very much everyone um so yeah I'm going to talk about my experience using make um or formerly integromat in bu in a business operations team at onto um and my background is in um operations customer operations and Leadership um and really my focus of the last 10 years has been on building the glue to solve um C customer and operations problems I'm going to give a little break in the middle for questions may not be many but I I'll check and then again at the end but if you have a burning question just like raise your hand and I'm happy to answer um hopefully I'll go into enough detail so you get a grasp of what make provides and what we've done with it but um if you want if you want to grab me afterwards then please do right so on two what is it on two is an electric car subscription um but it's not like Zip car so um you don't use it for an hour and then give it back think it's like your own car as it's an alternative form of car ownership so you can put your baby seat in the back and you can use it for as little as one month and as much is kind of two three years and you can switch cars and um it offers loads of flexibility to to customers and the chance to see what it's like to drive electric um and so here's an example car um it costs you know what 500 and something quid a month and we also bundle in all of those um Services you need so um Insurance um servicing and repair even there's a charging option included so that includes some charging networks um that mean you can get free charging and um that includes fairware and tear right no not damage which we're going to talk about today but fairware and tear um and we work with a number of we're a tech business really and we work with a number of suppliers um Automotive companies who will turn around and repair those vehicles at the end of life so when you return the return your vehicle or switch um we're going to rely on a number of different companies uh third parties to to deal with that that vehicle and the thing I want um you to hold on to is that we're managing a customer who's a subscriber and an asset the car and those things aren't always joined up or at least the automotive industry didn't really have a concept of a customer it only had a concept of a car um uh so some as I said some charges are not included um this uh I swim every Friday morning um in Hamstead I drive from West London so um and this is The Junction at which I had a crash um uh there was a lot of traffic lot more traffic than in this screenshot um and I did The Honorable thing and and made room for another car to come out here and um reverse into a van um and fortunately I I I caused minimal damage but um I'm going to get I'm going to pay at the end basically the end of the subscription when the car goes back that's when I'm going to pay um and so that's damage exceeding Fair wear and tear but other ad hoc charges include accidents excess mileage so if you drive more than however many miles per month and PCN soof for if you park in the wrong place and Council um finds you we don't pay that fine you pay the fine um and we so at the end of life we we authorize and pay for those repairs and then we Tred to recover the spend from the customers who who owe us the money and before make that process was a mess uh we we're working with three different suppliers Who provided three different types of data each um each each vehicle um that had damage on that was custom attributable took about 45 minutes of pure admin to to to to manage I don't mean a value ad admin where people who know things about cars decide what is damage and what isn't that's real uh inside and work I'm talking copying data from one place to another um or other calculations or trying to scratch your head to work out what customer should be charged and customers were taking it was taking about 20 days on average to notify customers um we had a 30-day regulated deadline to get that information to the customer so 20 days is not great and risk us not even be able to charge that customer at all and our recovery rate on a spend of about uh 200k a month um was about 20% so very very low um and I'll I'll highlight now that um in other businesses that I've been uh part of fruit doesn't always hang this low but uh it was when I arrived a year ago um and i' never heard of make um a year ago so um but I I was introduced and within uh I think within three weeks we were a customer and and then within two months we' rolled out um what I'm about to show you um so we we we the building blocks of of this invoicing process for an ad hoc invoice was to um create a ticket saying this uh this customer needs to be charged this amount um for this damage and then we're going to um email them them which was being done manually um but we decided that an automation slice was to to build a zenes ticket exactly as was being done before but enriched with more data um send that information to customer IO an email pipe and um tool that provides logic to to deliver emails in the way in an intelligent way and then stripe for invoicing um I'm sure you've come across these tools before um and it's really all of these are sensitive right so if the information is accurate um that's important you want to charge the right amount um you want to communicate that in a timely and sensitive way to the customer and then finally um charge them in the correct way um so our job was to link the car to the customer inform them and recover that that spend and this is um this remains um 10 months later the building blocks of this process excuse me but we've also built a bunch of additional layers that help us to manage this process so I'm going to give you a quick example of how we actually create the ticket um this is just the the First Slice so I'm not going to go into all of them because it' be too much detail but here on the left hand side um we're we're arranging the data so um every uh damage instance has multiple lines of damage per car so it's not just one simple cost um there are several columns that um influence what the customer should be charged and whether it is attributable or not attributable and so that data needs to be worked and that's this is where it's being worked um and then we're matching the car here to the customer um then creating a ticket and then finally adding some value by creating a log and um uh matching with historical tickets that may be of relevance to this to this instance and may cause the customers to dispute for example and we're filling about 30 fields in this sess ticket so it takes us no time but it used to take a lot of time to do that and with this Base information the team then can then check the evidence so this is the value ad piece where the the guys who know about cars or the people in the team who know about cars can check whether this customer really is attributable and whether there's evidence to support the claim we're then sending in HTML enriched um email notification um my colleague Laura actually created this table with um in my team and she received this damage notification yesterday from the company so she's been done by her own work um which she she was pretty pissed off about um uh and then finally we can enrich the ticket because the the the customer service team don't actually see that email that goes out blind right by our email agent so um the ticket provides the context on what the customers actually received and then for stripe um I don't know if you've come across this or looked at the documentation of stripe but it has a number of predetermined steps you've got to build an invoice then add a bunch of line items into it um kind of say you want to actually charge the customer and then charge them so you replicate those within make um if I just run back actually make is really I thinking of it like a a Transformer it's about transforming data um borrowing other bits of data from other places and then finally spitting out a results and that may be in another tool or it may be in the original tool but enriched with the data you've picked up along the way each of those um circles is a module and once you've set up a connection so zenes for example you add credentials you want administrative credentials rather than your own ones then you're going to perform a series of actions um you one might be create a ticket um and that's um that all happens via AP make controls those um API through Al of the Box Integrations and um you can also do custom ones yourself if you look up the documentation on how an API works you can then build your own Circle and interact with your database for example um and that means you can do lots of kind of cool things um and then data packets get sent down from the left to right and if if something makes down to the end it means usually something the output has been generated so for example with the multiple damage line items we grouping those by registration so that that's only gets treated as one event rather than charged eight times I'm going to show you the impact we had we we we took about a month to scope it out and and um get a stakeholder Buy in and and then we spent a month building and um and then released it there was some internal resistance early you know people adopting new systems always is going to be and um I felt like the quicker we went the more likely we were going to be able to overcome that resistance um rather than um listening to every Grumble um always find a champion and then just rely on them um and we reduced um the manual workload by 500 hours a month which is you know um two and a half people's work and that meant that they could do a lot more interesting things um we reduced the time to notify the customer by by half and recovered about 45k a month more so we think the net benefit is about somewhere between 500 and 800k of ebar a year um and this is you know a month of build um uh and what's great about that process is that if we scaled to a higher Fleet volume or a lower Fleet volume this process can go up and down without having to hire more people um or or or fire people if if there's fewer you know um so we saw that as a real positive um and people started to love it I'm going to share now um three insights so that's the case study I'm now going to share three insights about what it was like to work with make and what's worked best for us like where where have we generated the most value for for the business um and um it's just going to be simple illustrations and a little story so um the first Insight is that it's not just about the speed of the automation itself that's exciting it's about the speed of learning or has for us it's been about the speed of learning so one example is um we we charge customers automatically after seven days of informing them that they're going to to be um CH uh facing this bill um which they don't like um but what customers like even less is money being taken out of their account or some customers really don't like money being taken out of their account even if they owe it so we we saw this feedback and we thought okay well let's we can't change the way we build everything but we can uh provide an experiment that offers the ability for people to actually actively pay so to pay now um which counter which which for us was counterintuitive because it was seven day it was going to happen immediately rather than in seven days it's bad for cash flow for the customer but some people just want to know what money is going out of their bank and when and so um without bothering the tech folks are onto we we built an experiment that mimicked paying now it was just a button in the email that said pay now and then an Automation and make that ran every few couple of hours and then charge the customer that that amount um we had a 20% takeup rate for that pay now button which we thought was pretty high um and that meant better cash a better cash flow for us fewer disputes and better sentiment for customers and so the point um that I'm trying to make is that before putting something on the backlog for your core Tech infrastructure no and low code tools allow you to rapidly validate or invalidate your hypothesis and then your speed of learning increases which means you're going to get better and better um second Insight is that not all automations are born equal um so a dishwasher looks like a great automation right you put the dishes in and then um get clean dishes afterwards and you have to to wash any up um but um I don't think a dishwash is actually that great in automation because you have to load it each time and how well you load it really matters to how how well dishes come out afterwards and what I found the same with make automations is that if you if you load bad data like bad stuff happens um we had three different suppliers with three different types of data so we we took the time and I think I would repeat this to basically fight with our suppliers until they would provide the data exactly like we needed it uh and it even better is to get this straight from them VI AP API but a lot of these Automotive companies don't know what an API is um so you you have to deal with what you've got if you can get a data dump that works nicely do not copy and paste don't ask anyone to copy and paste in your team things will go wrong immediately um um so all the team are doing manually is just replacing the spreadsheet from from a Daya dump each day um which is pretty straightforward um and um yeah I think rethinking every step and especially those early ones otherwise you end up with dirty um and then the third Insight is from Aladdin um um it's that make is a great way of creating timely actionable insights um Aladdin in this story Aladdin has become Prince Ali through J gen's magic and he's trying to impress Princess Jasmine as he Wales in on like a huge parade of elephants and but he's too Brash and then jeie intervenes just the right time um as a bee in this case issuing a Mayday call mayday mayday and said says uh just be yourself um and love is restored and um he shows her the world on his magic carpet um and my point that I'm making here is that people do build amazing dashboards and I love dashboards um but in my experience senior people look at dashboards and no one else looks at dashboards um and so the senior people get a full sense of control uh where chaos is actually happening um so um all the data is there and yet you know the expected result isn't happening whereas what mate can do is it can you can you can stitch together information about from different platforms that's happening right now and ask questions think of what the question is produce a an alert that goes to the right person at the right time with the right context um more than any one person could be bothered to uh collate um and um prompts the the correction um and we found that amazingly helpful and those are the kind of layers that I've talked about that we built on top of uh the core modules um that's [Applause] it