Ten years ago, or even five, we would not have been here discussing artificial intelligence (AI). Now we hear all the buzzwords — AI, NLU (natural-language understanding) and NLP (neuro-linguistic processing).

What does it all mean, and where does AI fit in our governments? Other concepts fall into the spectrum of government, such as reporting, business process management and automation — but a key consideration is how they all integrate.

AI is only as good as what it is engineered to provide — for example, being able to get parcel data to residents and contractors in the field, or at home, or somewhere, 24/7/365, etc.

Another acceptable constraint with governments is the idea that we are to serve the public with the concern of budget. In serving the public, it is our obligation to provide as many ways as possible to access information and support including chatbots, websites, 508 ADA compliance, telephony and other such tools and services. This is where integration and platform-agnostic concepts begin to move into the spotlight.

AI by itself is rather boring and, in many cases pointless. However, when integrated with chatbots, user interfaces and telephony solutions, the use cases begin to expand exponentially.

Many problems are found with proprietary solutions, making them less agnostic and more difficult to integrate. Government agencies are bound by the obligation to serve, and doing so requires delivery of access using many solutions and outputs to residents and constituents. Moving forward, I see the need to provide such application support as API services that can be used by various AI systems and integrate with other products or services. API standards provide serializable data with JSON (JavaScript Object Notation), which allows for cross-platform integrations.

Google has done an exceptional job structuring their Cloud Datastore product. This allows data sets to be stored in fast NoSQL schema-less data structures, accessible from anywhere on the Google Cloud Platform (GCP) as well as microservices published on-prem or in the cloud. Security becomes a question, but many features of chatbots and AI are public-data driven, and security is less of a concern. However, internal processes would have to be filtered by separation of AI agents, and microservices to back-end solutions would require SSL and possibly additional authentication. Further, microservices can be written in C# .NET. 

Many government agencies are already equipped and staffed for this development scenario. To top it off, the C# microservices can be stored in the Google Cloud App Engine, which lives in the same cloud as the Datastore, further decreasing response times and increasing performance to the end user of AI interfaces. This eliminates the need for retooling or adding additional unbudgeted staff members. Return on investment (ROI) will become more important than ever to governments; private industry has been here for a while.

Next steps will be to bring AI and integrations internally into the operations and business functions of government, such as call centers, ticketing systems, HR systems (employee life cycle, onboarding and offboarding) and more. Now you can begin to envision how important portability and system-agnostic products will become in the future.

Did I mention that the future is now yesterday?