Conveyor Uses GPT-3 to Bring Automated Questionnaire Completion Out of the Stone Ages

The company, on a mission to end security questionnaires, is the first to use GPT-3 to apply generative answering to completing vendor questionnaires at scale.

Conveyor, the leading end-to-end customer trust platform for security and compliance teams, announces the release of its Questionnaire Eliminator – the world’s first generative answering product, powered by OpenAI’s GPT-3.

The increase in demand for vendor security assessments before purchasing B2B products means a security questionnaire is completed with every sales deal globally. This presents a scaling challenge for existing automation tools. Providing a response in the desired format and wording (which varies by customer) has always required manual human intervention. However, new GPT technology has, in just months, enabled dramatic advances by generating precise answers for novel, custom questions with speed.

“The ‘state of the art’ when it comes to AI is changing quickly, especially with new LLM architecture that’s unlocking new search, retrieval and generation pipelines that can probably surpass human accuracy this year with accuracy rates of 96-99%,” says Chas Ballew, CEO/Founder of Conveyor. “We’re really excited to be first in this space to unlock generative answering and solve the pain of custom questionnaires for security and sales teams.”

Marketing Technology News: MarTech Interview with Ian Liddicoat, CTO and Head of Data Science – Adludio

Unlike existing tools on the market that provide ‘top match’ answers a human still has to re-write, Conveyor uses generative AI to formulate tailored, accurate answers to the specific questions asked, decreasing time to review by 89%.

Now, entire questionnaires are completed in minutes. On average, 84% of answers don’t need edits and the time to review each question is significantly reduced from 5 minutes to 32 seconds. The ultimate value is not only speed but generating senior analyst-level accuracy to bring teams closer to having their software function as a human reviewer.

“Our hope is that one day, vendor risk assessments will be done instantaneously. This new product is a step towards that goal,” says Ballew. “Generative answering is going to change the entire landscape of cybersecurity assessments in 2023 and we’re hard at work building what’s next.”

Marketing Technology News: Marketing Teams Need A Digitally-Inspired Collaborative Buzz To Thrive In A Hybrid World