OpenAI launches Dots to target high-spend enterprise AI markets

From reactive to proactive AI

OpenAI recently took its Dev Day roadshow to London to pitch Dots to the city's developers. Unlike previous tools like Codex, which wait for specific user instructions, Dots are designed to be proactive assistants. They are built to learn a user's specific goals and monitor work tools and inboxes to initiate tasks automatically.

The company is positioning Dots as its flagship agentic offering for the enterprise market. By moving toward autonomy, the industry believes these personal agents will drive significant value for consumers. This shift mirrors how coding agents previously provided major utility for enterprise developers.

While Dots are not yet available in Europe, OpenAI is currently working with regulators to change that. The company is focusing on high-end, high-spend users first to establish a foothold in this new market. They believe these professional use cases will see the most immediate value from proactive technology.

Managing the cost of compute

Running proactive agents is significantly more expensive than standard chat interfaces. Adam Fry, who runs ChatGPT’s consumer business, noted that Dots are on the compute-heavy side of the technology spectrum. This has led to concerns from industry leaders about the sustainability of high running costs relative to revenue.

To combat these expenses, OpenAI is utilizing a strategy of token efficiency. They are routing some Dots tasks to a smaller model named Terra, which is a version of GPT-6.1. This allows the company to handle routine tasks without always relying on the more powerful GPT-6 Astra model.

Romain Huet, OpenAI’s head of developer experience, stated that the Terra model can perform better than Astra in specific areas like writing. He noted that the smaller model can perform these tasks at a tenth of the price. The goal is to deliver the most amount of value per dollar invested in the Dots platform.

The privacy and consent trade-off

Because Dots must have deep access to personal data to be helpful, they raise substantial privacy concerns. Critics argue that these agents require a significant trade-off between convenience and security. There are also growing questions regarding how consent is handled when an AI reads messages sent by third parties.

Meredith Whittaker, president of Signal, highlighted that people who message a Dot user never agreed to have an AI ingest their information. She pointed out that logging contact lists and communication patterns could be serious for those in high-risk environments. These concerns are particularly sharp for individuals working on sensitive human rights issues.

OpenAI has suggested that users can build trust with these agents gradually by connecting tools one at a time. For example, a user might start with a calendar before adding Slack or Gmail. Leadership also mentioned they are looking for an easy way for users to pick and choose which specific connections the Dots assistant can use.

Competition in the agent market

OpenAI is entering a crowded field of proactive and assistant-based AI tools. Meta has already launched Muse, which has proven popular with consumers. Other notable competitors include Grok Bot from xAI and the Duet assistant offered by Anthropic.

The competition is intensifying as companies like Apple and Goldman Sachs begin to cap employee AI spending. These spending limits could impact the revenue growth of leading AI labs. Consequently, OpenAI is focusing on the 'pro' and enterprise segments where the willingness to pay is higher.

The company's strategy involves proving that Dots can accomplish more advanced tasks than competitors with simpler user interfaces. By nailing the enterprise use case first, OpenAI hopes to demonstrate a level of capability that justifies the higher costs and data permissions required for the technology.

Future of enterprise AI spend

The broader market for enterprise AI is seeing a massive surge in spending. Reports show that enterprises spent $177 billion on AI in the year ending September 2026. This represents more than five times the amount spent in the previous year, with growth accelerating as companies adopt agents.

While the number of companies paying for AI has only risen modestly, existing customers are spending significantly more. This spending is often blowing out original budgets, with many companies spending two and a half times what they initially planned. AI agents are identified as a primary driver of this increased spending.

A few major players continue to dominate the deployment landscape. Anthropic, OpenAI, and Google currently capture more than two-thirds of the spending for enterprise AI deployment. As agents become more common, the focus will shift toward how efficiently these companies can manage the underlying compute costs.

Next steps for Dots

OpenAI will continue to refine the balance between model power and cost efficiency. The company plans to expand Dots availability to the European market as regulatory hurdles are cleared. They will also continue to develop features that allow for more granular user control over data permissions.

For enterprise clients, the focus will remain on high-value tasks that require proactive intervention. This includes managing complex workflows across multiple software platforms like Slack and Gmail. The success of Dots will likely depend on whether users feel the automation is worth the privacy risks.

As the market matures, we can expect more competition from other labs trying to solve the compute-heavy nature of agents. OpenAI's ability to maintain its lead will depend on its success in making Dots both affordable for businesses and trustworthy for individual users.