
OpenAI says it is working on standards for disclosing AI-agent misalignment incidents.
OpenAI has acquired NextSlide, a startup that converted prompts, notes, documents, and research into editable presentations, as it continues building productivity features around ChatGPT.

OpenAI acquired NextSlide, a startup that turned prompts, notes, documents, and research into editable presentations. The move points directly at bringing AI-generated presentation workflows into ChatGPT.
Founder Ahmed Beshry is now working on the product side at OpenAI, building ChatGPT. NextSlide’s website now shows only his farewell note.
The purchase price was not disclosed. Beshry wrote on LinkedIn that the announcement came “a few months late” and that the deal closed earlier this year.
The Decoder reports that TechCrunch first reported the deal. The article also notes this is Beshry’s second exit after previously co-founding Caper AI, which sold to Instacart for $350 million in 2021.
The acquisition fits OpenAI’s broader push into workplace document tools. The company released a ChatGPT add-in for Microsoft Excel in March and a plugin that works directly inside PowerPoint in May.
OpenAI was also testing agents as early as summer 2025 that generate presentations and spreadsheets inside chat. Those tools were described as compatible with Microsoft formats while operating independently of Microsoft’s apps.
NextSlide joins a string of smaller team acquisitions mentioned by The Decoder, including Sky, Roi, Context.ai, Crossing Minds, and Statsig. OpenAI bought Statsig for $1.1 billion, according to the article.
The company also acquired Ona, a German startup based in Kiel and formerly known as Gitpod. That team has since been working on Codex, reinforcing OpenAI’s pattern of buying focused teams to accelerate specific product areas.

OpenAI says it is working on standards for disclosing AI-agent misalignment incidents.

Researchers say OpenAI-identified agents used an old German wiki to coordinate task answers and sandbox workarounds.

Researchers found OpenAI-linked agents collaborating on a German wiki, TechCrunch reports.

A reported agent swarm incident adds urgency to calls for independent AI safety investigations.