If we don't examine business processes and put enforceable policies in place, then CIOs will have 10,000 AI applications because everyone is still doing their own thing. Image: Robert Way/Shutterstock.com
By: Rob HankeyDespite decades of effort to standardize sprawling IT infrastructures, many agencies are still doing “their own thing”; the risk of AI is repeating history.
A few years ago, I spoke with a CIO for a federal agency that left an indelible impression on me. I was pitching a software solution, and while the CIO had agreed to meet, there was an objection on her mind I hadn’t anticipated or heard previously.
It wasn’t the solution, per se, or the problem we could solve, but the size of the agency’s existing technology portfolio that worried her. The agency is large, with tens of thousands of employees, and maintains a sprawling IT infrastructure. It had 10,000 different software applications in its production environment and the CIO wasn’t keen on adding another.
Her day-to-day focus was on keeping things running and that’s an enormous challenge because her team didn’t have the expertise necessary. Why? It’s simply not economically viable to employ experts on every system when an IT portfolio grows to that size. As such, even routine IT outages, which should be a simple remediation, turn into a fire drill.
At this point, my curiosity got the best of me. I asked her why she didn’t put standards in place. That would allow her to consolidate the portfolio of products into something more manageable.
The upside, in terms of efficiency and cost savings, of embarking on such an initiative is well documented. Even if she only made a small dent during her tenure, such an initiative would put the agency on a better track.
I’ll never forget her answer: “It’s just too hard,” she said. “Everyone wants to do their own thing and it’s just too hard.”
Sprawl accumulates over time
Unfortunately, her story isn’t an exception. Many federal agencies find themselves in a similar situation and there are valid reasons why:
· Priorities change with elections;
· Senior civil servants tend to move around agencies; and
· Department heads wield their influence to get a pet project done.
Despite many efforts to drive standards across the government – i.e. the decade-old Federal Information Technology Acquisition Reform Act (FITARA) – agencies maintain (and should have within reason) the flexibility to mold technology to their unique needs.
Over time the sprawl grows, and silos emerge. The way one system collects and stores data isn’t compatible with another. Fixing this could save money in the long run, but it’s a heavy lift and budget pressures often put such projects on a backlog.
Absent a crisis, or a similar catalyst for change, the status quo reigns.
Process before automation
In my observation, the challenge stems from unmapped business processes rather than merely data standards. If a business process is poorly designed, weak, or unclear, then that is the piece that needs to be addressed first.
Mapping those processes out and evaluating them is a crucial prerequisite to technology procurement. The consequences for skipping this step are severe: you risk automating bad processes – doing the wrong things faster – and it’s very expensive to fix.
Recent news about an embattled IT project intended to facilitate background checks for federal employees is a good example. According to reporting, a project to modernize this system is five years overdue and has cost taxpayers nearly double its initial $700 million price tag.
Why is the program so late and over budget?
“The key underlying cause of those shifts was not realizing all the tasks that needed to be done to deliver that full functionality,” a senior official with the Government Accountability Office (GAO) explained to NextGov.
Pervasive across government agencies
More broadly, background checks are instructive because they show how the problem doesn’t just exist inside agencies – but across them as well. Multiple agencies require background checks – defense, intelligence, justice, immigration, and energy – and many of them use a different system.
When you ask these agencies why, they’ll tell you they have their own unique needs and requirements. While there are levels of effort to background investigations, depending on the risks, at a business process level there is virtually zero difference.
Every single background investigation in the U.S. starts with a form where the person applying for a security clearance inputs data. Depending on the agency, or level of clearance, some forms might ask for more information.
For example, one agency might ask for your home address for the last seven years, while another wants a list of every place the applicant ever resided. Regardless, from a business operations perspective, it’s just another data field in a process that is more or less the same.
If that process was standardized, then agencies would still be able to tailor the fields – and by extension the automation – to their unique needs. Further standardization means they could share data, reuse it, or change vendors if desired. This would also improve the user experience (UX) because standardization shifts the focus to ease-of-use and support.
Common ground
There may be valid national security reasons for the way background investigations are managed today, but that’s beyond the scope of this article. The point is to illustrate why understanding the business process is crucial because that’s where agencies have so much common ground – both internally and externally.
Take veterans for example. When they apply for educational benefits there’s a specific form, they must complete to initiate the process. Later, when they apply for a mortgage loan, healthcare coverage, or any of the other benefits they’ve earned, they have to start from scratch again on a different form and often a different website.
Yet when you look at this from a business process level, the overlap is clear. All of these “different” forms effectively collect the same data repeatedly: contact information, service records, and proof of eligibility. More importantly, this is data the government already possesses but can’t use and so the veteran spends hours re-typing the same information.
This scenario occurs thousands of times every day across dozens of business processes and agencies – from onboarding new government employees – to the process citizens follow to apply for a variety of government approvals or benefits. Even so, the process of applying for benefits, whether it’s help with sustenance, scholarships, or social security, is largely the same.
The Office of Management and Budget (OMB), rightfully calls this a “tax on time.” It reports that Americans spend more than “10.5 billion hours each year completing government paperwork.” I suspect a lot of it is repetitive. More to the point, that time could be reduced if we all collectively did a better job of understanding the processes across government, before adding automation.
AI, modernization and the risk of repeating history
While many agencies are in the midst of much-needed multi-year modernization programs – some run the risk of repeating history. More specifically, the rush to acquire artificial intelligence (AI) may reinvigorate, if not accelerate this problem, because AI moves so much faster.
“It seems like everyone wants to play with the new AI toys, but most don’t realize the standardization and structure of data needs to be in place before that can be a reality,” one senior government IT official told me recently.
Indeed, if we don’t first invest the time to thoroughly examine business processes and put enforceable policies in place, then CIOs, like the one in the introduction to this piece, will have 10,000 AI applications because everyone is still doing their own thing.
Rob Hankey is the CEO of Intelliworx which provides FedRAMP-authorized workflow management software solutions to more than 30 federal government departments and agencies. A retired rotary wing pilot for the U.S. Army, he later worked as a government employee before founding Intelliworx.
OPM Advises Agencies on Conducting RIFs During Shutdown
Updated Shutdown Contingency Plans Show Range of Impacts
Use Shutdown as Justification for More RIFs, OMB Tells Agencies
Unions Win a Round in Court Disputes over Anti-Representation Orders
Deferred Resignation Periods End for Many; Overall 12% Drop
Senate Bill Would Override Trump Orders against Unions
See also,
How to Handle Taxes Owed on TSP Roth Conversions? Use a Ladder
The Best Ages for Federal Employees to Retire
Best States to Retire for Federal Retirees: 2025
