How to get practical value from artificial intelligence
Connecting state and local government leaders
Deploying AI and machine learning “out of the box” allows agencies to quickly get benefit from the technology and avoid long procurement cycles and complex integrations.
For the past decade, cloud has been all the rage in federal IT circles, as agencies look for ways to decrease the burden of legacy IT spending.
Today, the IT modernization push continues, and agencies can now see the light at the end of the tunnel. So far in 2017, there have been positive signs from Congress and the White House that IT modernization will remain a critical part of our government’s priorities moving forward.
Earlier this year, President Donald Trump signed two executive orders -- one on IT modernization and one on cybersecurity -- that signaled his administration would be taking the push seriously.
Similarly, the House of Representatives approved the Modernizing Government Technology Act, commonly referred to as the MGT Act, to set aside funding for federal agencies to upgrade to new platforms, such as cloud, to drastically overall its IT infrastructure. This week the Senate too approved the MGT Act, as an amendment to the National Defense Authorization Act, so it will be taken up by the conference committee that hammers out the differences between the House and Senate defense bills before moving to final passage.
We have also seen movement from the president’s new American Technology Council, which released its own 51-page report in late August that summarized the state of federal IT modernization and outlined next steps.
What is most exciting about this wave of activity for IT modernization is that it is grounded in reality and looks at objectives that can actually be achieved. For too long, discussions about federal IT have focused on solutions or technologies that were not yet feasible or could not provide the instant impact the federal government needs today.
Nowhere is that more evident than the buzz surrounding the latest hot topic in IT -- artificial intelligence.
Understanding exactly what AI means
When “cloud” first appeared, it was loosely defined and even less understood. Likewise, the vague and ill-defined descriptions and promises of AI can potentially frustrate and confound public-sector technology leaders.
Understanding what flavor of AI technology actually provides which capability and how it might be applied in a specific environment to achieve a desired outcome will be 75 percent of the challenge facing government tech leaders over the next few years. As with all new technologies that we have slowly integrated into the federal technology ecosystem, AI is a great capability for public sector organizations to leverage. But are we really going to trudge down the same adoption path that we suffered through with cloud?
A common approach to integrating a new technology in government depended on execs first grasping its applicability to the mission and then actually acquiring, deploying and supporting it. This has typically been a long, arduous and expensive engagement that consists of multiple studies and the enlisting of several consultant firms.
Typically, an agency hires a third party to review the various flavors of AI tech, assess the vendors that provide the capabilities and propose how the technology might be used to achieve a specific agency outcome. Only then does the agency begin the daunting process of procurement, deployment and integration followed by the care and feeding, upgrading and ultimate retirement of the new AI tech.
This legacy process could take years and tens of millions of dollars before any benefit is realized. So why would agencies continue down this same path? In my opinion, they shouldn't.
Practical AI: A better path forward
Today, an agency can leverage “practical AI” that is already a fully integrated element of a cloud platform and can be employed to create usable information and deliver outcomes within months of an initial requirement description.
By leveraging platform-based AI, agencies can take the exploration phase out of the equation and shorten the acquisition and deployment phases. When AI features are native to the platform, they can use any other native application's dataset or any customer-defined workflows that reside in the native applications.
This practical approach to deploying AI and machine learning “out of the box” allows agencies to quickly get benefit from the technology and avoid millions of dollars of exploratory investment, long procurement cycles and complex integrations.
For too many, AI conjures up images of sci-fi movies and the future. Instead, AI can be implemented tomorrow in a very practical, very convenient way.