AI in CATI: the case for augmentation, not replacement
In CATI, the value of the method depends on more than asking questions by phone. It depends on reaching the right participant, validating their relevance and maintaining a credible person-to-person interaction. AI can help make that process faster and more efficient, but it should support the interviewer, not take their place.
Simon Glanville
Managing Director
8 min read
Conducting market research interviews by phone, known as computer-assisted telephone interviewing, or CATI, has long been considered one of the most controllable, high-integrity methods available to reach business professionals and other expert audiences, where ‘who’ you reach is as important as ‘what’ they say.
Through CATI, a trained interviewer can engage with, recruit, and validate an appropriate participant, while the scripted interview itself allows for maximum information to be collected and quality-controlled via visible, accessible response data and audio recordings.
The process is quality-led and, up to now, has required human involvement at every stage, with the associated cost and time implications. The use of AI has the potential to reduce the human hours and costs involved, while adding efficiency and speed to this process.
The use of AI, however, should always be considered as a way to assist the human-led process, not simply replace skilled individuals.
The core of interviewing is an essentially people-led and participant-led methodology, which benefits from, and is indeed only possible by maintaining, a person-to-person approach. CATI remains one of the strongest methods for reaching hard-to-reach B2B audiences.
The question is no longer whether AI belongs in CATI, but where it adds value without weakening trust or interview quality. For senior B2B audiences, full AI replacement is not the priority. The bigger opportunity is human-led, AI-assisted CATI: better sample validation before fieldwork, smarter support during interviews, and more targeted quality control afterward.
It can be useful to consider how AI can assist in the CATI process before, during, and after the interview.
After the interview
After the call, the value of AI becomes especially clear. Every CATI programme produces hours of audio, thousands of open-ends and a significant amount of quality control work. AI can accelerate transcript-based coding, standardise checks for inconsistency and support faster interim reporting, without changing the fact that a human conducted a credible interview with a senior audience.
Random quality checks can be supplemented by AI-led prioritisation of interviews most likely to need review. Flags related to speed of interview and quality of responses can be incorporated into the selection of interviews for quality control.
Keeping CATI human-led
The examples above are possible today through diligent use of AI. There will be more refined and targeted inputs that will develop as AI, and how to implement the tool, become more streamlined and focused. There are advantages in terms of the prioritisation and selection of audiences, in assisting with setting up and testing, aiding an interviewer ‘in the moment’, and in quality control and data checking.
However, CATI interviewing is essentially a human-to-human interaction and will continue to require skilled interviewer outreach, reassurance and involvement to maintain the rhythm of a well-conducted interview, particularly for senior-level expert research.
CATI interviewing is essentially a human-to-human interaction and will continue to require skilled interviewer outreach, reassurance and involvement.
Simon Glanville Managing Director, RONIN International
Need high-quality CATI research for specialist B2B audiences?
RONIN combines experienced interviewer teams, quality-controlled CATI workflows and AI-supported processes to reach the right respondents and deliver reliable B2B insight.
Explore our CATI capabilities