Our method: distinguishing facts, citing sources and calmly explaining what can really change for creators, independent and digital professionals.
IA
OpenAI launches a Data Agent capable of creating analytics and dashboards
OpenAI presents Data, a new ChatGPT Work agent that searches in natural language approved sources such as BigQuery, Snowflake, Databricks, MongoDB, Google Drive or SharePoint. It can search for the origin of a variation, produce an interactive dashboard, recommend actions and share results. Existing rights on rows, columns and tables remain in effect. The tool is provided in the ChatGPT Work plugin directory and must be installed or authorized by an administrator.
What it can change: Before connecting customer or financial data, define indicators, reference sources and access rights. Start with a simple dashboard — sales, reservations or conversion — and then manually check the calculations and the evidence.
Read the source : OpenAI →
IA
API Agents opens the cloud agent infrastructure of OpenAI to all developers
Available in public beta, API Agents allows you to create and run agents with a task, model, tools and environment defined by API. OpenAI manages the "harness" which maintains the context, orchestrates the tools and can coordinate sub-agents over long sessions. The environments can be hosted by OpenAI, operated on its own infrastructure or provided by a partner. Access does not add any specific costs: the models and tools consumed remain invoiced.
What it can change: For a product or small business, target a bounded and verifiable workflow — watch, content preparation, support or reporting — with execution log, cost limits, human validation before publication or payment, and a shutdown scenario in case of error.
Read the source : OpenAI →
Set-up
GPT‐Live‐1 brings simultaneous voice conversations to the API
OpenAI launches GPT‐Live‐1 in the API to build assistants able to listen and speak simultaneously, to better manage interruptions, silence and background noise, and to delegate reasoning or actions to a model in the background. The model allows to adjust tone, rhythm and style, supports telephone usage and offers more voices. It is now available at 0,05 dollar per minute for the frontal vocal layer, excluding the cost of the background model and tools.
What it can change: The most immediate uses are reception, prospect qualification, tutoring and support. Test first the understanding of accents, consent to recording, escalation to a human and the total cost per conversation.
Read the source : OpenAI →