Salesforce’s Vivek Kundra – ‘We need to rethink government institutions’


As a former CIO of the US government, Salesforce’s Vivek Kundra provides insights into the progress of digital transformation in the public sector.

vivek kundraOn a recent visit to London, Salesforce EVP Vivek Kundra took the time to sit down with me to discuss the latest developments in digital government. Kundra’s insights are always useful, not just because he has close dealings with the public sector at Salesforce, but because his previous job was CIO of the United States government.

Even more importantly, during his tenure at the US government, Kundra was responsible for some very forward thinking policies. For example, Kundra introduced the federal government’s cloud first policy – which given that it was 2010, and hardly the ‘peak’ of cloud maturity, was a pretty bold move.

During his time in London earlier this month, Kundra also dropped in to visit the Government Digital Service, where he said that he had a meeting to discuss the future of technology in the public sector. One area of interest was that of open data – but particularly in relation to how data needs to be extracted out of legacy systems to enable new digital technologies.

The UK has been at the forefront of the open data movement, being one of the founding partners of the Open Government Partnership. However, making open data useful and operational has always been a challenge. Not only this, but in my recent sit down with GDS Director General Kevin Cunnington, he emphasised the importance of getting the data out of these ageing platforms to enable a new data platform for government.

Kundra said:

The ability to provide much, much better services because you’ve been able to re-engineer a lot of those back office operations and are providing a level of transparency that has not been available in the past.

A couple of things. I think opening up the data is just the first step, and it’s a very, very important step because it’s a huge philosophical shift to create a more open, transparent, and participatory government, but that in itself is not sufficient. What you’ve got to be able to do as the next step is to make sure that you’re creating an innovation ecosystem. What I mean by that is startups that will actually start taking that data and building companies on top of it and encouraging startups

By having innovators build up platforms that are easy to fire up, easy to launch out so you can almost create this sort of app exchange, very much like you have an iPhone app store, imagine a government-wide app exchange where you’re developing these apps, whether it’s at the central level or it’s at the local level.

However, he added that it’s the transactional data that’s locked in the back-end systems that is important. Kundra said:

Secondly, I think what you’re beginning to see now is going after some of these large-scale transactional systems where you’ve got all this data, and now let’s say you’ve got a path to open it up. For the first time now you can because you have that open data, start building applications that are much more horizontal. I feel like sort of the last 40 years of tech investments, they’ve been very much around the back office. Inherently, the systems were closed. Now, all of a sudden if you start opening up data elements, creating new data models for those systems, you can actually begin to write code and launch apps that are horizontal, so you can almost abstract the departmental-level thinking.

Think about products

Kundra went on to talk about the difficulties in modernising government systems for the digital world – and getting to that locked-in data. He said that most of the modernisation is happening at the data abstraction layer, where data is being pulled out of mainframes. However, the next step for governments is to modernise those back-office systems – something with GDS Director General Kevin Cunnington has pointed to as critical for broad transformation.

However, Kundra notes that this is incredibly difficult to do. He said:

I think the path that we’ve seen where a lot of modernization is happening, it is in that abstraction layer. The reason you’re seeing that is because a lot of these transactional systems, they’re running on mainframes or minicomputers. The path forward has generally been to sunset those applications, but it’s the equivalent of changing an engine on a flying plane. This plane is going pretty fast, which is the agency mission, whether it’s around departments that are focused on healthcare or transportation or collecting taxes. You can’t just shut down and start the system in one day.

When you look at even GDS, they’re starting from the front end, which is great, very necessary, but now what you really need is the hard work and the tough conversations around what is the product roadmap when you think of agencies and their missions. What is the technology architecture? What is the data model? How do you make sure you’re building the systems ground-up for these new architectures?

However, Kundra also said that this is difficult because the majority of spend still currently goes on operations and maintenance of the legacy systems and that ‘digital transformation’ mostly still sits outside of mission budgets.

He said that if governments are going to take this seriously, they need to start thinking about these systems as products – in the same way that a technology company would think about the roadmap for their products. Kundra added:

Yes, they care about the front end, but the majority of their dollars are being spent on those systems. As a matter of fact, if you look at most governments, what you’ll notice is something like 95% of the technology spending is on operations and maintenance.

When you look at such a large number where majority of the tech spending across these governments is not on transformation. It’s not on redefining user needs. It’s not on a lot of the digital stuff. The next phase now need to be around upgrading and replacing and the journey of what does the product map look like. The reason I use the word product map is because a lot of government agencies don’t think of their systems as products.

For me, having come from government, you always would think of these systems as projects. A lot of IT systems, unfortunately, a lot of the ways they were run is people just keep throwing bodies at them, highly priced consultants, but not necessarily think of them like a tech company would. Say, this is our billion dollar product, what is the road map? How do we keep innovating?

That I think is sort of the next phase. You’ve got to move to a Cloud-first, a mobile-first world.

Changing the shape of government

tdgpromoDuring our discussion, I wanted to get Kundra’s take on what the future of government may look like. I raised the point that if we were designing a government today, the institutions wouldn’t look anything like they do now – many of which currently work in isolation, don’t talk to each other, and that the central-local divide may not have existed.

Kundra agreed, but his main point was that going forward, governments need to rethink their skill-base. He said:

It’s a natural continuum. The beauty is history teaches us that we’ve been through these migrations before.

From a policy perspective, too, we’ve got to be able to evolve in a tech era to be able to rethink a lot of these institutions. I think the skill set in the government now is going to be very, very focused on making sure that the people that are there understand procurement and how tech works, have fundamental understanding of platforms and the trade-offs that you’re making, having the ability to hold the vendor community to account and go after them when they don’t deliver in terms of lawsuits and in terms of making sure that taxpayers are getting the value for the money that’s being spent.

I think the past was technology was seen as this black magic that nobody really understood. Now, I think the beauty is because of the simplicity in many ways, you should ask the simply questions around why can’t I have that same experience on the consumer side with a deeper understanding that there’s a lot of complexity behind that simple search box, right?


Image credit - Images sourced via author

Disclosure - At time of writing Salesforce is a government and premier partner of diginomica.