Smart Bots for the Connected Future

The future of connected smart bots is happening right now in front of our doorsteps and at our fingertips - often unknowingly.

In this article I would like to explore and shed light on the advent of smart connected bots and how they are interwoven in the digital fabric of our lives.

There are several factors that lead to the proliferation of the use of smart bots for Internet of Things services orchestration and adding value to the smart connected devices ecosystem.

Taking advantage of new capabilities of smart bots however can pose architecture challenges for integrating with existing IT infrastructure, networking, software applications and processes.

The secret sauce to for the successful application of smart bots for connected devices, connected homes, connected factories and connected anything is applying a combination of real time data, deep learning and and cognitive intelligence to provide the right information at the time of now for a bot to make an intelligent decision and adding exponential value in the digital journey of AI and human interactions.

For service related bots that means providing real time information in the context when it is relevant to the human interacting with the connected device during conversational interactions with chatbots and audio connected smart devices.

The key is to have AI pre-processed data and deep machine learning analyzed knowledge bases ready for immediate real-time interaction and rules based answer processing in place to drive value add digital interactions with the human consumer via audio or digitally connected devices.

With today's technological advances what are different applications of smart bots in the commercial marketplaces that are adding value during customer journeys of interacting with smart connected products?

These applications include, but are not limited to:

1. Chatbots & intelligent agents interactions: Smart Chatbots understand natural language by the means of AI natural language processing algorithms that can interpret the intent of questions and pass instructions to the IoT gateway for processing the data running through deep learning algorithms and improving along the way providing ever better solutions during interactions with humans.
2. Crawlers & information bots integrations: these are bots that are running in the background and processing and classifying data as used in google search engine spiders, breaking news alerts and pricing assistants that monitor price changes such as in use in Expedia, priceline etc.
3. Transaction, Service & Niche Bots for connected devices: these bots are acting on the behalf of their human counterpart accessing any source that has an API to connect to which for instance is utilized in stock exchange type scenarios and automatically coordinating focused tasks such as coordinating a teams meeting schedule such as utilized by x.ai 
4. Master bots orchestrating multiple bot and non-bot interactions: these bots manage often seemingly disconnected services that need to be orchestrated as laid out in a McKinsey article for tasks that are focused on in home automation, but that are also applicable to bots in factories following the 5 Cs of cyber physical architecture and smart ecosystems such as connected buildings and connected cars

Example of the connected bot ecosystem in connected homes:

Real time closes the Gap between data and action

Components of a reference architecture for smart bot and connected device integration can be derived by applying the principles of the 5 Cs of Cyberphysical systems in conjunction with layered architecture like the worldforums IoT Architecture.

Combining that proven model with the Worldforum layered IoT reference architecture can be a guideline on how smart, self aware and self learning bots and machines need to interact in symphony with the layers of edge computing to data acquisition and human collaboration:

In order for bots to be effective in this highly interactive environment and serving up information in real time that is relevant to the user along the lines of these IoT reference models the Bot has to have these physical architecture model components in place:

1. Establish API connectivity with connected devices and other peripherals

When we think of connected devices we don't necessarily have to look far since connected devices could be mobile phones and smart devices in homes, factories, cars etc. connected to IoT hubs. The connected device could be even a webpage where interactions are processed and enhanced with service bots and chat bots. In order for a multitude of different types of Bots, smart devices and sensors to interact Bot developers need to have access to 3rd party REST APIs to drive machine to machine communications for streaming, structured and unstructured IoT.

Connected devices are often interaction together through Bot,, voice and service orchestration as this future vision of a connected car illustrates:

2. Authenticate and Authorize the user with the device or service:

Security Considerations for Bot interactions regarding authentication and authorization:

• User identity authentication: verified with secure login credentials, such as a username and password. These credentials are exchanged for a secure authenticated token that is used to continually verify the identity of the user.
• Authentication timeouts: revoked either by the user or automatically by the platform after a given amount of time.
• Two-factor authentication: verify a users identity through two separate channels (e.g., once by email, then again by text message).
• Biometric authentication: verify a users identity using a unique physical marker, such as a fingerprint or retina scan (e.g., Apple’s Touch ID).
• End-to-end encryption: only the two parties - the user and the IoT interface involved in the conversation can read the otherwise encrypted data e.g. Facebook messenger recently implemented this capability with a secret message feature
• Self-destructing messages: When potentially sensitive information is transmitted, the message containing this information is destroyed after a given amount of time similar to Snapchat

The Intel proof of concept of the enhanced privacy ID is an interesting approach to address IoT security concerns for companies and in this case specifically for TSA airline security, but this approach can be applied to other real world scenarios such as authenticating and tracking IoT connected devices in transit as laid out in the illustration below:

3. Gather, store and process the data gathered from IoT to human interactions

• Leveraging data protocols such as MQTT, CoAP, AMQP, Websocket and Node data is transferred and converted into machine readable information. The Bot often interacts physically through audio interfaces or chat with the human end user serving up solutions sifting through massive amounts of knowledge data bases after analyzing the questions through natural language processing. In order to ingest large amounts of data into a data store, at real-time or in batches and to process the data and feed it back to the user based on AI insights multiple components of an IoT storage and processing platform need to be in place. This includes cloud data ingestion layers and data lake cloud infrastructure to manage massive amounts of streaming data to be ingested even if we only care about certain events that are set in a rules event engine. Ad-hoc cloud blob storage, unconstrained bulk data ingestion and specific events based data ingestion require different tools and architectural approaches all leading to being able to gather, store and process the data.
• The following architecture can be applied irrespective if you are a Microsoft shop or rely on other cloud platforms such as Amazon, Googles or others:

4. Cognition Logic

Making all the data we now have "actionable" is the final critical step, towards increasingly personalized services. Being able to segment users based on the data and scenarios and perform actions automatically when certain criteria are met.

In order to set up an interactive and AI driven Bot service we need to continuosly and in real-time monitor data in the background and respond automatically at the right time and place.

Imagine a world where our homes, cars and wearables are connected to flexible and scalable cloud infrastructure and AI driven predictive intelligence that can provide real time feedback and even start a conversation that adds value to our everyday lives.

If you think this is a far fetched future vision - think again since we are being disrupted with AI driven bots as we speak as shown in the short video below:

~~~

Disclaimer: The opinions expressed in this post are my own personal views and don't necessarily represent the views of current or previous employers.

If you like this article please click the like/ follow button. I look forward to your comments and insights regarding this topic.