Has Cvent implemented any AI features in their products?

Yes, Cvent has implemented CventIQ.

What is CventIQ?

CventIQ is a collection of powerful artificial intelligence capabilities embedded across the Cvent platform. It combines powerful AI capabilities with our unmatched industry expertise, proprietary insights and analytics, smart platform, and intelligent infrastructure to help marketers, event planners, and hospitality professionals move faster and smarter, collaborate better, and deliver more engaging, high-return experiences.

What powers CventIQ features?

CventIQ features are powered by third party LLMs such as Azure Open AI and include GPT4o, GPT4o-mini. However, future enhancements will support models like Anthropic Claude & AWS Bedrock.

Do you develop any LLMs in house?

No. Cvent leverages third party LLMs such as Azure Open AI and include GPT4o, GPT4o-mini.

How does the AI system ensure security?

Cvent ensures security of our AI system through a comprehensive, multi-layered approach outlined in the Cvent AI Security Standard and the Artificial Intelligence Development Policy based on industry standards such as SOC 2, ISO 27001 and ISO 27701. See https://www.cvent.com/en/security

Is human oversight enabled in the AI system?

When reasonable, Cvent provides human-in-the-loop functionalities where customers can review the output of a generative AI model before publishing/saving the output. In some cases, human-in-the-loop may not be reasonable such as Attendee Chatbots where attendees would interact with Chatbots to learn more about an event without any human review.

How is AI trained and what data is used for training?

The LLMs used in CventIQ are trained by our third-party providers. Cvent does not train any LLMs. We evaluate the third party’s Security controls around the development and operation of the LLMs provided prior to onboarding. It is ensured during the evaluation that no Cvent provided data is retained and no Cvent provided data is used by the third party for training the LLM. We rely on their expertise to ensure the models are robust and effective for our application needs.

How does the AI functionality handle biases?

Cvent does not have any input into how the models are trained. The models trained by OpenAI, Anthropic, and others have some built-in violence, hate speech, prompt breakout, and other protections, as evaluated and validated by Cvent’s security assessment of the third party’s security practices prior to onboarding the LLM. Cvent does not currently provide any additional processing to aid in bias detection. See https://openai.com/index/gpt-4o-system-card/

How do you ensure your AI complies with legal and ethical standards?

Our third-party providers adhere to industry regulations and ethical guidelines, and Cvent implements a comprehensive Artificial Intelligence Development Policy that establishes a secure, responsible and ethical framework for AI development and use.

What is the data retention policy for Cvent IQ?

Cvent will retain logs of questions our users ask and the answers they receive for a period of thirty days for quality assurance purposes only.  The logs will be automatically deleted after 30 days.

Are customers able to opt out of AI functionalities?

Yes. Customers can turn off AI features or selectively enable/disable specific AI functionalities.

Is the AI system able to make automated decisions?

No. Cvent does not build any sensitive automated decision making into our products using the outputs of generative AI. Cvent customers however may use the output at their discretion with the proper opt-in consent from their attendees.

Is the AI system GDPR and CCPA compliant?

Yes. Cvent systems undergoes privacy assessments based on ISO 27701 and follows internal policies aligned with GDPR and CCPA requirements. Cvent has policies to prevent the entry of sensitive personal information (special category data under GDPR) and has opted out of Microsoft storing or logging prompts and completions to minimize unauthorized access to customer data.

What type of data can be inputted into the AI system?

CventIQ is built to enhance the event planning experience and processes event data such as event titles, event descriptions, session titles, exhibitor information. Customers are expressly prohibited from entering any sensitive personal information into the AI system.

Will our data be used to train AI models?

No, neither Cvent nor our LLM providers will use customer prompts nor responses to train the AI model.