Ask me a question
Let consumr.ai interpret the reports and textually answer your questions.
Last updated
Let consumr.ai interpret the reports and textually answer your questions.
Last updated
"Users don't know how to draw the right inferences from one or a series of reports built for a purpose"
We built Solve to ensure that a user does not have to struggle with using omnibox to the right effect. Instead they can type in a question and get the series of reports to run that can give them the answer.
However, once we execute the report there are few issues a user faces:
I have executed 7 reports how do I study all of them and draw the right conclusion?
Is there a view that can help me compare all 7 reports?
I am not analytical in nature and drawing inferences does not come easily. What do I do?
Out of 7 reports, how do I know which report gives us the best learning?
"To build an feature that can learn from the output of a series of reports, across modules and provide inferences based on the questions of the user"
Imagine a situation where you have 7 to 20 reports that you have executed and want to analyze it. Would you know what is the difference between, your A-B behavior report, A intersect B, and B-A? Would you be able to spot the nuances and summarize the difference? Maybe you can, but can our client? But I am sure that our clients sure have a lot of business related questions that he has but wouldn't know how to infer them, matching 5 to 7 reports side by side.
This feature is an extension of Know. The default view gives you access to the reports table. But when you want to get down to finding answers to your business related questions, Switch to "Ask me a question" (currently named "Questions")
The current version considers reports from the last 1 month and feeds it in to the intelligence engine. Here it analyzes the data output and answers it for the user within the context of the question.
For example in the screenshot below, we have fed in a series of reports for Metro shoes and then gone ahead and asked it a simple question: "What are the differences between the behaviorism and intentions of Metro consumers?"
The answer it provides is enlightening for the user
" Based on the reports, Metro consumers' behavior is mostly driven by interests in shopping, apparel & fashion, art & design, and drinks. They have a strong inclination towards online shopping, interior design, luxury goods, and jewellery. The age group of 25-44 is the most active. In terms of their intentions, the consumers are often searching for specific products like 'metro footwear', 'metro shoes near me', and 'metro sandals for women'. They also have questions about the brand itself, such as its origin and the founder. Comparisons with other brands like 'mochi shoes' also come up frequently."
While the aesthetic changes will be brought in soon enough, it gives us a quick summary in form of an answer for the question. However, for the curious minds if you want to see the rationale, you can click on the "See rationale" button below and it gives you the details including the report ids it got the information from.
Currently, please see this as a phase 1 of this feature, where we have brought in a series of reports by default in learning. This means in Phase 1 the user does not have a choice of selecting the reports they need to consider. This is like an initial version of a feature that we want the clients to use and based on the learnings we can decide the next course of development.
However, the following are the development plans in Phase 2 and 3, that can be combined or developed separately:
A capability to call in Projects in to learnings. Each project will consist of reports. A project can be built as per a study or a campaign.
An intelligence within the feature to identify missing reports required to answer and suggest it/them immediately. You can look at it as calling in Solve within Know
Suggest next course of action and what intelligence need the next action and what will it be
An intelligence to also run in Campaign, Adset and Ads related performance data to suggest the next action
Coming Soon
