Derek Holt, Digital.ai | DevOps Experience 2022

ai. i. I think many in the industry have as we think about this inflection point of where we are at from an industry.

And as we think back towards the agile transformations of the early 2000s through the devops Revolutions of of 2010 and Beyond and now more and more looking in the role the data is going to play and the role that value is going to play in the next turn of the crank when it comes to how we build the liver and optimize software Investments within within our organizations. And so I'm excited to have a chance that Charlie a little bit today and explore some thoughts and ideas of regarding what we're seeing in the industry and and ultimately maybe predict the future a little bit as we as we progress on and so it's interesting, you know as Here today, it is over 20 years, which is a little hard to believe since the signing of the agile Manifesto and and subsequently roughly 10 years later and since famed venture capitalist, Mark Andreessen penned his famous Wall Street Journal op-ed why software is eating the world the concept that every business, right? i.

We have the opportunity. We work with over 50% of the Fortune 500 with over a thousand large scale Enterprises around helping them accelerate their digital transformation and helping them ultimately harmonize the business process and software development and delivery to do the development and ultimately deliver on the outcomes that that they expect that their shareholders expect. At their customers expect and what I see more and more whether that be Banks or insurance companies or travel companies or retailers is that there's a more and more realization that every company in many many ways will become a software company and again not just in the ways that they interact with their customers, which is obviously the one that we mostly experience and as I jump on a flight here later this week, I'm not gonna print out tickets.

I'm not gonna very likely even talk to an agent. I'm gonna pull up my mobile app. I'm gonna I'm gonna ultimately check in I'm gonna get a digital ticket and and I'm gonna kind of scan my way all the way to my to my destination in many ways.

Um, but it's also not just through the lens of how we interact with our customers. It's also if you think about it's how we work and if anything one of the interesting observations from the obviously extraordinarily challenging times over the last few years in terms of the pandemic was the ability to quickly go home. And work from work remotely, right?

And and obviously there was a tremendous amount of challenges and loss and other things with the pandemic but we have been able to continue to do business if you will and in many ways, I like many others. I'm sure set up at least a bit more of a comfortable home office and and I was able to kind of quickly adjust my way of working and software was a core part of why I was able to do that. So ultimately Mark andreessen's prediction, you know, 10 years ago of every software eating the world I think is in many ways becoming more and more true every day not just in high tech but in every single industry as as software permeates across the across the board and yet As we sit here today in 2022.

ai. We think a lot about the Enterprise our core focus of the enterprise. We love startups.

We love small businesses many of us like myself. I have done a series of venture back startups, but when it comes to how we think about our portfolio and and who are Target end users are it's the Enterprise. It's scale.

It's heterogeneous environment. It's it's not dozens of developers. It's it's tens of thousands of developers spread all around the world and previously spread around the world in different development now spread around the world in a hyper distributed way around many, you know, working working from home or at least in a hybrid work environment.

And so if you have 10,000 developers, you may actually 10,000 lives because everybody's everybody's home is their individual lab, but as we think about the last 20 years while we've made Tremendous amount of progress and we've written a tremendous amount of code and we've we've matured from the way that we we plan and the way that we iterate and the way that we deliver in the way that we test and secure. We also know that they're continues to be challenging right it particularly again the areas that we focused on these challenges that are typically related to scale and so I think about these three and these whole true I get the opportunity to spend a lot of time with cios and and ctOS and and I like these almost every time I have a meeting and at least two of the three I'm getting ahead not if not all three and and so, you know, even though we are here before 20 years after the signing of the agile Manifesto and we have made tremendous amount of progress around becoming less waterfall more agile at the particularly the teen level there's still a tremendous amount of work to do to adopt agile practices at scale that means teams of teams, but also tool chains of tool change so across End of end and kind of the broader from idea all the way to running code. There are still a lot of opportunities for growth and we saw that you can see a light quote.

Here we run. A an annual state of agile survey. This was our 15th year.

Just just last year. We'll launched the 16th here coming in the next couple months and we saw during the pandemic. I think many ways that Miss Destiny an even broader acceleration of teams self-identifying is as using agile practices and you'll see some other data from that survey throughout today's presentation.

We've all also see more more organization say hey over the last 10 12 years really due to the fact that software became so much easier to play for credit card and consume. I feel like I've got a lot of different tools do a lot of things. In fact, I CIO tell me the other day.

Yeah, right. I got five planning tools Etc and you are now seeing the pendulum swing back into the ability to or the want to be able to unify tools together and I think one of the interesting Dynamics there is we actually believe Does not mean rip and replace what that means is to for individual Enterprise companies to kind of get themselves out of being the tool integrator. They don't want to be that that's not a value add for them but to work with with companies that can help bring together some of their existing Investments and then obviously bring more value on top of that and that's a big part of this is not a commercial for digital that AI today, but that's a big part of what we spend our time thinking about along the way and then lastly one of the things that I'm most excited about you're gonna hear a lot about this to today's presentation is for decades.

Now these tools both in the development side and more and more on the production side with this the standardization around ideas and support that and and user Telemetry Solutions Etc have been throwing off a tremendous amount of data and we as an industry and I use this as a self-criticism we as kind of vendor industry historically have not done that much interesting with it in my opinion. Yes, we can set off alerts. Yes, we can.

You burn down charts velocity Etc. But it's a lot of the team level and it's a lot to me of data data. You kind of domain specific data lakes and and lightweight dashboarding.

There is a huge opportunity and this again is something that we're we're delivering on today to use all of that data to and applied modern data lakes and machine learning and AI do not only be able to get a more panoramic view of how we develop and deliver software but predict the future based on the past and type technical outputs to business outcomes more directly than we ever before and so these are the big trends and again, depending on what customer we talk to one maybe more important than the other but but I will tell you these three are very very kind of top of mind for for kind of key key Executives. And so that then warrants the question. Okay, if that's the case, how do we how do we accelerate beyond that over the next 10 years, right if the 2000s were about agile and taking that off which obviously A journey we continue on if 2010 and Beyond was about Automation and doing things at scale and governance and compliance around around the the business process of software development delivery.

How do we accelerate into this next decade and how do we address some of the challenges I highlighted before so number one. How do we adopt agile practices of scale number two, and I think this is really really important just because the software side of the house the faster doesn't mean the business goes Pastor. So how do we start that adopt agile practices everywhere?

Not just within thought we're development teams. And what does that look like? How do we start the unify the devops platforms as I talked about before not in a rip and replace way but in a in a evolutionary way that starts to tie kind of core existing assets together, but also enhances those workflows through open Integrations.

How do we bring this augmented and analytics and Ai and ml both to the process of how we build software but as we'll talk about a little bit later also about what software we build is it valuable that our We're gonna happier or more upset. Did we make more money or less money? How do we tie these software Investments?

Not just to builds and test coverage and how many tests pass and and cycle time and all of that. How do we tie back to actual dollars and cents and then lastly along that we how do we kind of drive on this route on this kind of Journey towards what you know, some call Value stream management. We certainly highlight that but at the same time this is just about that transition too high back to to business value and I think in many ways moving past something we used to say a lot in the past which is a business and it alignment to now say business is software and software is business.

These things should be very very tightly integrated not just aligned which I think is is the next step of that Journey. So I want to walk through each of these just to kind of give a few thoughts and and then we'll we'll hit on a few other elements before we before we close up and so, you know here as I highlighted earlier this ability to adopt agile. Is a scale has been challenging and I think some of this is rooted in a lot of the team level tooling that is in place.

But also just the the more complexity when you have teams of teams and and need to manage dependencies and planning and whatnot above that it some great obviously some great Frameworks have emerged over the decades like safe and others and and there's a lot of different best practices that teams can can pull from but the requirements are very very key. Right we need to be able to do this across heterogeneous tool chains and environment. So it can't just be for the latest thing in Mobile or container or kubernetes development.

What about our Mainframe Investments? What about our our Colo Investments? What about our more traditional Cloud Investments?

It needs to be agile across the board because of course in these composite applications that exist today the dependency expand all of those all of those realities we need to be able to do in a hyper distributed way, right? We in many ways maybe not certainly on the days that hey we're gonna get a hundred and fifty people together in a big Planning section that's harder to come by these days. And so how do we use more and more collaborative tools?

But our collaboration tools but make those collaboration for more domain and and and operationally aware along the way and as I highlighted earlier this business and software as one no longer, it's me. It's insufficient to Simply have business and quote it alignment. This is about acknowledging that that's software is quarter our business and business, you know, the end goal of the business is quarter our software investment then and unifying these these teams one and you're seeing this obviously in the tooling but you're also seeing in the way folks are organizing their team and I think that will continue to evolve as we continue to progress.

You're the second one is agile everywhere. This is one that was really really exciting for us as we did as I mentioned earlier our 15 date of agile report one of the things that bubbled up that I had not expected. But once we saw it, I wasn't surprised by it made a lot of sense was during the Dynamic this was taken right around the first end of the first year of the pandemic.

One of the things that was was kind of forced upon many organizations was to be more agile within other within other parts of the business, of course. Software development and it are leading the way there but think about being more agile around operations and around marketing and around sales and around other parts of the business. And again the pandemic sort of accelerated this because there were points in time if we all think back where we didn't know what was gonna happen the next week or the next day or the next month, of course, and and so with that out an accessity folks started to adopt agile practices making the entire business not just software part of the business with the entire business more in Nimble more Dynamic and what what's been exciting is folks have started to see the value in that and they've not kind of gone back to the way things were before.

So I expect these numbers to continue to grow and again, it's critical we can be as agiles and you business is an agile. We're not gonna see the ultimate benefits that that we that we need and ultimately that we expect and so I mentioned this earlier but one of the things again if I'm old enough they myself Here, but I'm older not to remember when we were buying kind of one size fits all one product that fit everything and it was like a centralized buying decision. This was gonna be late 90s early 2000s.

The pendulum certainly is swung over the last 10 years around kind of the proliferation of more and more software Solutions and very nichy capabilities often, but also a more autonomy for teams that is created a bit of a tool sprawl. It has also interestingly enough from the positive side. It has created better Enterprise software.

So you've seen sort of the consumerization of enterprise software the tools are better easier to use more enjoyable to use a great thing, but we're gonna start seeing the pendulum come back and Gartner highlighted here by by 2024 60% of organizations are gonna move more towards platforms then to spare tools that they have to stitch together. But again as I mentioned earlier and others disagree with this, but our view in the Enterprise in particular is that does not mean that people are gonna get rid of all the tools they have today and start Over what they're going to need is a platform approach that is an open platform that is going to allow to integrate and that's why you know, I know one of the things that we invest heavily heavily into the to the tune of hundreds of third party Integrations is that those types of Integrations out box because we find that our Enterprise customers expect that and and it's and it's a prereq given the scale and the heterogeneous nature the environment but you were gonna see more more this vision of being together from idea all the way to running code. And then more more that feedback loop coming from our our production side of the house.

All of that time together is going to be key. I think just large scale Enterprises are gonna not want to do those Integrations and spells they're gonna want vendors that can get help them with that. And again, this is something that we are certainly into The other area and I talked about this earlier.

It is one that I could not be more excited about is is ultimately the role that data is going to play and again, I I over generalize here, but I think it's actually pretty fair is that you know, we have historically as an industry that you buy a product that I buy a new testing product. It comes with its own. Domain specific data warehouse that comes with its own reporting some of its own dashboarding and that sort of the extent of it so I can suboptimize within that one domain.

But if I want to 360 degree view of the end and business process, it's really really impossible in many organizations. ai does and some ways that the AI and digital that AI but imagine a world where we can democratize the data not just on the development side, but also on the production side bring that data all into a universal data like not just a data from the tools that maybe we provide or or another vendor provides but all vendors across the heterogeneous environment and then we can use that date. It's not only give that 360-degree rear view mirror perspective so we can identify risk earlier and see opportunities for improvement.

But what if we started to use the fact that this is very repetitive data, there is a lot of data here and it just is Taylor Made From machine learning and a and we can start to apply those those machine learning models to predict the future based on the past to be able to yes, of course have Dora mattress, but what if we had change risk prediction where we can predict whether or not we're going to have a negative impact on production and thus the door metrics through machine learning or yes, we have flow metrics, but what if we could predict how would accelerate our flow Based on data? Where should we be focusing our attention to drive and continuous improvements? It's really interesting.

We as software engineers and developers are the ones that are building all machine learning and AI models that exist in the world and we have not, you know for the most part started to use Ai and machine learning on how we work. And so there's this really great moment in time here. We're starting to pull all of that together and and democratize the access the data again both on the production side and on the on the development side.

It's a really really exciting time and like many other parts of the business whether it's supply chain or it's Finance or other sales Etc. I do believe the Enterprise that are you there data in information and that information into insights that are driving business outcomes are gonna be the ones that have a huge competitive advantage in the next decade and then lastly and I touched on the earlier I think for many of us that the broader category management. Maybe it's been a little bit Fusing and and but but ultimately I think of a very simplistically right to me value stream delivery is about how we build software and value stream management is is that software that we build valuable and the valuable is obviously that word can be can be a complicated complicated one, but it's an inch.

It's a very important one to highlight right what we're doing anything, you know, we're making changes to a system whether that's an internal application or a customer facing application and we know when we make changes the end results either get better or worse performance is either Better or Worse quality is either Better or Worse customer satisfaction is either better or worse and so on and so ultimately starting to be able to not only Upstream have those those discussions so that we are very much online and and as I have that early kind of business software integration, we're very much aligned on what the outcome we want to drive and then just as importantly having the Telemetry and the analytical views to know whether or not that has happened I think about Than in my earliest days as a software developer, I would you know get a new user story or maybe a defect I would I would pull up my work management system. I pull up the code. I do my work I'd make sure the build ran the test ran and you know, everything was checked in and I closed the the ticket and I for all intensive purposes pack up and go home, right?

I had no real visibility to when the heck that was gonna get to a customer and I certainly didn't know when the you know, whether or not on the flip side whether or not when the customer did get it whether they were happier or less happy whether in some cases whether or not they even clicked on the button that I created or the new Wizard that was was added and so we now live in a time when all that Telemetry data exists and so it's super interesting not just and again, I think it's important when I say democratized data is yeah breaking down the data Silence of the past, but it's also making sure the data is not just kind of pulled into dashboard. Just for kind of management or executive levels. It needs the permeate the entirety of the team because in the end when I was a developer, I didn't need that to go to my managers manager and things into my manager and then to meet I needed access to that data to say hey Derek remember that that change you made a month ago.

It's now in production. Oh great. I now I know and then ideally, you know being able to look a week or two later or depending on the application, maybe an hour or two later and to be able to determine whether or not that was good.

And then of course applying even broader approaches to things like, you know, Progressive deliveries and feature flagging you know, the thing that would allow me to run multiple tests that wants but that's are great. But you have to be able to measure whether or not a is better than be or be with better than I and so forth and so super exciting I think from my perspective that this topic is becoming more and more prevalent across organizations. And again, whether you call an outcome based delivery or value management or you know, there's a whole bunch of different terms here, I think The the driver is is very much very much the same and then the one I think it's important to note.

And and I think this is always at least my perspective is always been. Look these things are not binary right this is not like we're gonna wake up one day and say tomorrow. We're doing value stream management or tomorrow.

We're doing all augmented intelligence and AI it is ultimately a journey, right? And so like all revolutions. This will happen very evolutionarily when we do now and we have a conversation five years from now much like I kind of generalize the two they're really Transformations and the 2010s and the age of Automation and devops and scale.

The reality is there were incremental adjustments and changes that were happening every day every week every month and we see that to be the case. It's one of the things that we really want to encourage organizations to say you could look at all of this Ai and management and you could be on the left hand side of this this slide and that's okay, right? And because really great thing about this journey is that there's value to be derived real business value real technical value to be derived at every step of this journey.

And you know, it's interesting for us. We build tools to help people obviously. continue to mature their software development and delivery practices, you know across planning and testing and devops and and security and and so forth, but we're also a development organization and so we are Somewhere on this on this maturity Journey as well.

And so I think it's it's been actually really rewarding thing. We ought to use our own tools to drive this and so we we kind of get to learn along with our customers around the opportunities and challenges that each one of these steps along the way. So before we before we close out, I do want to kind of hit on one more topic and I alluded to this earlier but I I really started to start to think about the way as a industry as the way we think about our our way.

We deliver ultimately the way we deliver software and I think we have spent a ton of time over the last 20 years in the first one which is how do we build software? Right? How do we plan?

How do we code? How do we how do we secure? How do we releasing deploy?

And that's really really important right? I know this is a bad analogy. But but let's use a manufacturing analogy.

If I can't build the car effectively with high quality with repetitiveness with reasonable costs. It doesn't matter how amazing the design is and how great the marketing is, right? You got to actually be able to build it or I can't build the iPhone, you know, effectively and assembly line.

It doesn't matter whether or not the man is really really high. So there's a prereq you have to be able to do that well and day to play the key role in that second to No it just because I'm really good at building. It doesn't mean people want to buy it or use it or engage with and so we've got to marry these two different but highly related optimized how we build software but then ultimately optimize the value that that software delivers.

And so we think about it in a few ways. i told this but regardless we need to be able to have we want to get good at how we build software. We need to have really good end and for planning all the way to deployment with testing and securing and all of that in between we need end in analytical views not just within each of the circles on this side, but overall the entire circle and we need to apply AI Solutions on top of that and so we need to be able to and this is actually how I look at the world.

a remember we serve the Enterprise the first thing we did when we Unified Our toolsat is to go end-to-end Integrations and then to pull out the data from all of our tools. Yes each tool has its own operational reporting but we need to be able to have Enterprise reporting and so we we put this data Lake in the middle and then some real bi not just asked what real bi and eventually ML and AI on the right hand side, but then we said hey, our customer is the Enterprise. And the reality of the Enterprise is they're never gonna just have us and this is a frankly a limited view of the logos.

That would exist is really probably if not over a hundred nearly 100 different kind of large pliers that exist across this tool chain. So we said hey, we need to knock this through our own tools. We need to be able to do it for all of the tools that gives people that that 360-degree view and again, not just the testing data not just planning data, but what happens when we bring those two data sets together and all the other ones, how can we find correlation and trends that that are kind of between the between the practice areas if you will and by the way, not surprisingly that's where a lot of the magic both the outside a positive returns, but also the outside is negative impact of not doing certain things along the way so we say, hey we need to pull that together and do that across across ahead of Genius environment.

And then we said it can't just be development tools. It's got to be production. So we need to be able to bring in use your telemarket tools.

We need to be able to bring you. Sentiment tools we need to be able to bring in customer success and support and Erp and Cloud cost all of those things need to be brought together and unified into into you know into single single perspectives along the way and so you get the idea here of the of the direction that that we need to be able to go. So unify left to right unify across the data expand that data set to include all types of heterogeneous environments and then we and I think others will apply machine learning in AI to that that really tuned in data set to be able to drive these these outside impacts.

And so ultimately this allows us to to create these environments where we can where we can ultimately optimize not just how we build software. There's a example I won't go into all the details here, but then we can also we can also optimize Business value of the software that we are are ultimately building and so this was a really good example, I of course sanitized all the data, but but this was an e-commerce customer they are saying hey we want to reduce the Bandit card. The very reasonable thing to do as an ultimately a business goal, but we want to do it without on the left hand side of the screen we want to do it without causing problems.

We sell other priorities. It's not the only priority so we don't want to reduce abandoned carts and have a lower NPS score or we don't want to reduce the band in cars and and increase customer acquisition costs, which that one would be hard but you get the idea. So they're able to look at the kind of stuff business goals and the kind of core one that's right in front of them.

And again, we see folks starting with things as simple as okay ours to drive that alignment and and the ability to measure the key result others are going you going further than that, but ultimately we are doing this in production today so as much as I'm kind of defining this is the next 10 years of Software development at delivery. A lot of this is live today. And in many ways your your competition in many ways.

Your local companies are likely already on this journey. So to summarize that again, I appreciate everyone taking the time hopefully folks found this these observation to be interesting. Hopefully you agree with most of them.

I'd be shocked if you didn't agree with some of them and that's okay. This is about having that that type of a dialogue but when I think about the things that are gonna really ultimately Drive the next 10 years, I think it's really about that end and looking at the tooling as enabling and business process of software delivery. I think you're gonna see people move away from sub optimizing and the boards optimizing you're gonna ultimately see the role that data play to be extraordinarily critical in in everything that we do at the practice area amongst the practice areas leveraging software development and delivery tooling but also production tooling and To start to get to the point where we can really really not just automate and optimize but predict the future based on the path.

And then lastly I think is a challenge for all of us. Doing these things that in general is hard doing these things at Mega scale and Hyper distributed environments is really really hard. And so there's a set as always it's about people process and pools that to an Empower those individuals.

There's a set of opportunities there to make sure that we're doing that. I firmly believe in this the latest and greatest tools and the latest and greatest approaches cannot be reserved only to the new development efforts because as much as they get 95% of the press, they ultimately in the Enterprise typically result in five to ten percent of the investment. And so we need to be able to of course empower the future but we need to be able to backport and and apply these best practices to the 10 15 20 30, maybe 50 60 years or the software archeology that has been built up of which yes, I may be interact with the application through my mobile phone, but that mobile application very likely is hitting a Mainframe right is very likely hitting through layers of composite of course, but but Get the idea.

And so we need to be able to do this across the Enterprise and to do it at the level of compliance and governance and security and all of those things that the Enterprise expects and so with that I Think I've reached my time again. I hope everybody got something out of the discussion. I think it's an extraordinarily exciting time to be in this industry.

I've had the opportunity to kind of live through a bunch of these ways both as a truly vendor, but also on the other side consuming tools and building out sovere development teams. And and I think it's a really really interesting kind of collision of a set of megatrans that I think is gonna really create a lot of opportunities to accelerate, you know, ultimately go faster and innovate fast and run test faster but to do so with confidence and with Clarity and with proper governance and all of those things, I think we can have both of those things two things that sometimes we're at odds with each other right? We wanted to go faster than that.

We had that be willing to have more risk. I think we're gonna be able to kind of balance that risk and speed a little bit better through a lot of this automation through a lot of issues in the day that connectivity and visibility on and on and so again, thanks so much for the time. I hope you enjoy the rest of the presentation and look forward to chatting more soon.

Thanks so much.