A survey of 1,076 business-to-business (B2B) decision makers published today finds that more than half (51%) of buyers of software now begin researching their purchase using some type of artificial intelligence (AI) chatbot.
Conducted by the market research firm G2, the survey finds well over two-thirds of respondents (69%) also ultimately changed their initial preference away from one vendor to another as a result of guidance from an AI chatbot, with one-third acquiring software from a vendor they have not previously known.
According to the G2 survey, nearly three in four (71%) B2B software buyers rely on AI chatbots for software research, compared to 60% seven months ago. Only 3% of respondents said AI chatbots haven’t meaningfully changed their research habits. Half (53%) of B2B software buyers feel that research done with an AI chatbot is more productive than traditional search, up from 36% seven months ago.
Four out of five respondents also noted that AI chatbots accelerated their purchasing decision, with 83% reporting they now feel more confident in their final choice. A full 85% said they think more highly of a software vendor when an AI chatbot mentions them in a recommendation.
However, 64% also note they have encountered inaccurate AI chatbot recommendations often or very often, with nearly half (45%) noting that citations from software review sites are the most confidence-inspiring signal in an AI-generated response.
Comparing vendor strengths and weaknesses is the #1 use case for AI chatbots in software research (41%), with an equal percentage using Deep Research tools regularly for software evaluations. The ChatGPT AI chatbot from OpenAI remains the dominant tool for B2B software research (63%) and nearly two-thirds spend six or more hours per week using AI chatbots for work.
The impact AI chatbots are having on purchasing is clearly now far reaching, says G2 CIO Tim Sanders. However, when it comes to research, not all AI chatbots are created equal, he adds. ChatGPT, for example, has been clearly optimized for research when compared to the Claude AI chatbot from Anthropic, he notes.
It’s not clear what ultimate impact AI chatbots will have on B2B purchases, but the days of trying to parse a list of individual search results based on keywords as a starting point are rapidly coming to a close, says Sanders. That shift is making it easier for many smaller providers of software to become better known, he adds. “They can break through from obscurity,” says Sanders.
Exactly which AI chatbot might prevail when it comes to researching purchases has yet to be determined. While OpenAI has a clear lead in terms of capabilities, AI chatbots from Google and Microsoft that are more easily accessible from a productivity application are often more likely to be close at hand, says Sanders. If and when those vendors improve their research capabilities, many purchasers may opt for different tools.
Regardless of approach, this may only be the tip of the AI iceberg when it comes to transforming how purchasing decisions are made. The only thing to be determined now is how a software vendor can best raise their visibility in an era where, thanks to the rise of AI coding tools, there is likely to be more competition than ever.
