OpenAI releases personalization and customization features for ChatGPT including custom instructions, memory, and custom GPTs
Feature update ● Confirmed 92% confidence first seen
OpenAI launched several new features enabling users to personalize ChatGPT through custom instructions, memory capabilities that persist across conversations, and the ability to create custom GPTs—specialized AI assistants configured for specific tasks without coding. These tools allow users to store preferences and context, create reusable workflows, and deploy standardized assistants for tasks like customer support and content generation.
Decision brief
- What changed
- OpenAI released a set of personalization and customization features for ChatGPT, including custom instructions, persistent memory across conversations, a 'skills' feature for reusable workflow templates, and custom GPTs—no-code AI assistants configurable with instructions, knowledge files, and external tool integrations.
- Why it matters
- These features lower the barrier for non-technical teams to build standardized, repeatable AI workflows (e.g., customer support, content generation, data analysis) without engineering resources, potentially accelerating internal AI adoption and shifting some automation work away from custom-built internal tools. Organizations that deploy custom GPTs at scale may see productivity gains but also introduce new questions about data handling, consistency of outputs, and governance over who can create and share these assistants.
- Evidence
- All three sources are OpenAI's own blog posts announcing the features, so coverage is entirely self-reported by the vendor with no independent third-party verification of adoption, performance, or real-world impact.
- What remains uncertain
- It is unclear how memory and stored user data are secured, retained, or governed, and whether custom GPTs introduce new data leakage or compliance risks when deployed across teams; adoption rates and actual productivity impact are unverified since all claims originate from OpenAI's own announcements.
- Monitor next
- Watch for independent reporting or enterprise case studies on custom GPT adoption, security incidents related to memory/data persistence, or OpenAI guidance on governance and access controls for organizational deployments.
Analytical support, not advice — assumptions and open questions stated above.