there we go hey good morning good afternoon good evening wherever you are and whenever you are welcome to the creator toolbox this show is for servicenow admins builders and developers of all skill levels and that includes you we want you to be our advocates for servicenow welcome to the show uh in each episode real quick we're going to explore a servicenow platform feature around the app engine and creator workflow so stick with us we'll be doing some integration hub at some point we'll be doing some app engine stuff we'll it's it's going to be a lot of fun this show is tuesdays at 8 a.m pacific 11 a.m eastern time so i hope you can join us there subscribe like turn on the notifications you know what to do you're smart youtube people you're savvy let's get into the introductions andrew where do you begin oh thank you chuck i'm andrew barnes developer advocate here at servicenow i've been in this role for uh two and a half years and been working in the ecosystem for about six and a half years as a customer a couple of times an implementation partner and now here at servicenow and i just love building on the now platform um you know custom apps integrations and managing large-scale development deployment we'll do this in abc order so next is our man brad who just came in out of the shackleton expedition from the antarctic apparently hello everybody my name is brad tilton i am a developer advocate uh the developer program here at servicenow i have been developing on the platform for about 12 years and been an employee for a little over the last four and it is historically cold here in southeast texas and so i may i may lose power at some point during the show but i am here now glad to have you and my name is chuck tomasi senior developer advocate with servicenow been here for about ten and a half years coming up on 11 real fast faster than i'd like to admit i was a customer for a couple of years before that so my history goes back into the mid 2008 era and uh when they named the releases after seasons so appropriately this probably would have been winter 2008 at some point around here i don't know okay let's get into today's topic i'm going to turn over to andrew he's driving today so tell us what you've got on deck oh thanks chuck so this week we're focusing on the now intelligence parts of the platform uh so we've got uh some blogs and this episode of creator toolbox and then on friday our usual live code happy hour will all feature things around now intelligence when i was looking you know through what was in quebec for now intelligence um there's a couple of things that really stand out and we're going to do one of them on friday which is ai search on live code and today the other thing that really jumped out at me is a brand new uh feature called process optimization so having been a product owner and a platform owner and done implementations for servicenow and and other enterprise applications working on improving processes is something that i have been you know working in and familiar with and doing uh for a long time and know how expensive those types of uh you know exercises are and so when i saw that we were deploying something to uh you know help automate that process i i was just like well i've got to go see if this is as good as as i want it to be um is is this actually helpful and so that's what we're gonna look at today is uh process optimization um in the platform and what we're gonna do is well we have some processes ourselves so the you know one of the apps that that we build and we use on a daily basis is our artifact manager which manages all of our content generation and deployment process and i said well let's let's see if there's some things that we can learn with this process optimization engine on our own process in our servicenow instance so that's what we're going to do today we're going to create a new setup for this and we're going to inspect our own process and see if we can define some learnings um and how long that takes us i've got a bit of anxiety here andrew it's about how dumb am i so are there any questions from either of you because you've never seen this right before we dive in have heard of it very eager to take a look at it and anxious as well so yeah let's let's jump in and take a look all right yeah i'm excited to see how it uh how it's going to work i've not seen it and i've i've only read like the docs page uh so let's see what happens i'll be curious to see because we've got some stuff that is scripted logic we've got some stuff that's in flow designer we've got some stuff and we're managing in effect artifact manager has multiple processes there's a process for webinars there's a process for live streams there's a process for blogs so i'm really curious to see how it ties into all of that in and what the outputs are absolutely so let's dive in i will start my screen share so uh you know talking about first the uh you said you'd gone and read the docs so that's always a good place to start um with a new feature right is going and reading the documentation about it how to get it installed and set up and what you need to do to get rolling with a new feature like this so i just wanted to pull this up real quick and and show you this is step one go read the documentation then step two is go to your instance and you know make sure that you have a process that has um auditing turned on for the table and fields that you're going to to inspect so in our case how do you know if auditing is turned on that's a really uh good question chuck i know that there's a property to tell it not to do it uh i'm trying to remember how you tell it too there's an audit checkbox on the table in the dictionary ah okay if you turn on activity history it will turn on automatically so if you've got a table that's got activity history turned on it will be turned on audit is not one of those check boxes that shows up on the default view so you're going to have to use advanced to find it and you i don't even know if it's still on the form or not you may have to do a little digging to put it on the one thing this is intelligence week and since it's on my screen i thought i'd mention it which is we have a new box here in the list view that looks uh looks different doesn't it a little speech bubble thing yeah is this uh nlq it is [Music] nlq right there in my list view i'm going to have to get the dev program and dev program dev instances so i was on type video so it kept that i suspect if i pull that off i might actually get something uh and there there were none created in the last week um so so this this is this is roughly what i wanted it to do so you can just click this little little chat box here refresh your demo data and go um so this is really cool i'm i'm kind of excited about that um so yes as you said you need to uh you know make sure that that's uh auditing is on um and then go to um uh process option process optimization so there is a guided setup um and then if you're like me you probably uh will probably skip the guided setup because i'm not i don't follow guided setups very well we just dive manual either right into the analyst workbench did you have to enable a plug-in for this or did it was already enabled i don't believe that there is a i believe it's enabled by default i was going to ask the same thing good question um and so you launched this analyst workbench or work whatever it called workbench yeah and you create a new entry here okay for your new process now if you're using um itsm or csm they both have plugins that you can enable that that jump start like incident process optimization um already has one that ships clearly servicenow didn't ship an artifact manager process optimization um out of the box but there are some for itsm and csm processes that come with some plugins so you'll see when i hit new in that workbench i was landed here in the process model definition i'm just going to move that up there so i can see it and we can name this so artifact we don't need a short description um we're going to you know inspect a table for this one but you could do a a data source and our table is the artifact table and then you can apply any filtering here now i'll caution you don't apply too many filters at first here at this filter level there's some other places where you can configure the model definition to tweak it um so don't go overly restrictive on this filter at first you may want to get you may want to filter on what things that are completed or in our case published or potentially but you you may not want to because you may want to be so it's what part of the process you're inspecting and you don't necessarily need to limit it like that um because you can inspect different parts of the process later once the model has run it and you can say oh well let's just focus on the ones that are you know in the resolve state if it was an incident for example and you don't need to to apply a filter here to do that so this filter is telling it go look at this subset of data in order to optimize the process absolutely and so i want that to be fairly generic yeah so start with not many conditions here and then um you can narrow it down afterwards if you need so i'm going to go ahead and hit save here and we'll get our related uh lists uh down at the bottom and these two are uh ones that we we absolutely have to uh populate the no definitions and the breakdown definitions if you have child table definitions this is also where you do that but so to start with we need no definitions so let us create um and so these are the fields that we want to gather information about uh along the process so a really common one would what would you think a really common one would be the assignment stuff so assignment would be one state is like number one go to how long does it take to move between each state sure state changes are absolutely a definite required one for any task sort of uh defined process and we run all our automation based on different states so and types are important and and then for us type is going to be really important right since we have different processes for the different types so let's go ahead and define that one and i i'm not sure we would divide a lot of insights on the assigned too we could throw that one in here okay yeah it doesn't change much if it's your blog article to write it's your blog article it's pretty true now we could learn that through this inspection you know if if there was a manager coming in who wanted to look at it and this might actually help them with that um so that seems like a good start what is what are the fields filtered on obviously that's not all our fields that is a really good question there's some reference fire in there and i'm hoping it's not scripted but it probably is so if you do a it's definitely show uh a filter so let's see what it uh what it has choice and reference looks like choice and reference potentially because i had parent and type let's see of course that's really uh good script include name your name script includes very descriptively all right what do we think prom and prominent yeah that's that's interesting so uh we can go dive into that if we wanted it doesn't really jive with performance optimization does it it might have been one of those things that had a different product name a different name and yeah remote tables used to be streaming tables so all the tables are st something so the next step uh is our breakdown definitions so you're going to want breakdowns for you know different fields so you can learn things about them so i believe that i am going to grab state so these are kind of like the performance analytics breakdowns yep let's grab assign two and you'll see you have a lot more choices with these on which ones you can pick and then and how about updated by or something maybe updated by there might be something useful there and i'm i'm really just throwing some darts at the wall with these i'm not sure which ones are really relevant for us right now um but we have now at least some breakdown definitions so we can return back to our model definition so we're on the artifact table we've got a couple of breakdowns and a couple of nodes defined which is what we need to get started so you can start with you know a sample set to see what the outcome is to see if you need to tweak some of the filtering you know define a different breakdown definition or apply some filtering to one of the breakdowns our data set is small enough we can probably just say go run the full the full thing right now so what this is doing is it is inspecting both the record and the audit table for all of the things that we have specified so if you have child tables defined it would go and do the same for them um and then once it collects all that data then it actually you know builds the model based off of the the data that it is just inspected um so i don't think this will take too long but while it's going um let's go take a look at something else hmm that's when i i meant to take a look at the nlu because i can get distracted with this really real easy yep i did i did take a look and i do think you have to enable a plug-in to get this to work oh great um let's see ross says optimization which makes sense let's see if we can get the use process optimization and set up there we go yeah that should tell you here's the plugin activate the process optimization plugin and it does note here what i uh mentioned which is you can load the itasm process optimization content pack and the csm optimization content pack which you know don't name these the same same way come on [Laughter] i don't care which way it's named just name them the same way hey look we distracted ourselves long enough uh that we can now uh go open our analyst workbench to inspect um what we just got from that uh that model definition and inspecting all of our tables um so this will be an entirely new uh view that we're able to uh take a look at here and and and i'm excited to show you uh because this is really cool um i'm like i like it a lot uh just the way it does the visualization uh is pretty neat uh it is taking it a hot second to load come on you can do it you want another distraction no okay there it is there we go okay what is this i've seen screenshots of this help us understand what we're looking at absolutely so we've loaded this in our workbench and this is a really really condensed view of my model right now i'm going to you know close out some of the distractions here and you'll see that right now this doesn't look very useful does it not much of a problem this doesn't even look like a process what's going on here and the answer is under this uh this little refine is it starts out with no nodes and you have to add or no arcs and you need to add them uh so it collapses these arcs so it's not always based off of yeah okay the that how many of these things took right and so you've got here you can see we've got states type and assign two and what were those it from our definition what were those nodes or breakdowns well let's take a look those were our nodes oh your nodes okay [Laughter] so that's what we'll be able to see on here is all of the different nodes um and the path between those nodes so if i want to pull out one of them for example like i want to pull off the um assigned to the assign two for right now and i have to remember where to do that which is it's got to be one of these filters from that's a breakdown that's the breakdown i need conditions uh maybe if we just i'm trying to remember where to do that um and it's not just not the conditions under advanced filters did you try that one [Music] underwear uh advanced filters conditions on the left did you try to try that one yeah that's what i just clicked okay and this is to filter the data set so this is why i said don't be overly restrictive on the the model definition because you can then narrow it down here inside uh of the analysis tool um what happens if i just yeah that's going to leave all of them i want to turn off one of these [Music] there's a way to do it i'm sure i just don't know how but that's okay because we're learning um you know this is this is a brand new tool but let's take a look at what um this can give me um actually there's a transaction filter transition filter i can trans get rid of yeah let's do ready to work and needs research so this is allowing me to add some nodes that i want to follow the transition between these so like i could go from ready to work needs research to published and apply that so what this will do um and rightfully so is none none of our things went from ready to work to need research to publish no i don't think we've used needs research ever we may want to consider deprecating that state but that's an interesting insight is hey you're not using some of this stuff that you thought you wanted so let us just take out all right so here are all of the things that got to the published for example and we can we can now take a look at what some of these paths are to get to that and see if there's anything in these that we want to you know learn about one of the really handy things let me collapse that side panel is right here in the primary metric we can do things like say oh take a look at the average duration of these transitions and we're going to want to oh wow that down a bit there we go so sliding this arc thing is really useful because it you know starts to group like transition um branches so if i read this right after 471 days something got a sign from dave slusher to nobody [Laughter] that sounds about right right there on interstate yeah right here this is fascinating stuff and you know if you say oh i there's something interesting going on here and hopefully it'll be more than one occurrence you can say oh well we'll go look at that record yeah or records poof and then we can go uh inspect those in the list view um so let us take a look to see if there's something on maybe the total occurrences helps us um find a path that we want to inspect with something else [Music] so we've got 42 go work in progress and then hmm there's 64 that jump right to ready to publish that seems uh that seems interesting doesn't it like yeah we've got work in progress and ready to work and there's more that go that way than there are that go to work in progress yeah so that's an interesting insight right there right and then check this guy out 218 write to published go right to published oh you know what that might be is the imports that i was doing because i just imported a whole bunch of stuff from another from the community live stream app and the tech now apps so there may be something in there uh potentially and we could uh if we could get the zoom rate we could go with our show where is our short record that's what i was looking for oh it's only on the transition yeah not on the paths that actually makes a little bit of sense but uh it would be kind of nice to have it there let me zoom back in for you um so there's immediately some insights that i'm getting just from our fairly simple process right here going oh well what's going on there why why is that and can i learn something about these records maybe there was um you know a type or something that uh was was causing you know you know maybe if i filter to blogs i'll get a different insight versus live streams yeah just for joking yeah can we do a type webinar and see if that number stays about the same or filter it out and see what happens so we could do type um we still want state and we want uh conditions so we could just type is and you said you wanted to narrow down to what webinar webinar well that wasn't our major offender yep this is not the major offender uh over here in this club interesting because a lot of those records were created as published um but this may not have all of your uh because this is a oh right right right fairly right yep old clone here's another caution [Laughter] when you clone are you bringing over the audit histories that is absolutely required if you want to test this out in a development environment um yeah and i did make sure that i cloned with my audit histories um for this purpose otherwise we wouldn't get very many insights yes not in debut won't oh we seem to have lost brad he must have a rolling blackout um so we have no blog series because uh we don't do blog series we do blog posts and we can it's interesting because um brad's ui builder thing probably should be a blog series so we can take a look here and see that we have a lot going straight to published 47 here we get 32 of them ish over to work in progress and then they move ready to publish so we don't get a lot of ready to publish for for blog posts so this is the kind of thing where we could actually start to remove some of the state transitions for our types to make it easier as the consumers of this to to flow in the way that makes sense for our process right so there's two things you get to learn from an inspection like this is what's happening and how to direct it the way you want it to be right so you get inspection into what's actually happening and then that helps you have insights on how to adjust the process to be more like what you want it to be i'm curious what end represents we don't have a state called end um start and end nodes uh here don't represent anything in particular okay everything has to end up there everything has a start and everything has an end okay all right and and over on the left you've got view as graph right next to the board yes so this is the breakdown here so we can view this in a graph type or in a list view type so if we switch over to like [Music] so you can see the different breakdowns allow us to and you know a little graph yeah a little graphic eye and we can apply a filter right there from this breakdown which is which is very nice so i can say clear the filters oh that's enough and then and then come back here to type so this is a different way i could have filtered blog posts it said oh there's you know 60 routes and 178 of these go apply that filter and then i can you know inspect this uh process for blog posts differently than the other ones so there are um you know some great insights that you can do with average duration of these and oh my goodness to merge them again yeah probably one step up so what would be nice to me is if you know i was able to sort of tell it you know give me like these are are clearly uh you know things i have to adjust frequently uh when i'm switching the view um but i feel like i could you know say be similar to this one no matter which transition i go to um you know of granularity but uh or remember what it was when you come back if you're flipping between them saying that that would be awesome that's how i like to see this particular setup and when i come back here leave it that way so if we take a look at this one um so this is uh blogs and we're doing average duration of the transitions here and so we can take a look to see that you know we we spend a lot of time uh in the in the ready to work state here an average of 79 days apparently you know the big black line on the right is assigned to nobody uh yeah that makes sense um the transition from ready to work to ready to publish is uh very short right um and let's take a look down here um what can we see yeah so i really i need to figure out how to pull out the the nodes i don't want and i'm just not familiar enough to know how to do that those aren't all yeah we want clickable right say turn this off yeah from this node off yeah and i don't know how to do that um so there so we've been inspecting these uh you know transitions from the the graphical representation here but you can also take a look at them here in this this transition expander allows you to you know try and find these you know numerically um oh look 26 of these and then it will allow us to filter the whole graph based off of that transition and inspect that so now we've narrowed it down a blog post we've narrowed it down to the ready to publish to publish transition and we can inspect just that part of the process because you're you're not going to learn very many insights from you know the the topmost all of the records all of the transitions possible you really need to drill into uh some of these specific scenarios in order to gain the insights that you need to adjust your process um and these these tools like i've already so i took a look at this data um earlier and i realized um what i said earlier which is um our our states some of them are never getting used or almost never getting used depending on the the uh the artifact type and so we could just remove those as options um to make it easier to use i hope peter computer audio and i believe there we go yep sorry about that everybody um we lost our zoom connection for a moment but we're back here looking at our process and learning how to improve our process with the new process automatic optimization process here in servicenow in quebec and um i had just applied some filter transitions so we are now filtering on blogs and inspecting a particular transition path which is the ones that were moving from ready to publish to published for blogs i've added a secondary metric um so i'm looking at average duration and then total occurrences uh and so you can see you know how many occurrences are going up here we've got you know 27 days with one occurrence um you know 10 days with four occurrences and what you can see here is that this is really allowing uh for that that process owner in engineer team behind a process to inspect this data and make some process optimizations based off this so what do i do now that i've narrowed it down to a visualization that you know is is revealing uh something important to me which is uh we don't need a lot of these states um for this process so we can uh you know put in a request to fix that so what we can do here is we can say create an improvement initiative ha ha and we can you know tell it which business service and you know what what we want to do and how important this is um to improve this process and we can see these improvement initiatives uh here in our process optimization workbench area there's actually a little little guy right there to pull that right in while i'm working here so i can see the the current initiatives and filter them and find them right here so you know if a different process owner had already realized something i should check to see if someone has already put in the one uh that i want to do uh before going and creating it and we haven't created any yet so there there weren't any in the list but that is like step one is make sure no one else has already suggested this improvement those are tied to our table or how does it know which ones to show in there that is that is a question i don't know the answer to um so let us answer the things and save that and see filling in only the mandatory fields i love it hey they're mandatory for a reason right um and uh um there we go yeah so we've got this one here uh but i'm not sure uh all linked ones and the current model so uh okay so when i hit create it it does uh link um from the model definition that i was working from uh to the initiative so that is handy well that is cool so with that chuck is there anything else that you wanted to take a look at on the screen or ask about um i know there's a lot to dive into here um and there's there's tons of tons of cool stuff to inspect um but that is what i wanted to highlight uh today on creator toolbox so the the question i've got to relate this back to what people are doing every day is developer is going to be building the models process owner or process engineer or business analyst somebody's going to be viewing these maps and and doing the analysis possibly in creating these improvement initiatives we saw it from an artifact manager perspective i could easily see i i'm curious to see what the with the itsm and csm ones bring up in terms of hey here's some activity packs to help you get started but we won't go into that today what would be some common next steps for somebody who wants to take a look at this or you know some their boss says hey i heard about this process optimization stuff yeah so that's a good question you'll need as we noted earlier you know if you want to try this out in a subprime instance you need to make sure that that subprime instance has been cloned with the audit tables so that'll be the first step is make sure you're using an instance that has got um audit table entries you need to enable the the plugin and then you need to set up the model like we did today and so you can get to this step which is actually inspecting it so those are the those are the steps to like get back to here and then once you're here you can what i would do you know if i was still at a customer is i'd actually take my most recent you know our own analysis of our process so like change management for example is the thing that we we're constantly improving um and take the take the learnings from our last one and and see how it compares to what this engine gives me like did it give me the same insights do you give me different insights and compare that to what we had been doing for our process uh you know improvement without these insights and see how i could have defined the same things and if there's different learnings that i could gather from that so that that's how i would get started and choose a process um that has some good data in it so itsm and csm demo packs um those come with demo data and you can use those in case you don't have like production data so choose choose one of those two are the most likely ones that are going to give you enough data and enough process interactions to show you something about how this tool actually works could take a while to build those out if you've got obviously if you've got lots and lots of change requests that you want to show people uh it could take a minute yeah um but that's uh one of the reasons why there is that uh content pack for things like change is there's already you know so the servicenow team has already put together a model for you to use against change so that really should help you bootstrap to applying that to change management for example because they've already set up the model for you to just run against your data and get some insights quickly yeah if you've been a customer for a few years though you might have a while to wait before it analyzes all the audit histories from all your past 100 million incidents that have come in uh that is that is very true another question that i thought of is uh performance analytics is also a way to analyze your processes as you're supposed to put you know your your indicators on those decision points in your process and then you can determine are there process optimizations are there any tie-ins here this might be a heath question that i should have asked him on the break point that's coming out tomorrow there nice plug for break point so uh i don't know if you've got any thoughts or insights on that as to how performance analytics would work alongside this or what you get out of each instead of yeah and i don't have any insights onto that but i would think that you know performance analytics is is how i was trying to optimize my processes before and this seems to be a more powerful inspection tool because um i didn't set up the granularity in performance analytics to be able to uh track it at the all of the the audit steps and transitions along the way um i could have but i i never did for any of my processes set up the granularity that i'm able to get with this and the speed um just because i don't need those uh breakpoint you know those those data sources most of the time i really just need it when i was optimizing my process and not for the running of the process um so i didn't set those up because i like to get dual use out of things sure um but but i see some similarities there um and this just allows me to uh visualize these things in a different manner yeah and then pa did um and uh with that note on uh so you triggered something in my brain when you said uh you know if you're a long-running customer uh you're gonna have a lot of data mm-hmm one of the important things there is when i said don't be overly restrictive on your filter at first for your model i will actually rescind that a little bit for some of those large tables and say if you've changed your process at a time period that you know about narrow down the filtering of the model for that window what about just saying last hundred days or last the last 365 days something that would make it a little more reasonable even if they haven't changed their process in four years and they've got two million incidents out there they like to look at that do we really need all two million of those to get a good view i would think not i would think something like depending on your volume something like six months or you know the last year would be the right amount um and the thing is is you can you know build the the model based off of you know being a little bit more restrictive uh to start with and and making sure you're getting the insights you need and then open it up if you're not but something like six months for a every day used process like change management and incident should give you enough insights to to analyze your process and move forward from there um so uh with that um i was so excited about uh showing this off and playing with it and i've i've i've created a couple of stories for our backlog to to make our process uh better just from the you know little bit of time i've been able to to inspect this and and i want to to use it uh and learn a little bit more um how i can you know get more analysis out of this easier but this this is a cool tool and i'm i'm excited about it and it actually does look like it's useful like i gained insights that i didn't have into our own processes um you know within minutes of of running this i hope that our customers get the same instant gratification that we did as well thank you andrew and thank you dear viewer we'll see you again now next tuesday there's no creator toolbox so the 23rd of february we'll be doing a tech now so if you want to catch that it is on workspace configuration jeremy duncan's going to be leading us through that so meet us here back at the same time 8 a.m pacific 11 a.m eastern we'll see you there if you want to register for that bitly tn 84 reg tn is in nancy or tech now get it episode 84 i try to make this simple for me as well so bitly slash tn84 rage if you haven't already signed up hope to see you there otherwise you can catch it a few days later on the community youtube channel and we will cross-post it to the developer blog as well so until next time everybody take care be safe and we'll see you again soon you