While Nvidia publishes its NEMO microservices to integrate AI agents into corporate workflows, research has revealed that almost half of the companies see only minor gains in their investments in AI.
Nemo microservices are a set of tools, some already available, that developers can use to create AI agents capable of integrating into existing applications and services to automate tasks and manage the life cycle of agents to keep them up to date if necessary with the latest information.
“There are more than a billion knowledge workers in many industries, geographies and locations, and our opinion is that digital employees or AI agents will be able to help companies do more work in this variety of fields and scenarios,” said Joey Conway, NVIDIA Senior Director of Generative AI software to enter.
Nemo microservices are also included in the will NVIDIA AI ENTERPRISE SUITE developer tools.
Components include Conservative Nemo for corporate data collection, which is transmitted to Nemo CustomizerDescribed by Conway as a microservice which “takes the latest advanced training techniques and teaches models of new skills and new knowledge so that we can make sure that the models supplying agents remain up to date”.
The Nemo assessor is intended to verify that the food model the agent has really improved instead of regressing, while Nemo railings Try to keep the agent on the subject so that he works as expected and avoids security and safety traps.
Nvidia envisages these microservices working in a circular pipeline, taking new data and user comments, using this to improve the AI model, then redeployed it. Nvidia calls it as a “data steering wheel”, although we cannot prevent ourselves from feeling that it is overlooked by a real steering wheel.
Visualization of this complete “data steering wheel” with an avatar Jensen Huang frightening in the center – Click to enlarge
Conway described Nemo microservices as “essentially as a docker container”. The orchestration rests on Kubernetes, with additional features such as Kubernetes operators to help you.
“We have software today to help preparing and preserving data. There will be many more,” he said.
Nvidia claims a large software support for its new AI toolbox, including corporate platforms such as SAP, ServiceNow and AMDOCs; Software batteries like Datarobot and Dataiku; More other tools such as Datastax and Cloudera. He also supports the models of Google, Meta, Microsoft, Mistral AI and Nvidia himself.
The examples where Nemo microservices are already put to work include AMDOCs, which work on three types of agents for its customer operators of telecommunications, said Nvidia.
These include a billing agent, a sales agent and a network agent. The billing agent focuses on resolution of queries, while the sales agent works on personalized offers and customer engagement as part of the agreements. The network agent will analyze newspapers and network information between geographic regions and countries to proactively identify service problems.
Developers can download Nemo Microservices from the Nvidia NGC catalog, or deploy it as part of NVIDIA AI Enterprise Suite.
Research today is also published in the United Kingdom which claims that companies spend an average of £ 321,000 ($ 427,000) on the AI in order to improve the customer experience, although many do not yet see significant gains. He found that 44% of business leaders said that AI has so far had only a slight improvement.
Despite this, almost all the respondents (93%) said that their investment in AI had offered a good return on investment (king).
The research was ordered by Storyblok, a supplier of CMS software for marketing specialists and developers, who have declared that companies must look beyond the implementations at the surface level and integrate AI in a way that stimulates a significant transformation.
He revealed that the most popular use cases for AI among British business leaders are the creation of website content, customer service, marketing analysis, translation services and marketing content. ®