23.9 C
Tuesday, May 28, 2024

LILT Unveils AI Controls and Multimodal Translation

Must read

Self-Service Management For Models, Modals, and Data Enables Enterprise Security and Control

LILT, the leading AI solution for enterprise translation and content creation, announced the company’s newest enterprise AI deployment and control features at its fourth-ever AI Day, including LLM orchestration, multimodal translation capabilities, and content generation prompt templates for teams and use cases. 

“As organizations move beyond AI pilots into production, scalable enterprise control becomes a non-negotiable,” said LILT CEO Spence Green. “Our newest features enable this for our customers, with full enterprise control of proprietary data, models, and workflows across models and business systems.”

During the 90-minute live virtual event, LILT announced several new features for enterprise control and scale, including:

Multimodal Translation Capabilities
Through translation capabilities directly within the LILT platform, users can now upload multimodal file types, including image, audio, PDF, and video, in a source language and immediately submit it for translation into a target language. The new UI also increases user flexibility for text translations, allowing users to select their preferred Verified or Instant translation for a text file. Expanding translation file types enables businesses to scale LILT to broader teams with more expansive use cases across media, images, social content, and more.

Enterprise AI Controls
A new broad set of enterprise controls allows organizations to define when and how their employees use AI, all from directly within the LILT platform. These controls enable authorized users to securely control the deployment of AI across the organization, including:

Workflow Control and Design
Authorized users can now determine and select distinct AI- and human-powered workflow combinations by content and Connector type, including AI + Review, Verified Translation, Instant Translation, and Customer Review, to ensure output alignment to business goals and standards. 

Connector Implementation and Oversight
Authorized users can now activate and manage Connector setup, monitoring, and activity in a no code application right from within LILT, empowering teams to deploy AI with fewer technical resources.

Data and Linguistic Asset Management
Authorized users can now manage, curate, import, export, and refine all data sources in LILT including translation memories, termbases, and glossaries, in order to control the quality of linguistic assets.

LLM Selection and Monitoring
Authorized users can now enable and configure integrations with third-party LLMs directly within LILT, selecting and overseeing LLM deployment and orchestration across their business, facilitating the deployment of different LLMs for diverse business use cases and quality needs.

Model Fine Tuning
Authorized users can now directly control and fine-tune cadence and direct the training sources for LILT and third-party LLMs from within the LILT platform, enabling tighter data control and bias oversight. 

Content Generation Prompt Templates
Through newly released predefined prompt templates, teams can now immediately generate content in any language for a diverse set of use cases, including:

  • Marketing: Newsletters and emails, social media, press releases, product comparisons, media pitches, blog posts, and longform
  • Support: FAQ responses, user guides, and customer support chat
  • Sales: Customer emails, presentations, and sales pitches
  • Legal: Terms of service, privacy policies, contract and GDPR clauses, and legal statements
  • HR: Training courses, job descriptions, and employee onboarding content

This new set of use case prompts enables businesses to activate users around the world across numerous teams and needs, all leveraging the same fine-tuned LLMs and historical data sets.

As organizations scale AI deployment, LILT heard and acknowledged its customers’ need for ability to manage that deployment in a controlled way, with defined user roles and access and tight data security and data selection for model training. As the world goes digital, customer demand for video translations has grown exponentially, and LILT is excited to meet those needs.

More articles

Latest news