losses in the first quarter. A successful consumer AI product could help offset those losses and justify the $125 billion-plus annual capex. Meta shares have gained 34% year to date, trading at 24 times forward earnings.

This article is for informational purposes only and does not constitute investment advice.