When you’re trying to understand your customers’ experience, product reviews are a great place to start. This qualitative feedback contains a wealth of knowledge, but the sheer volume of reviews is often overwhelming. It would take hours of time (that you and your team likely don’t have) to read and translate each piece of unstructured feedback into actionable insights – and more reviews are added every week. So, you skim to the best of your ability, resigning yourself to missing key insights.
No longer: Enter Viable’s GPT-3 natural language reports.
By leveraging OpenAI’s GPT-3 end-to-end, we’re able to digest and analyze each piece of qualitative data from hundreds of sources, and generate an easy-to-understand report that reads like a human spent hours crafting it.
To highlight the power of our software, we scraped 5,000+ Google Play, iOS App and Reddit comments about the Nest App. In no time, our AI analyzed thousands of pieces of feedback, and pulled the most salient themes and insights into an easy-to-digest report that goes way beyond your average sentiment analysis. But why tell you when we can show you.
Our technology breaks this report into three sections: Compliments, Complaints and Requests.* Unsurprisingly, the Compliments are flagged as “low” urgency for the company because they don't require anyone to address them. As you’ll see, the AI identified four key “Compliment” themes based on 822 data points that, in the case of Nest, validate what the app and its products enable the consumer to do well. Even though reviewers are talking about different physical Nest products – the thermostat, doorbell, security camera, etc. – our AI is able to differentiate between the various features and identify these key themes. It can even understand which product reviewers are referring to when customers use different terminology to talk about the exact same product (e.g. “thermostat” vs “temperature sensor” vs “app” vs “it”).
Boost customer satisfaction with precise insights
Surface the most urgent topics by telling our AI what matters to you.
The “Complaints” section has been flagged as the most urgent, and offers insights that are easy to overlook when you’re strapped for resources. Most notably, it calls out that the Nest thermostat is unreliable, and that the app’s notifications are sometimes inaccurate or delayed. The report allows you to click into each complaint to see the data supporting each theme – when we do this, we see that a user relayed that their Nest App did not notify them when intruders entered their home. When your consumer relies on your product to keep them and their valuables safe, this is a huge problem.
In other reviews, users shared that their notifications were delayed, and that whatever activity had tripped their security camera often ended before the app loaded the live stream. Understandably, this is leaving users feeling that the Nest App and its associated physical products are not living up to their expectations. Through the hundreds of reviews, our AI identified this issue as “Severe”, and provided reasons why, including quotes directly from consumers in its synopsis. To discover this issue, let alone have it explained at this fidelity, would take a consumer insights team days. Through Viable, you can have access to this critical feedback quickly and monitor any changes with a weekly report emailed directly to your inbox.
As we move on to the “Requests” section of the report, our AI found that a large volume of users want more control over their physical Nest products via the app.
While customers enjoy the available features, they’d like to have increased functionality, like custom settings for different sensors, control over the humidity in their homes, and more specific times for scheduling their devices. In the report's metadata table, we found that the majority of the people making these requests were based in Georgia and were operating on Android. These details could help Google identify hidden connections between their quantitative and qualitative data that would easily be missed by human analysis. And this capability isn’t limited to location or operating system – the AI can use any meta tags available within your data set to identify which users are providing any given feedback.
As if these insights weren't enough, the report also has a section where you can ask follow up questions to our AI directly. It will reply in natural language and you’ll be surprised by just how human it sounds.
Without any human interpretation, our technology was able to surface and highlight all of these insights directly from product reviews across three sources. This is only one example of our software's ability to analyze qualitative customer feedback. Our AI can analyze everything from Zendesk tickets to employee surveys, and even sales call transcripts to present the most important details for you in no time.
Curious to see the Nest report in full or run a report with your own data?
SCHEDULE A DEMO
*Our full reporting capabilities also include a fourth section called Questions but it’s typically more relevant to customer support data such as Zendesk tickets. Since product reviews typically do not contain many questions, the AI did not generate a Questions section.
%20(10).png)
%20(9).png)
%20(8).png)