Automating the verification of records using machine learning

AI automation

NovaceneAI’s machine learning prediction model enables market research firm redirect staff from repetitive tasks to higher value-creating work.


A market research firm in the Educational Technology space needed to increase its capacity to verify, approve or reject, tens of thousands of records collected as part of their market insights services. The verification was carried out by a team of verifiers who struggled to keep up as new records became ready for verification. The firm needed to prioritize the records and process those most likely to be valid first.


NovaceneAI worked with the client to review the data and develop an approach to meet the project objectives. Following the planning stage, NovaceneAI developed a machine learning classifier that takes in signals from the records and predicts whether a record is valid or invalid. Following fine-tuning, the selected model achieved the performance requirements set by the business.

The prediction model built by NovaceneAI is saving us 200 hours of work per month. We can now use that time to work on a backlog of tasks.

Justin Ménard, Founder and CEO


The classification model was delivered through the NovaceneAI Platform and it is accessible by the client 24/7. Non-technical staff from the market research firm can now classify new records and retrain the model with new data at the click of a button.