How G2 is Experimenting with AI to Supercharge Software Review Insights

At G2, we continue to challenge ourselves to embrace AI to work smarter and bring new innovations to market for software buyers and sellers alike. While AI-generated reviews are not accepted on G2 to maintain authenticity, AI can still have significant and positive contributions to the world of reviews in different ways

Our Product R&D team is always exploring new ways to glean additional, more helpful and streamlined insights from the 2.3 million+ trusted reviews on G2. This abundant content and data, brimming with invaluable customer feedback regarding the software products and B2B services they utilize, appeared to be an ideal candidate for AI utilization. Therefore, following our standard procedure for all Product R&D initiatives at G2, we conducted testing with the goal to learn.

Three new AI-powered features for streamlined researching reviews on G2.com

Over the past several years, we have been experimenting the best ways to apply AI to G2 reviews – from the review form itself to the insights produced from reviews submitted. Based on findings from G2’s 2023 Software Buyer Behavior Report, we know that when it comes to review sites, the ability to filter and compare products according to different review attributes are the most helpful to buyers. We’ve also seen the demand for easily digestible software attributes, both in qualitative feedback from our buyer audience and in keyword search volumes on popular search engines. 

With these buyer preferences in mind, here are the three new AI-powered features we’ve been testing which aim to help software buyers research software and services on G2.com through more robust and streamlined review insights:

• Pros and Cons: G2 Profiles and product categories now feature Pros and Cons to help buyers distill software reviews into critical insights, with the ability to filter reviews grouped by contextual themes. Leveraging AI, we conduct thorough sentiment analysis and provide advanced filtering capabilities, helping buyers learn more about the category and product  — while delving deeper into themes that are most important to their purchase decisions and filtering out the noise. 

Considering that G2 is a strong traffic magnet (80M annual visitors!), we want to support however users choose to discover software – whether through our funnel or via search engines. Regardless of their path, Pros and Cons gets buyers to the content they need for confident software purchases.

Ultimately, this new feature enables software buyers and sellers to view specific strengths and weaknesses of products relative to their category. Not only will this concise sentiment help buyers make better decisions, but it will also help vendors more quickly identify their top selling points and areas of opportunity across sales, support, marketing, service, and more. Pros and Cons are currently available across 35 categories, with plans to expand to additional categories soon. 

• Category Snippets: Category Snippets provide product overviews with pros and cons extracted from user reviews. Derived at the Product-level using AI, Snippets insights directly relate to the reviews of the product — compared to Pros and Cons insights (above feature) deriving at the Category-level, which are based on themes and paint a more broad picture of the product as it relates to multiple categories. 

Rolled out as a test across 80 category pages on October 17th  and appearing within the “What G2 Users Think” section of the category page, Category Snippets concisely distill themes from reviews, shedding light on product strengths and weaknesses so buyers can be informed and confident in their software purchase journey on G2. And with monthly updates from a six-month pool of reviews, vendors can easily spot the highs and lows in their reviews – staying informed on what customers appreciate and where improvements can be made at a glance. 

• Review Depth Checklist:  Powered by ChatGPT, the new Review Depth Checklist on the G2 review form encourages reviewers to discuss specific topics we know other buyers are searching for. Here’s how it works: as a reviewer is typing their response to the “love and hate” questions on the review form, an AI prompt analyzes the text behind the scenes to determine if any of the copy meets the checklist criteria. Then, in real-time, a topic in the checklist will be marked complete if the user covers that topic in their response. 

Using the authentic G2 reviews, the team refined and trained the prompt to fit our needs. This new feature is now live on all variations of the G2 review form and has replaced what previously was called “thought provokers.” Knowing that buyers searching on G2 are looking for lengthy, quality reviews that cater to their needs, this AI review depth checklist aims to produce the most helpful insights they’re seeking to inform purchasing decisions. 

Gathering insights on the AI software experience

While this next advancement isn’t necessarily  “powered by AI,” it’s a new capability that provides anyone visiting G2 unique insight into the impact of AI across the software landscape. Over the summer, we updated our G2 review form, adding questions that capture insights into the experience of buyers using generative AI capabilities in over 200 software categories.

While there are 21 AI software categories on G2, vendors are rapidly introducing AI features to solutions across the software landscape – from CRM to Restaurant Management, Retail Analytics, and so many in between. With the updated G2 review form, buyers can now rate their experience with AI features including Text Generation, Text Summarization, Text-to-Image, Text-to-Speech, Image-to-Text, Text-to-Video, Text-to-3D, and Text-to-Music when they leave a review for a product in the category. 

As more and more vendors add true AI capabilities to their products, G2’s Market Research team will be able to capture this satisfaction data – uncovering new trends and better advise software vendors on what buyers are seeking from AI functionality across software categories. Additionally, those visiting G2 will be able to view and compare AI features between products using the comparison tool on G2. 

The power of testing

Entrepreneurship is part of the overall ethos at G2, and in the Product R&D team we aspire to  introduce products that customers love. In fact, it’s one of our company’s values. As part of flexing our collective entrepreneurial spirit and always staying curious, I’m proud of our G2 Product R&D organization for bringing these new AI innovations to market within the realm of reviews. 

Stemming from team brainstorms and Hackathon events, these latest AI-powered features and ideas are currently in test mode. Some have only been rolled out in a limited number of select categories, for example. We do this to gather insights on what’s working and what’s not – and then adjust as needed before doing a full go-to-market roll out. This approach of testing and iterating has been key to Product R&D strategy and success at G2. 

As these latest advancements are tested, we aim to improve the experiences for reviewers, buyers, vendors, and anyone visiting G2.com seeking insights from reviews. And as the world’s largest and most trusted software marketplace, we’re continuing to gather more data and insights to serve as the ultimate resource on what’s happening across the software landscape. 

Stay tuned for additional updates – whether they’re new experiments we’re running or results from those already underway. And I hope you’ll be curious about G2, just as we continue to be!