Petavue - Intelligent Assistant for Enterprise
Petavue - Intelligent Assistant for Enterprise
Petavue - Intelligent Assistant for Enterprise
Petavue - Intelligent Assistant for Enterprise
A conversational business intelligence assistant to help deliver every day value to teams including assisting them by importing, cleaning, and analyzing data to draw conclusions and make decisions.
A conversational business intelligence assistant to help deliver every day value to teams including assisting them by importing, cleaning, and analyzing data to draw conclusions and make decisions.
A conversational business intelligence assistant to help deliver every day value to teams including assisting them by importing, cleaning, and analyzing data to draw conclusions and make decisions.
Responsibilities
Responsibilities
Responsibilities
Responsibilities
Ideation, User Flow, Information Architecture, Wireframing, Prototyping, Visual Design, Iterations
Ideation, User Flow, Information Architecture, Wireframing, Prototyping, Visual Design, Iterations
Ideation, User Flow, Information Architecture, Wireframing, Prototyping, Visual Design, Iterations
Ideation, User Flow, Information Architecture, Wireframing, Prototyping, Visual Design, Iterations
Duration
Duration
Duration
Duration
April 2023 - March 2024
April 2023 - March 2024
April 2023 - March 2024
Petavue’s Team
Petavue’s Team
Petavue’s Team
Petavue’s Team
2 stakeholders, 1 product manager,
1 project manager, 2 frontend developers, 1 designer
2 stakeholders, 1 product manager, 1 project manager, 2 frontend developers, 1 designer
2 stakeholders,
1 product manager,
1 project manager, 2 frontend developers, 1 designer
2 stakeholders, 1 product manager, 1 project manager, 2 frontend developers, 1 designer
Petavue is a Natural Language Engine that can add everyday value to enterprise teams while giving users full control to speak their business language.
Petavue is a Natural Language Engine that can add everyday value to enterprise teams while giving users full control to speak their business language.
Petavue is a Natural Language Engine that can add everyday value to enterprise teams while giving users full control to speak their business language.
Petavue is a Natural Language Engine that can add everyday value to enterprise teams while giving users full control to speak their business language.
What was the problem we were trying to solve?
What was the problem we were trying to solve?
What was the problem we were trying to solve?
What was the problem we were trying to solve?
Complexity of SQL Querying For Data
Complexity of SQL Querying For Data
Complexity of SQL Querying For Data
Complexity of SQL Querying For Data
Extracting specific data from SQL tables is a labor-intensive and time-consuming task.
Requires a deep understanding of SQL and complex queries often involves thousands of lines of code.
Organisations rely heavily on developers or data analysts to perform this task.
Extracting specific data from SQL tables is a labor-intensive and time-consuming task.
Requires a deep understanding of SQL and complex queries often involves thousands of lines of code.
Organisations rely heavily on developers or data analysts to perform this task.
Extracting specific data from SQL tables is a labor-intensive and time-consuming task.
Requires a deep understanding of SQL and complex queries often involves thousands of lines of code.
Organisations rely heavily on developers or data analysts to perform this task.
Learning Curve of Tools for Data Insights
Learning Curve of Tools for Data Insights
Learning Curve of Tools for Data Insights
Learning Curve of Tools for Data Insights
Gaining insights from data using tools like Tableau presents another hurdle.
Learning to use such tools effectively can be time-consuming, even analysts may require several hours to create meaningful visualisations and reports.
Can hinder timely decision-making.
Gaining insights from data using tools like Tableau presents another hurdle.
Learning to use such tools effectively can be time-consuming, even analysts may require several hours to create meaningful visualisations and reports.
Can hinder timely decision-making.
Gaining insights from data using tools like Tableau presents another hurdle.
Learning to use such tools effectively can be time-consuming, even analysts may require several hours to create meaningful visualisations and reports.
Can hinder timely decision-making.
Addressing these challenges was crucial for organisations seeking to harness the power of data-driven insights efficiently and make informed decisions in a competitive landscape.
Addressing these challenges was crucial for organisations seeking to harness the power of data-driven insights efficiently and make informed decisions in a competitive landscape.
Addressing these challenges was crucial for organisations seeking to harness the power of data-driven insights efficiently and make informed decisions in a competitive landscape.
Petavue’s Goal
Petavue’s Goal
Petavue’s Goal
Develop a solution that streamlines data extraction process from databases
and simplify the creation of data visualisations.
Develop a solution that streamlines data extraction process from databases
and simplify the creation of data visualisations.
Develop a solution that streamlines data extraction process from databases
and simplify the creation of data visualisations.
Develop a solution that streamlines data extraction process from databases
and simplify the creation of data visualisations.
Reduce the reliance on specialised skills needed
and minimise the time required to derive insights from various sources of data.
Reduce the reliance on specialised skills needed and minimise the time required to derive insights from various sources of data.
Reduce the reliance on specialised skills needed and minimise the time required to derive insights from various sources of data.
Reduce the reliance on specialised skills needed and minimise the time required to derive insights from various sources of data.
Create a simple and intuitive interface to get elaborate, detailed, personalised and contextual insights from enterprise data.
Create a simple and intuitive interface
to get elaborate, personalised, detailed and contextual insights from enterprise data.
Create a simple and intuitive interface
to get elaborate, personalised, detailed and contextual insights from enterprise data.
Create a simple and intuitive interface
to get elaborate, personalised, detailed and contextual insights from enterprise data.
Solution
Solution
Solution
The system is driven by human language instructions.
Users can fetch information from most of the commonly-used systems and tools in a jiffy.
Users can ask questions over the information to get deeper insights.
There is a limited need for training, as users can choose either the grid view or guided view.
Grid view is for users who know the data they are working with, and want to modify it and extract more information from that data.
Guided view is for users who are more new to the system as the system will guide the user with related questions/options to select.
View detailed workflow executions and outputs.
View detailed workflow executions and outputs.
View detailed workflow executions and outputs.
View detailed workflow executions and outputs.
Users can view the system's detailed workflow in real time as it processes their query.
This transparency allows users to understand the system's thought process and identify any potential errors or areas of confusion.
It helps users improve their prompts for more accurate results.
Dashboard to view metrics and insights
Dashboard to view metrics and insights
Dashboard to view metrics and insights
Dashboard to view metrics and insights
Users can add any widget to the dashboard.
They can also decide which data or which metric to add to the dashboard.
Users can ask follow up questions on any specific insight.
Dashboard can be shared with employees to keep track of the data.
No-config system, driven by human language instructions.
No-config system, driven by human language instructions.
Fetch information from most of the commonly-used systems and tools in a jiffy.
Move beyond low-code systems and drag-drop editors.
With a limited need for training, your entire organisation, from leadership to individual reps, can unlock value almost immediately.
Users can fetch information from most of the commonly-used systems and tools in a jiffy.
Move beyond low-code systems and drag-drop editors.
With a limited need for training, the entire organisation, from leadership to individual reps, can unlock value almost immediately.
Who we designed for
Who we designed for
Who we designed for
Who we designed for
Team Leads
Team Leads
Users responsible for guiding a team during a particular initiative or towards a specific goal.
Users responsible for guiding a team during a particular initiative or towards a specific goal.
Teams
Teams
All users in an enterprise team depending on the data and insight requirement.
All users in an enterprise team depending on the data and insight requirement.
Business Users
Business Users
Business Users
Business Users
Non-technical users who access business accounts as part of their daily responsibilities.
Non-technical users who access business accounts as part of their daily responsibilities.
Non-technical users who access business accounts as part of their daily responsibilities.
Non-technical users who access business accounts as part of their daily responsibilities.
Data Analysts
Data Analysts
Data Analysts
Data Analysts
Users who clean, collect and analyze data to find solutions to different problems.
Users who clean, collect and analyze data to find solutions to different problems.
Users who clean, collect and analyze data to find solutions to different problems.
Users who clean, collect and analyze data to find solutions to different problems.
Team Leads
Users responsible for guiding a team during a particular initiative or towards a specific goal.
Teams
All users in an enterprise team depending on the data and insight requirement.
User Journey
User Journey
User Journey
User Journey
Information Architecture
Information Architecture
Information Architecture
Information Architecture
Features we wanted to include in the first iteration to test with our potential users:
AI Models: Users can connect different models like OpenAI, DeepInfra, Cohere etc. and configure them.
3rd party integration: Applications like Salesforce, Slack, Zendesk etc. can be integration as a lot of businesses store their data in those platforms.
Data Sources: Data can be uploaded in any format like API, document, texts, and 3rd party applications.
Users will have access to a list of configured apps they can run on the platform in the form of a chatbot
Users can create their own app by connecting their data sources and configuring their AI Model
Features we wanted to include in the first iteration to test with our potential users:
AI Models: Users can connect different models like OpenAI, DeepInfra, Cohere etc. and configure them.
3rd party integration: Applications like Salesforce, Slack, Zendesk etc. can be integration as a lot of businesses store their data in those platforms.
Data Sources: Data can be uploaded in any format like API, document, texts, and 3rd party applications.
Users will have access to a list of configured apps they can run on the platform in the form of a chatbot
Users can create their own app by connecting their data sources and configuring their AI Model
Features we wanted to include in the first iteration to test with our potential users:
AI Models: Users can connect different models like OpenAI, DeepInfra, Cohere etc. and configure them.
3rd party integration: Applications like Salesforce, Slack, Zendesk etc. can be integration as a lot of businesses store their data in those platforms.
Data Sources: Data can be uploaded in any format like API, document, texts, and 3rd party applications.
Users will have access to a list of configured apps they can run on the platform in the form of a chatbot
Users can create their own app by connecting their data sources and configuring their AI Model
Features we wanted to include in the first iteration to test with our potential users:
AI Models: Users can connect different models like OpenAI, DeepInfra, Cohere etc. and configure them.
3rd party integration: Applications like Salesforce, Slack, Zendesk etc. can be integration as a lot of businesses store their data in those platforms.
Data Sources: Data can be uploaded in any format like API, document, texts, and 3rd party applications.
Users will have access to a list of configured apps they can run on the platform in the form of a chatbot
Users can create their own app by connecting their data sources and configuring their AI Model
Mid-Fidelity Wireframing
Mid-Fidelity Wireframing
Mid-Fidelity Wireframing
Mid-Fidelity Wireframing
Minimum Viable Product
Minimum Viable Product
Minimum Viable Product
Minimum Viable Product
These were some of the screens from the MVP of the product. Moving forward, a lot of these designs changed, but this was a good starting point and gave us a clear picture on which direction to head.
These were some of the screens from the MVP of the product. Moving forward, a lot of these designs changed, but this was a good starting point and gave us a clear picture on which direction to head.
These were some of the screens from the MVP of the product. Moving forward, a lot of these designs changed, but this was a good starting point and gave us a clear picture on which direction to head.
Data Sources
Data Sources
Data Sources
Users could connect any data source with our platform.
They could upload various different sources of data like Documents, APIs, Call Recordings etc.
Users could connect any data source with our platform.
They could upload various different sources of data like Documents, APIs, Call Recordings etc.
Users could connect any data source with our platform.
They could upload various different sources of data like Documents, APIs, Call Recordings etc.
Apps
Apps
Apps
List of all the apps can be viewed from the apps screen.
Apps can be categorised into different groups depending on the domain.
Users can pick any app and run it. Each app would work like a chatbot.
List of all the apps can be viewed from the apps screen.
Apps can be categorised into different groups depending on the domain.
Users can pick any app and run it. Each app would work like a chatbot.
List of all the apps can be viewed from the apps screen.
Apps can be categorised into different groups depending on the domain.
Users can pick any app and run it. Each app would work like a chatbot.
Create new apps
Create new apps
Create new apps
Users can configure their own app and give it a specific name and select the category it comes under.
They can pick a few data sources which the chat application would access.
Users can train the system by giving a specific prompt and example questions.
Users can configure their own app and give it a specific name and select the category it comes under.
They can pick a few data sources which the chat application would access.
Users can train the system by giving a specific prompt and example questions.
Users can configure their own app and give it a specific name and select the category it comes under.
They can pick a few data sources which the chat application would access.
Users can train the system by giving a specific prompt and example questions.
Templates and Chat Screen
Templates and Chat Screen
Templates and Chat Screen
Inside each app, users can ask any question related to the field and the data sources.
The templates in the screen are system generated templates, which users can pick and run at any point.
Inside each app, users can ask any question related to the field and the data sources.
The templates in the screen are system generated templates, which users can pick and run at any point.
Inside each app, users can ask any question related to the field and the data sources.
The templates in the screen are system generated templates, which users can pick and run at any point.
Feedback From Users After Testing
Feedback From Users After Testing
Feedback From Users After Testing
Feedback From Users After Testing
Data Sources
Data Sources
“I don’t want to add specific data sources each time as it’s very time consuming”
“I don’t want to add specific data sources each time as it’s very time consuming”
“I don’t want to add specific data sources each time as it’s very time consuming”
“I don’t want to add specific data sources each time as it’s very time consuming”
“Is there a way the system can recognize these sources through the instructions we are giving?”
“Is there a way the system can recognize these sources through the instructions we are giving?”
“Is there a way the system can recognize these sources through the instructions we are giving?”
“Is there a way the system can recognize these sources through the instructions we are giving?”
“Is there a way the system can recognize these sources through the instructions we are giving?”
Templates
Templates
“I want to add a set of instructions to obtain data that is relevant and save it”
“I want to add a set of instructions to obtain data that is relevant and save it”
“I want to add a set of instructions to obtain data that is relevant and save it”
“I want to add a set of instructions to obtain data that is relevant and save it”
“I want to be able to give instructions and see the execution so I can understand what’s going wrong”
“I want to be able to give instructions and see the execution so I can understand what’s going wrong”
“I want to be able to give instructions and see the execution so I can understand what’s going wrong”
“I want to be able to give instructions and see the execution so I can understand what’s going wrong”
Run Their Own Questions
Run Their Own Questions
“Do we always have to create a full new application before running anything?”
“Do we always have to create a full new application before running anything?”
“Do we always have to create a full new application before running anything?”
“Do we always have to create a full new application before running anything?”
“Can’t we ask Petavue any question like ChatGPT but on our own data?”
“Can’t we ask Petavue any question like ChatGPT but on our own data?”
“Can’t we ask Petavue any question like ChatGPT but on our own data?”
“Can’t we ask Petavue any question like ChatGPT but on our own data?”
Dashboard
Dashboard
"We analyse our data to see the metric changes"
"We analyse our data to see the metric changes"
"We analyse our data to see the metric changes"
"We analyse our data to see the metric changes"
“Since the app that we are configuring with data is analyzing data anyway, why cant we have a dashboard overview of our metrics?"
“Since the app that we are configuring with data is analyzing data anyway, why cant we have a dashboard overview of our metrics?"
“Since the app that we are configuring with data is analyzing data anyway, why cant we have a dashboard overview of our metrics?"
“Since the app that we are configuring with data is analyzing data anyway, why cant we have a dashboard overview of our metrics?"
Some of the features that the users requested were given priority and built. The others were pushed to future versions due to time constraints.
Product - Version 1
Product - Version 1
Product - Version 1
Product - Version 1
Some of the features that the users requested were given priority and built. The others were pushed to future versions due to time constraints.
More testing and iterations
More testing and iterations
More testing and iterations
We released the first version of the product to a select group of users to gather initial feedback. After reviewing the input and going through several iterations, we identified key areas to focus on for improvement. While we received a wide range of suggestions, only a few were prioritized for further development.
We released the first version of the product to a select group of users to gather initial feedback. After reviewing the input and going through several iterations, we identified key areas to focus on for improvement. While we received a wide range of suggestions, only a few were prioritized for further development.
We released the first version of the product to a select group of users to gather initial feedback. After reviewing the input and going through several iterations, we identified key areas to focus on for improvement. While we received a wide range of suggestions, only a few were prioritized for further development.
Flowchart when the system is answering the questions
Problem:
Users struggled to understand the flowchart generated during template creation because it was too technical.
Solution:
Providing text-based process was more effective.
By simplifying the language and hiding technical jargon, we communicated the system's behind-the-scenes processes in clear, natural language that aligned with how users interacted with the system.
Templates
Problem:
During initial user interviews, users indicated they would use pre-built templates.
However, after the product launch, we observed low engagement with this feature.
Solution:
We decided to remove the templates section
Users action showed that they preferred auto-generated questions over fully built-out templates.
Data Dictionary
Data Dictionary
Data Dictionary
Problem:
The system struggled to accurately interpret business terms and jargon, resulting in errors and poor-quality responses.
Solution:
We implemented a data dictionary feature that allows users to define key metrics and train the system to better understand their specific business terminology, leading to more accurate and relevant results.
My Learnings
My Learnings
My Learnings
My Learnings
Adaptability
Adaptability
Adaptability
Adaptability
It's crucial to remain adaptable, as the final result often diverges significantly from the initial concept. I've learned not to grow overly attached to my designs, as continuous changes are inherent to the product development process.
It's crucial to remain adaptable, as the final result often diverges significantly from the initial concept. I've learned not to grow overly attached to my designs, as continuous changes are inherent to the product development process.
It's crucial to remain adaptable, as the final result often diverges significantly from the initial concept. I've learned not to grow overly attached to my designs, as continuous changes are inherent to the product development process.
It's crucial to remain adaptable, as the final result often diverges significantly from the initial concept. I've learned not to grow overly attached to my designs, as continuous changes are inherent to the product development process.
Early testing and market validation
Early testing and market validation
Early testing and market validation
Early testing and market validation
These are pivotal steps prior to product launch. Ensuring your product aligns with market needs can save time, resources, and enhance its chances of success.
These are pivotal steps prior to product launch. Ensuring your product aligns with market needs can save time, resources, and enhance its chances of success.
These are pivotal steps prior to product launch. Ensuring your product aligns with market needs can save time, resources, and enhance its chances of success.
These are pivotal steps prior to product launch. Ensuring your product aligns with market needs can save time, resources, and enhance its chances of success.
Communicate with stakeholders early
Communicate with stakeholders early
Communicate with stakeholders early
Communicate with stakeholders early
Effective communication with stakeholders ensures alignment and collective efforts in pursuit of an enhanced product.
Effective communication with stakeholders ensures alignment and collective efforts in pursuit of an enhanced product.
Effective communication with stakeholders ensures alignment and collective efforts in pursuit of an enhanced product.
Effective communication with stakeholders ensures alignment and collective efforts in pursuit of an enhanced product.
Design Process
Design Process
Design Process
Design Process
The process is not always linear. We went through multiple rounds of iteration and testing to reach the final product. We are still continuously trying to improving the product every single day.
The process is not always linear. We went through multiple rounds of iteration and testing to reach the final product. We are still continuously trying to improving the product every single day.
The process is not always linear. We went through multiple rounds of iteration and testing to reach the final product. We are still continuously trying to improving the product every single day.
The process is not always linear. We went through multiple rounds of iteration and testing to reach the final product. We are still continuously trying to improving the product every single day.
Thank you for reading!
Thank you for reading!
Thank you for reading!
Thank you for reading!
Other projects
Other projects
Other projects
Other projects
InSynk Studios
InSynk Studios
InSynk Studios
InSynk Studios
Website
Website
Website
Website
Grail - Verification Portal
Grail - Verification Portal
Grail - Verification Portal
Grail - Verification Portal
A trusted platform for decentralised application verifiers to access and authenticate customer's identity.
A trusted platform for decentralised application verifiers to access and authenticate customer's identity.
A trusted platform for decentralised application verifiers to access and authenticate customer's identity.
A trusted platform for decentralised application verifiers to access and authenticate customer's identity.
Freelance
Freelance
Freelance
Freelance
Mobile App
Mobile App
Mobile App
Mobile App
Website
Website
Website
Website
AG Fashion - Cloth Rental Application
AG Fashion - Cloth Rental Application
AG Fashion - Cloth Rental Application
AG Fashion - Cloth Rental Application
A peer-to-peer apparel rental marketplace that aims to promote sustainable practices.
A peer-to-peer apparel rental marketplace that aims to promote sustainable practices.
A peer-to-peer apparel rental marketplace that aims to promote sustainable practices.
A peer-to-peer apparel rental marketplace that aims to promote sustainable practices.
Let's
craft
something
extraordinary! 🤝
Let's
craft
something
extraordinary! 🤝
Let's
craft
something
extraordinary! 🤝
Let's
craft
something
extraordinary! 🤝
Let's
craft
something
extraordinary! 🤝