Making Conversations Actionable with Google

Fig: Image and video shows a conversational dashboard solution built on Google, being interacted over Facebook App

Common Challenges

We see the following significant challenges with clients across industries for solutions that needs multiple human touch points to drive business outcomes.

Either they are too complex, cluttered, or monoliths without any versioning

Asks for deeper understanding over domain and functionalities

Longer time to analyze/ build/ market/ value for the use cases

Unavailable over common channels (social, collaboration, chatting, phone calls, etc.)

Limited reusability of existing functionalities and features constrains ROI

?Inflexible to frequent minor changes, diverse parameters, and filters on the fly

Inconsistent and non-standard responses missing guardrails and governance

User access management becomes utterly complex for large organizations

Here comes a sample use case solution built on Google for conversational dashboards.

For this use case, I asked myself a few very basic questions before constructing the solution. There can be many from you as well, and I would really appreciate them as part of your feedback. Maybe your inputs will help discover newer areas/ use cases and will be able to collaborate and co-create more robust and comprehensive solutions together.

1. With so much boilerplate intelligent capabilities available from Google cloud, why can’t we make use of them to ease real world business and operational tasks?
2. Why can’t the solutions be just driven by human natural language conversations and extending the utterances and unspoken too backed by social sciences?
3. Why does the business need to know so much about the domain and go thru confusing filtration criteria to get the desired information?
4. Why the dashboard solutions need to be so much inflexible and complex?

"Fortunately, there are ways, and we can pretty much speak/ chat over a phone or any other device like we would interact with a human agent using natural language and get our desired dashboards ready in seconds."

Here I am trying to present an easy and exciting way to build dashboards on the fly based on some publicly available sales dataset. This solution is built using Google services (can be on other public cloud providers as well e.g. AWS, Azure, IBM, etc.). The solution uses minimally invasive coding to create various integration points. This solution shows, how we can build highly cost-effective, enriched experience, faster time to market, and a smoother solution triggered solely through human natural language conversations. We can expand such solution beyond, applied to the following different use cases as well.

• Conversational financial reporting/ dashboards/ insights
• Conversational operations reporting/ dashboards/ insights
• Conversational business reporting/ dashboards/ insights
• Predictive, prescriptive, and descriptive customer/ product insights
• Conversational DevOPS/ infrastructure automation
• ?Conversational simulation of customer/ products/ marketing/ finance/ operations/ almost anything journey-centric
• and much more...

Technologies Used:

• Google Cloud Services
• Google Data Studio
• Google Dialogflow CX
• External Telephony/ Collaboration Platforms (Teams, Slack, etc.)/ Messenger App (Facebook App in this case)

High Level Solution Overview:

1. Users use natural language to interact over flexible channels without having to worry about the domain level understanding
2. DialogFlow CX understands the intent from the spoken/ unspoken, and extracts the embedded parameters to invoke the appropriate actions
3. Back-end decisioning, actions, and resolutions communicated back to the users through email as well when the output is ready or failed in real time
4. User clicks on the link that was shared as part of the communication at step #3 and views the dashboard (assuming user is already authorized)

Advantages

• Time to market/ value/ build for dashboards in many scenarios eliminated or significantly reduced
• Utterances can be well understood with google social and natural language capabilities
• Translation, transcription, conversion, ascent recognition, etc. all available as boilerplate features from Google further lowers the cost of ownership
• Single pane of glass for multiple dashboards and possible parallel execution
• Highly secure, scalable, performant, and resilient solution for varying conditions

Learnings:

1.?????Human languages to appropriate training phrase conversion and parameter substitutions are critical aspects of such solution to bring overall performance efficiency, scalability, reusability, and cost-effectiveness

2.?????Better requirement coverage bringing comprehensiveness to our scenarios is a must, to avoid high overheads in reengineering and future cost implications

3.?????There are massive opportunities for low code/ no code platforms where human skills can only focus on analytical thinking and problem solving to grow the business

Let's explore together new opportunities where we can go conversational.

Feedbacks, comments and any kind of inputs are appreciated.. Thanks.

Happy reading, Happy Holidays!!