Assistant Cards are the heartbeat of the Assistant as they show any information at a glance and can be consumed by the user on any device and for any purpose; to view information, explore data, trigger actions or respond to requests.
Cards are made with Google Polymer Components and our open-source components that can be quickly assembled to build lists, forms, buttons, etc.
Values are dynamically inserted into the Card markup and the layout itself adapts itself using Liquid as the templating engine.
The components of a Card can be assembled and reassembled infinitely ensuring low-maintenance code once, deploy everywhere development.
Designed for at-a-glance information, Cards can show rich media like images, galleries, dashboards and gauges.
Cards have multiple configurable Actions, including opening a Modal, triggering a command, returning a response to an API, getting suggestions from the AI engine and more.
Cards are written once, but can also be deployed as voice only, as Microsoft Adaptive Cards, as HTML5 Components or as Slack elements.
Cards can be triggered through a multitude of events, either on a schedule, through a webhook, by a user question or when the AI identifies a notable event.
Companies can define downtime where most Cards would be aggregated without triggering notifications. In addition, users can pause Card delivery during busy periods like during a meeting.
Cards a user stumbles across in Search or embedded elsewhere can simply be added to their Board, where they will be persistently visible.
Digital Assistant has two modes: a stream of notification Cards that the user can work through with the goal to be done, or their personal Board where they can always see up-to-date information on their favorite Cards.
Users can browse Cards that are available to them and subscribe to or unsubscribe from them. Some Cards may be pre-enabled based on their roles and Cards can also be mandatory for certain audiences.
To preserve the overall clean look of Cards, they can be expanded by either revealing entire section, showing more items of a list or showing more detailed information in a modal.
Just like a YouTube video, any Card can be embedded on any other website dynamically allowing you to expand the functionality of other websites like Intranets or SharePoint pages.
The built-in and browser-based low code Card Designer allows you to make easy changes to any of the Components the Card consists of without having to open code editors.
We provide a number of template components to help you make new Cards from scratch, including lists, buttons, modals, notifications and more. Most Cards can be developed in half a day.
One of the most radical concepts of Digital Assistant is Portable Services. That means you create a business case end-to-end once and you can port it to any device, experience, environment you desire. Here are all the ways you can use it:
The Digital Assistant app is a browser-based, responsive design PWA app that works on any device and initializes with smart caching, offline mode and “Add To Homescreen” support.
Supporting multiple native app wrappers, you can deploy Digital Assistant through your MDM.
The PWA supports push notifications on Android, Xamarin for iOS apps and the Chrome extension offer desktop push notifications.
Supporting Chrome, Firefox, Safari and Edge you can invoke your Digital Assistant using the browser extension which even features a badge count of the number of new Cards for you.
Each type of Card can be sent to just one or multiple channels the user is using, i.e. Sales info only to their desktop.
If the AI notices a notable change in the data it can send an update on all Channels regardless of the preferred Channel.
Leveraging the Adaptive Cards standard all Cards can also be natively shown in compatible containers, i.e. Microsoft Teams, Skype for Business, Cortana, etc.
Any device with a smart assistant built-in, like Alexa or Google Assistant, can also invoke and read out responses from Digital Assistant.
Supporting a wide range of chatbot channels you can use Digital Assistant natively within your favorite messaging platforms.
Everything that gives Digital Assistant it’s smartness, predictiveness and generally contributes to the ease of use and UX of our product is summarized by the term adenin AI.
When the user starts typing a question the AI suggests questions it knows answers to and gives hints what it is missing (i.e. a name) to make a question result in a valid search result.
The AI can check for preconfigured value ranges, check conditions and test regressional data and – if any notable conditions are met – displays them right on the Card for more contextual decision-making.
The Natural Language Processor (NLP) can be trained in multiple languages simultaneously so users can ask questions in many languages interchangeably.
As part of the AI’s training you can construct markup that automatically flags part of a question as being the entity or intent, thus helping separate question from fulfillment.
The AI learns when you most often consume a certain type of content and the algorithm will show that information more timely, when other kinds of information might be shown in summaries only.
Looming deadlines, unexpected spikes in data, action words in messages you receive and more can all trigger an ad-hoc notification on all your devices. It’s the Assistant trying to catch your attention when it thinks you may miss something notable.
With the use of Machine Learning algorithms Digital Assistant will build profile from usage analytics to suggest content to users that are similar to profiles the algorithm created. Admins can also inject their own recommendations as part of the ELK stack.
The built-in Analytics ELK stack allows you to explore the usage data and workings of the AI engine. By running reports you can see which phrases the AI couldn’t match and thus allows you to re-train it for better accuracy.
Leveraging the transcription services by Google or Amazon you can offer users to ditch the keyboard and interact with their Assistant from their Alexa, their car or even with the voice input button in the Assistant’s app.
Many organizations achieve a good basic coverage of employee questions by incorporating FAQs or Helpdesk tickets into simple question and answer bot questions which gives basic personalization options.
Through the use of NLP you can start any request with a simple command to pre-populate fields before their submission, e.g. “Schedule meeting with Kevin at 4pm tomorrow” would pre-populate the invite with all the info the Assistant could gather.
The entire AI engine is available for installation on-premises so no data has to leave the network. Alternatively, even for a hybrid environments requests can be routed to private servers for ultimate security.
As a platform, Digital Assistant comes equipped with everything large and small enterprises require to reliably and effectively deploy the solution to their workforce.
All data Digital Assistant integrates can be transformed with a node.JS layer which means you can alter the incoming JSON or other data formats before you insert it into your Card.
Many enterprise applications offer workflows and approvals so Digital Assistant aggregates all those into one list of all an individuals’ approvals for convenient management.
From adjusting the verbiage of the UI, to complete translations and LTR support you can completely adjust the Digital Assistant web app to your needs.
Audiences are recommended user groups that would get preconfigured Cards and better recommendations for search and preferences.
To conserve bandwidth we use service workers that manage real-time updates to the user’s Cards and transition between offline and online mode.
Digital Assistant has a Stream of updates and notifications that the user works through as well as a Board of persistent Cards that the user can always see.
Supporting worker, Redis and database clusters, webfarms and a separate ELK stack integration you can seamlessly scale Digital Assistant to fir your demands.
All our branding is optional and we encourage customers to find an exciting new name for their Digital Assistant implementation.
Whether you search or poll APIs, all processes happen across all integrations simultaneously showing the first results as soon as they come in and not until the last service responds.
You can create your own declarative integrations for any custom databases, RESTful APIs, SOA, RSS feeds and many more formats.
The users view is always in sync between all their open sessions. Ask a question on a smart speaker, and get the visual response on an ambient display alongside the voice output from the speaker.
If you want to integrate Digital Assistant Cards into your own Enterprise Search or chatbot to serve the user relevant Cards then you can query the Assistant API.
Cards on the user’s Board can be freely rearranged to create a highly tailored personal dashboard for the user.
You can add a number of Cards to your own collection of Briefings that the user can request specifically, i.e. as a morning briefing, a mid-day check in briefing, etc.
To reduce the amount of distracting Emails you can redirect notifications to a personal Email dropbox where your Assistant aggregates them into a list of updates.
You can choose to deploy Digital Assistant as a SaaS application, in your private cloud or install it on-premises.
For reliable deployment in the enterprise you can rely on a plethora of safety related features that ensure data gets to anyone who is allowed to see it, and no one that isn’t.
Every user action and data access is logged in a secure container so it can be accessed and audited anytime.
You can integrate Digital Assistant into your federated identity service to allow seamless access of the app but also secure access to the data Digital Assistant integrates.
Multiple failed login attempts lock an account and send out an Email to the administrators.
The entire app is naturally HTTPS and SSL encrypted.
To protect outside attacks Digital Assistant can be run inside your DMZ where it dynamically loads via a proxy only the data that user is authorized to access in that moment.
Cards and data are both protected by permissions you can define yourself or remap from existing user groups from your identity service.
Users will only ever find items that they have permission to view in the original source, too.
Both as an on-premises application and a SaaS provider we comply with industry standards and we conduct our own continuous audits of our security. We share all reports on request.
*Some features require at least the Medium plan.