Removing technical debt can be crucial for small counties like Santa Clara. In this AMA, CIO Ann Dunkin lays out the biggest challenges an IT leader when dealing with public sector data.
What is holding back a CIO in 2020?
That’s a really interesting question. It depends a lot on sort of the type of organization you’re at. If you look at these big organizations with lots of legacy systems, the answers are really different to these little nimble startups that basically have no IT.
For big organizations there’s this tremendous opportunity with all these new technologies in the cloud. The fact of the matter is most of us are sitting on a tremendous amount of legacy hardware and software that continues to generate technical debt. You have to constantly balance taking care of those things with being able to move forward.
You can’t stop supporting that legacy to move forward. So there’s this constant struggle for resources between your legacy systems and the new systems that you would like to implement or the modernization you’d even like to do to those systems. So balancing all that noise of new projects with what’s truly strategic is a huge challenge.
Is this balancing act the case at Santa Clara too?
We have a lot of legacy systems. We still have a mainframe and an assessment system that is quite long in the tooth, as do most of the California counties. Large counties are struggling with this. There’s a solution that has been working for the small counties, but the large ones are really struggling to solve that problem.
There’s definitely a tremendous amount of technical debt. Large organizations that provide services like hospitals and things like that have this physical infrastructure that we have to have and that doesn’t go away. All public services–the hospitals, the police departments, the jails–all those things have to stay. Meanwhile, we’re trying to get off the mainframe and trying to modernize the system and move things to the cloud and balance all that with the ability to keep operations running in an emergency.
What’s the biggest problem you face in attracting talent that can help you remove technical debt?
The problem is our system. It’s not the willingness of people to work for the county. It’s our ability to run our hiring process fast enough to get people who are looking to decide to stick around and wait for our offer as opposed to the three offers they got last week. Our processes are just slower. That’s partly due to civil service rules, but partly there are ways we can and are trying to speed that up.
But we can offer a competitive base salary and people can then choose to have a better work- life balance by working for us. They’ll be able to choose to serve the citizens of the county, instead of choosing a private sector role. They’ll have ongoing development opportunities, which has been less the case in the private sector than it used to be.
How quickly can new technologies be rolled out in the county, like 5G for example?
5G is an odd one because the places where you’re going to put up 5G antennas don’t belong to the county, they belong to the cities. We only manage unincorporated areas for the most part.
But the bigger point is that there’s all sorts of new technologies–AI, machine learning robotic process automation, tons of IoT– it’s a mixed bag. How fast we are able to implement those things is important. Our CTO runs an innovation lab where anyone in the county can propose something to it. We’re able to do zero cost and up to $25,000 projects without competition. Those prizes are proof of concepts and if you want to implement something we have to go through a competitive process.
That process really allows us to jumpstart the ability to look at a new technology and determine whether it’s going to serve our needs well in the county or not. That’s one way that we’re able to bring innovation in more quickly.
Another is just trying to build up, strengthen the architecture program, which really didn’t exist until a couple of years ago. Strengthening the role of our architects so that they’re able to be forward-looking about what technology they believe they should be implementing in the future in the organization. Give them the ability to build that into our roadmap. There’s a constant stream of requests from other parts of the organization for things that they very reasonably want right now, most of which aren’t new technologies. In some cases, we can look at that and say, there’s a better way to solve the problem that you’re looking to solve using new technology. But in other cases, you know, it’s pretty pedestrian things that people need to keep their lights on.
What kind of dialogue is going on about the sharing of data in the government?
In the county, because of the types of work we do, it is a really complicated problem. We have data considered public and subject to a public records request. But there’s also criminal justice information. So services in house health and hospital data are subject to HIPAA and a bunch of other regulations. A big part of the conversation that goes on right now is how do we find ways to use data that serves the community while protecting the rights of our citizens?
Every time that the police are called to this address, this person ends up in an emergency room. Three hours later, how do we figure out how to get some services to that person so that it stops happening?
We’ve got people who are trying to get off the street and into permanent housing. When there’s part of housing that needs a bunch of wraparound services, how do we bring all that information together? A dashboard that someone can actually look at? Because by the time you pull the data from the dashboard, there’s no one who’s authorized to look at the data from all the different departments.
So it’s a really complicated problem to be able to figure out how to use that data while protecting people’s privacy and not about not breaking the law.Just recently saw something cool happened in a collaboration with UCLA, where they gathered data together to be able to identify people who were likely to slip into homelessness, based on financial events and be able to intervene with them with cash assistance, to keep them in their homes.
I was very impressed by that. That must have been really complicated not only to get the data in a meaningful way, but to get the data way that someone could actually share it.