Dec 2024 & Jan 2025
updated on Dec 5th 2024
On Beta : Dec 18th 2024
Live on App
Live on App
Live on App
Live on App
Live on App
Live on App
Knowledge
Live on App
Live on App
ive on App
Live on App
Defining Competition inside a portfolio
Live on App
How will the release work?
On 18th Dec, app will get a pop up when someone logs in announcing the availability of Beta. The user will have the option to switch to Beta.
The second release on Beta will be on 30th Dec 2024 where new features as mention in the table above will be added to Beta.
On Jan 15th 2025, app.consumr.ai will have the new interface, flow and features. The user will have an option to switch back to classic if needed. But the primary look of app will be the new interface.
March 1st 2025, Classic will be deprecated and the new interface will be permanent.
Details of the Release
Guided Flow for AI Twin
The most important aspect of creating an AI twin is to incorporate the appropriate persona and memories. The output of an AI Twin is highly dependent on the quality of data input. Therefore, it is essential to add the correct persona and memory to an AI twin. Furthermore, achieving the right persona and memory requires selecting the proper inputs by working backwards.
To create an asset that functions as a persona or memory, one must link the conversion stage of the AI twin and select inputs accordingly. Based on observing user interactions with AI twins and their creation processes, it is clear that new or intermediate users often struggle to plan ahead when generating reports.
The guided flow for AI twin is an intelligent wizard designed to understand brand and category details from the portfolio. It identifies the conversion stage of the AI twin and searches for the most relevant assets for this stage and portfolio. If one or more assets are missing, it prompts the user to create a new asset and uses an intelligence engine to analyze and recommend the best inputs needed to develop an asset. This guided flow will be available in the managed section through a form and also via an agent within a conversation.
Meetings and Conversations
In the past, we developed various types of insights to unlock and simplify different data sets through visual displays, aiding users in creating narratives around a brand. Over time, however, we realized that users often struggled to interpret the output as intended. This prompted us to explore ways to simplify these datasets through conversational interfaces.
This innovation led to the creation of AI twins, which enable users to interact with personas possessing specific memories. Additionally, we discovered that bringing together multiple AI twins to discuss a particular topic could uncover deeper insights.
We envisioned this concept within a chat platform, similar to Microsoft Teams, Google Chat, or Slack. Consequently, we introduced the idea of conversations and meetings. A conversation can occur between a user and an AI twin or between a user and an agent.
A conversation can occur between the user and a group of AI twins and/or agents. When it involves two parties, including the user, it is referred to as a conversation. However, if it includes a group of AI twins, agents, and the user, it is termed a group conversation. In both scenarios, all conversations are unstructured and ad-hoc.
Conversely, a meeting is a structured form of planned interaction among AI twins and agents, facilitated by a copilot. The user only participates in initiating or planning the meeting or analyzing and interacting with the meeting's outcome once it has concluded. Structured meetings offer options like conducting custom meetings, focus groups, and brainstorming sessions.
There are three types of structured meetings:
Focus Groups: A group of AI Twins is assembled to answer a set of predefined questions. These responses are analyzed using specific models and presented to the user.
Brainstorming: A group of AI Twins is given a problem along with a method for discussion. The solutions discussed within a predefined framework are collected and presented on data cards.
Custom Meeting: A group of AI Twins is assembled to discuss a problem statement without any set structure. An objective-led problem statement must be provided, and the AI Twins will discuss the issue. The output will include Minutes of the Meeting, a conclusion, structured cards, and detailed minutes.
An AI twin can be considered as an external party that represents a group of users who either intend to use or have previously used your client’s product or a similar product within the same category. On the other hand, an agent is viewed as an internal team member who can provide specific insights regarding the data involved. Essentially, one can consult an AI twin for its perspective as a consumer, whereas an agent can be queried for detailed information and opinions related to the data, akin to consulting a knowledgeable colleague with a specific skill set. Each agent is designated according to their expertise, ensuring ease in identifying their particular strengths.
Retail and Professional Environments
We currently have three main reports: Behavior, Mentions, and Intent. Using these reports, we created campaigns, an optimization section, conversations, and AI Twins. However, we identified a need for more data dimensions for certain business models like retail and B2B.
Our existing data types are mostly online, focused on behaviors, intentions, and mentions. Therefore, we established a separate environment for online reports. Additionally, we integrated income tax data and household reports from the US into the retail environment, which also includes behavior, intent, and mentions data. The market report was moved into retail as it fits better there. Lastly, LinkedIn-based data of professionals will be used to create behavior reports in the business environment.
As a result, we now offer three environments with specific omniboxes tailored to different input types and options.
Including three different types of environments will enable users to generate reports from various perspectives and create retail twins, online twins, and business twins. These can be combined in group conversations or meetings to maximize the use of the diverse data each twin provides, with each twin representing hundreds of thousands of consumers from a market. Currently, household and income tax reports are only available for the US market.
Manage Section
The manage section was last updated about a year ago. Since then, many new features have been added, and access to them was still provided through the manage section, leading to several failure modes. The update to the manage section aims to address these issues. The following are the changes being implemented in the manage section:
Bundles & Constraints
Bundles
To provide complete flexibility, it will now be possible to build bundles within an organisation. A bundle can be named anything and will include parameters such as environments, markets, and platforms. Environments will include sub-parameters such as Online, Retail, Business, eTail, Campaigns, Conversations, and Meetings. It will be possible to create one or multiple bundles per organisation or sub-organisation.
Constraints
Constraints are rules associated with a bundle. They assist users in defining the usage limits for a particular bundle by an organization. A constraint includes rules for:
AI Twins: A PW user can set a constraint on the number of twins an organization can create, including the number of updateable twins and limits for audio call minutes and video call minutes.
Meetings: A PW user can define a limit to the number of structured and custom meetings an organization can use.
Users and Sub Organization: These two parameters will remain unchanged.
Design
To implement these changes, we had to modify the design. As a result, it is now more user-friendly. Additionally, the aim is to provide organizational admins with access to management features.
AI Twins Memory Update
Global Asset Table
Before discussing the purpose of the joint asset table, it is essential to understand what environments are. This update introduces new types of reports, tailored to different environments. Going forward, the environments will be classified as: Online, Retail, eTail, Business, Campaigns (formerly pre-flight), Conversations, and Meetings.
Campaigns do not have a history, aside from change logs, which will be managed separately. Conversely, Conversations and Meetings will be continuously managed by users, with conversation histories maintained in a chat interface format. Consequently, we have six types of assets created at specific points in time, such as Behavior, Intent, and Mentions reports within Online, Retail, eTail, and Business environments. Additionally, there will be market reports in the Retail section.
Furthermore, external files and nuggets saved from Conversations and Meetings will also be included. All these elements will be featured in this section.
Nugget Board
You will now be able to right-click on a conversation bubble or select a sentence or paragraph within a conversation bubble to save that piece as a Nugget. A Nugget is a piece of important information that a user can save for later use. This feature is similar to research questions, where the function was to ask a question and save the answer. Now, with a dedicated section for asking questions to an AI Twin or an agent, users can save useful information nuggets to use later in another conversation or incorporate them into a campaign.
The Nugget board will be a dedicated section within a conversation where you can view all the nuggets related to the portfolio you select in the Global Navigation at the top. All the nuggets will be arranged chronologically for easy recovery.
Using Nuggets to create creative campaigns
Nuggets are crucial pieces of information from conversations that can be used later in other discussions or campaigns. In the asset table (History) or Nugget board within a conversation, you can select nuggets and take action to incorporate them into a creative campaign. The rest of the creative campaigns remain unchanged.
Video and Audio calls with Twins
You have been able to chat with the AI Twin textually. We are now introducing the option for users to have a voice call or video call with the AI Twin. The process is straightforward; simply click on the call icon in the top right corner of the conversation screen to initiate a call.
In a voice call, both your speech and the AI Twin's responses are recorded and transcribed. However, playback of the call is not possible, though you will have access to the transcription.
Currently, this feature is not available to all users. ProfitWheelers will use it for demonstration purposes. Depending on user interest, we may introduce a per-minute rate card for our clients, similar to the pricing models used in the telecom industry.
Defining Competition inside a portfolio
Understanding the evolving consumer landscape is crucial for your brand, but recognizing and analyzing your competitors is equally important. Brands do not target all competitors simultaneously; it is essential to focus on direct competitors and clearly identify which ones to prioritize.
Enabling clients to predefine their brand’s competitors allows us to recommend relevant reports and gather intelligence not only for the brand but also for the identified competitors. You can specify your competition when creating or editing a portfolio. Additionally, we have introduced a competition recommender to assist users in this process.
Last updated