Chalkline helped Pemberton deploy Microsoft Power Apps to bring low-code automation to their portfolio management system.
Matthew Stibbe:
Hello, I'm Matthew Stibbe from Articulate Marketing and I'm here with Harry Boyne and David Smetana from Chalkline, co-founders of Chalkline, and Jacob Myner, who's joining us from Chalkline, where he is a SharePoint consultant. And David Fanning, who is Head of Technology at Pemberton. So, those are the introductions. The purpose of this is to really talk about the exciting work that Chalkline's has been doing for Pemberton but before we get into that, David, tell us a little bit about Pemberton and tell us a little bit about your role there.
David Fanning:
Okay, so, to put it succinctly, Pemberton is a leading European credit manager, backed by one of Europe's largest insurers, Legal & General and essentially, we provide a number of financing solutions for borrowers and investments providers to institutional investors. And my role is head of technology development and my mandate, if you like, is to oversee the application of technology to improve the firm's effectiveness and efficiency as it scales to meet its ambitious growth plans.
David Smetana:
I mean, it's quite interesting because we've been working with Pemberton for... Is it 18 months now? Is that about right Harry? I can't remember. Something like that.
Harry Boyne:
That's three and a half years, but...
David Smetana:
Directly. And we've been on a bit of, I guess, a journey with Pemberton. So, that's been leveraging a lot of the Microsoft Cloud stack. So, in the backend there's Microsoft 365 now, we've got WVD, we've got SharePoint, so there's a lot of new technology there. I'm interested because there's a lot of opportunity in applications out there in the marketplace. So, your role is, I guess, to look internally at the requirements and try and map those requirements to applications that are out there. I mean, we've worked on and completed a piece of work around a trade application, so I'm interested to hear what the requirements were for that application and why did nothing off the shelf fit those requirements?
David Fanning:
So, Pemberton had an incumbent trading system when I joined the firm in 2019 late, and essentially spent the next kind of five to six months trying to complete the implementation of what was a fairly poorly implemented trading system, portfolio management system. And essentially we came to conclusion that we wanted to replace and retire that system. Meanwhile, we were developing ideas for a data hub or a data platform in-house.
And so, the two ideas came together and presented an opportunity to extend that data platform concept with a trade capture facility to essentially apply as [inaudible 00:03:55] from portfolio management department, the verification of those trade details by the operations area and then the fording of those trade details to the loan admin on an automated basis.
But really, it's all around... Because we were obviously in a regulated environment managed by the FCA, we need to demonstrate separation of concerns in each of the areas and separation of responsibilities and also demonstrate good risk management practises in terms of ensuring that the information is accurate as input once it's verified and as far as possible automated end to end. So, the actual import on the loan admin side is automated, there's no opportunity, again, for any kind of fat finger opportunities. On the loan admins side, just simply received the data files that we [inaudible 00:04:51].
David Smetana:
Okay, and did you look at off the shelf software available in the market? I mean, I guess, there must be numerous different types of software that could fit these needs.
David Fanning:
Yeah, we did review the market, what we didn't want to do was get into a whole portfolio management straight trading system procurement exercise, given the time constraints that we are subject to. So, essentially, the contract was due at the end of the year or within six months time at that time, so we needed to come up with a solution that was feasible within that time period. We did talk to a couple of different providers, but again, they were fairly heavyweight systems. And really, what we implemented, in some ways, is a MVP approach of a viable product, but actually, as it turns in a lot of ways, it's 80% of what we needed it for anyway. So, it's a fairly efficient and effective solution.
Matthew Stibbe:
And relative to the others, presumably cheaper?
David Fanning:
The in-house option is, yes, cheaper than what would be a fairly large implementation, yeah. I'm not sure it's that much cheaper in terms of time on our side because we think the same parties would have had to have been involved, but in terms of actually implementing it and getting up and running, it was pretty easy actually.
David Smetana:
That was one of the thing that, I think, I've heard. Is that a lot of the off the shelf stuff, if it can do 100 things, you only really needed 10% of the functionality, but you've obviously got to pay full whack. So, you're paying 100% of the cost for 10% of the functionality and usage but, I guess, what we've been able to do together is get those requirements and deliver a niche product to fulfil your specific requirements.
David Fanning:
Yeah, I think, when you're looking at products and service providers, it's always difficult to pick out a particular service because usually it's bundled in with half a dozen or a dozen other services, but really what you want is this. So, I think what we've done is built a solution specifically, it is bespoke, it is specifically for what Pemberton wanted it for and it's got the attributes that we need needed in there, et cetera.
So, yeah, we got exactly what we wanted without the need to go through market, go through that procurement process and implement what would be a much larger piece of software, and in the timeframe that we had, I don't think it would have been feasible really to get that. We did also have a look at some offerings from the loan admin, but they weren't of the maturity and the ease of use that we wanted. So, again, the in-house solution just seems more effective and efficient.
David Smetana:
I know from experience, because I've seen the amount of work that was put in from your side and from Pemberton, in terms of giving us the requirements, which I know Jacob was very thankful for. So, I guess, I'm interested to know what you felt the engagement was like with Chalkline and what that experience was like for you. Obviously, there was a fair amount of due diligence and requirements that you got internally from the business and put that down on paper to articulate it with us, but on an ongoing basis, Jacob's on the call, please feel free to be open and honest. What was that engagement like, working with Jacob?
David Fanning:
I really enjoyed it, actually. So, my background is engineering and the last few years, it's been more engineering management or management around technology teams and perhaps more business architecture, higher level. So, I've actually quite enjoyed the opportunity to be hands-on again, if you like. So, the way that Jacob and I split the work was [inaudible 00:09:23] very high level, the initial design document. I tried to put it together as if I was building it and then basically reviewed it, discussed it, check who came back with his design and the estimates, et cetera.
And essentially, the way that we split the work thereafter was, I bought out the data platform pieces. So, the database, where the data comes from where the data has got to be stored, et cetera and then Jacob built the power app implementation on top, where there was integration needed to the loan admin. Again, that was quite a good opportunity for me to try out my ageing coding skills and code the Azure functions, which actually worked quite well in the end, around that. So, yeah, it was a real collaborative effort. I thought it worked really well, actually. I found it very rewarding, actually.
David Smetana:
And one for you, Jacob. I guess, what we try to do, as a solution provider, is to make solutions as simple as possible. So, we've leveraged as much as we can from native Microsoft "no code" using some of the Azure functions, but what has that process been like for you and what's some of the... Why do you think the platforms that we've used were the right set of tools for this job?
Jacob Myner:
I think, it even surprised when we finished the trade out, how much we were actually able to do within power apps itself. It's got so many nice components there and it was certainly a bit of a learning experience from our side as well, how much that can actually extend to do. I was really happy with the interface in the end and again, like you said, it's the no code, or I should probably say low code-
Harry Boyne:
The age old argument, isn't it?
David Smetana:
It's a myth, we can debunk that.
Jacob Myner:
Yeah, it takes a massive element of it though, right? There's no messing around having to create a user interface, which again brings the amount of time that you spend on the project down massively. So many of those components were just there for us to use already. All of the data connectors that Microsoft have for SQL were fantastic, which is where the data's being stored at the moment.
As that SQL data's in Azure, it's really kind of plug and play to go and grab that data out of there and start working within the application itself. And then, I guess, we used, other than power apps, we used a bit of power automated as well to manage that connection and manage the process flow through the application itself. And again, it's plug and play, just fit the little components together and try and make it all work. Infinitely easier than creating a whole bespoke application and developing it yourself.
David Fanning:
Yeah. I mean, I was going to say, in our previous firms, if I was to do this, I've done this kind of trade capture system, this will be the fourth time in its incarnation, and this has been the simplest by far, I think. Previously, I would have had a small team of developers and database developers working together and they would have build the solution. They would have been hosted in docket containers, this, that, and the other, to handle all the middleware integration that we probably would've had to have done.
This is a much simpler stack, it's a much simpler solution and I think it's hopefully very easy to manage thereafter. Really, we built upon the Azure stack totally, we've tried to leverage all Azure technologies as far as possible, generally within Pemberton and we've certainly embraced things like the server-less databases, the server-less ETL and Azure functioning is, again, server-less. We're trying to reduce the amount of maintenance and overheads that are normally associated with producing software platforms like this and power platforms is just another example of that.
David Smetana:
And correct me if I'm wrong, Jacob, but I know there was, as we've previously said, there was some requirements in the background from David before you presented to us. So, I'm sure there was some groundwork there, but I think we completed it in about 15. Is that right? Something like that.
Jacob Myner:
Yeah. I think it was within that.
David Smetana:
Okay. That's good, that is good.
Harry Boyne:
And, I guess, one of the interesting things that you take as a given when we talk about 365, but there's considerations in terms of when we talk about it from a backend perspective. So, authentication that we don't... Of course, you need to consider it from a, "Who needs access to what?" But we're not worrying about libraries, we're not worrying about MFA, it's just there, it's just sorted, it's already done. So, that's considerably taken the amount of effort we've needed to put in out of that as well.
Matthew Stibbe:
Because you're already doing that to get onto Microsoft 365 and active directory stuff.
Harry Boyne:
Exactly that. So, hard work's already done for us there and one less security concern for us to worry about, one less penetration test.
David Smetana:
Security built in by design from the onset.
Harry Boyne:
Exactly.
Matthew Stibbe:
If I could follow up on that point, because I'm curious about this. You said at the beginning, obviously, there are security and compliance requirements and how did that impact on the project?
David Fanning:
Say certain functions where the application can only be done by certain business teams. So, we can have an operations user, for example, putting in a trade that has to be done by portfolio management. So, we just use security groups to control who had access to what function in the certain part of the application or a certain part of the workflow. And because, obviously, everything's integrated, you're relying, at that point on, groups that have been set up in active directory and benefiting from security around the two-factor authentication, et cetera.
David Smetana:
Okay. So, that leads me onto my next question, which is, ultimately, what benefit has this brought to the business and the users? What are they seeing, in terms of business outcomes?
David Fanning:
Well, effectively allowed us to retire quite an expensive portfolio management system, on time and on budget. So, that was the primary goal or the primary deliverable from my perspective. Other goals, it's provided the separation of responsibility that we needed for the workflow, so [inaudible 00:16:34] that portfolio managers entering the trade, operations reviewing and approving or rejecting the trade. So, it's defined and set in stone, those operational requirements and at the moment, it's the only way you can get certain trades into the system or into our loan admins, that is the only process. Which is also very good, that we got no fat finger opportunities with emails and things like that. It reduces the overall operational risk of trade execution. That's essentially what it does.
David Smetana:
So, I'm going to ask a self-critical question, of all of us. So, you said this was your fourth iteration. What, if anything, could we have done better? Let me go to you, Jacob first, and then we're finished with David. What could we have done better? What lessons learned for the next one? We'll save that for the end, because I know we're working on a new application with David, but what could we have done better with hindsight? If anything. If it's perfect, we can say so.
Jacob Myner:
Yeah, I think there's a few things that we want to try out and especially, I think, we've considered a couple of things with the next application that we've looked at. One of them being the CDS, now called Dataverse from Microsoft, which is Microsoft's own data platform. We're exploring it using that full part of the process for the next one, rather than just kind of chucking all the data straight into the database, almost using it as a temporary data store. So, I'm not sure if it's being able to do it better, but it's leveraging enough a piece of the Microsoft 365 stack without potentially having to set up a SQL tables going forward.
There was an opportunity there and I think, from my side, from the power apps part, we didn't actually do it the first time round, we didn't use the inbuilt power apps forms, and I'm really keen to try those out for the next application or the next iteration of this. Where it creates a lot of the submission through to the database for you, again, that could just be reducing more time on the project. We had our reasons for not doing it this time round, but for the next one, it looks like that could be really helpful, especially when you're dealing with entities and sub entities and as they get more complex, those forms can be more and more useful.
Harry Boyne:
The benefit and the downside of using the Cloud and using all these tools is by the time you hit the publish button, there's a new feature out, which is already bigger, brighter, better. So, it means every new app has added additional benefits. I mean, it's great and it always means there's something to learn.
David Smetana:
Yeah. I mean look, that's one of the challenges that we have as an MSP. I mean, if I asked Harry to start divulging the details of our R&D list, you'll start to see him sweat profusely, because-
Harry Boyne:
How much time have you got?
David Smetana:
It's never-ending and at some point you do have to draw a line in the sand and obviously start to work on things, but it is, it's ever evolving, and from your side, David, anything that you think we or you could have done better to improve it?
David Fanning:
There were a lot of features that we didn't implement that we could implement in the future. So, there's the opportunity to have those allocations calculated rather than being manually calculated on a spreadsheet and then pushed into the application. But overall, it does, as I say, what we need it to and I think we're happy with it. There's some additional tagging that we could've done. So, allowing the portfolio managers to tag certain trades in certain ways, but these are all the features that we made a design decision on. Basically to get it in on time, we deliberately cut some of these features. I don't know if you want to talk about the future applications-
David Smetana:
Yeah, yeah, please do.
David Fanning:
I think there's a number of technologies that would have been useful to implement that I've implemented at other places. So, using the IM channels as a means of handling the handoff between business teams. So, where you have a business process that effectively spans two or three teams, we have an IM channel to support that business process. So, the business teams can post messages to each other throughout the business process and essentially we can also hook into this from the system side.
So, when you need a trade to be reviewed, you pop a message into the channel, the appropriate person picks it up from the channel, hits the link and the form springs open with the trades that they need to review. So, it's using that connected technology in a more integrated way, as opposed to manually messaging via email or asking them to pick up the phone and look at a trade on a certain blotter. Again, if we can build these and use these technologies together in a more integrated fashion, then it just becomes fresh. It goes back that to efficiency and effectiveness, you've got less people waiting around for things, they can react to them as in a more real-time way.
David Smetana:
I presume you're thinking about using Teams [crosstalk 00:22:18]. Yeah, okay.
David Fanning:
Yeah, previous firm we used it a lot and it was very effective, especially when you can embed parameters into the links that you share in that channel and hit that link and it jumps open to the right, the right page with the right permissions for the user and they can just carry on the workflow that way.
Harry Boyne:
We can always embed the power apps inside Teams as well, in a tab within the channel, so... Have you thought of that?
David Fanning:
I have been reviewing that and we are seeing user cases for that. Yeah, I think, we just need to look at the requirements, really.
Matthew Stibbe:
And perhaps if I can pivot to a more positive note to maybe bring us to a close. What went particularly well?
David Fanning:
I think the overall development timeline went well, actually, we had strong support from the business teams, they were really involved on the testing. The turnaround from Chalkline was very good, where we needed to tweak amendments. So, I think we were just overall very happy with the timeline and they're very happy with the collaborative nature between the two firms in delivering what was a very important deliverable for that year.
Matthew Stibbe:
Fantastic. It feels like we've covered all the questions that I had. David, did you have any more?
David Smetana:
No, I mean, I think, we've pretty much covered all of it. I think we've got a good understanding, so I'm happy.
Harry Boyne:
We did speak about what's next, didn't we?
David Fanning:
Yes.
Harry Boyne:
We've covered that.
Matthew Stibbe:
Well, a little bit. David Fanning, what does this open the door to? What might be coming down the pike? What would you like to automate next?
David Fanning:
Yeah. More complicated processes.
Harry Boyne:
Are there any spreadsheets we can get rid of? That's the big question.
David Fanning:
We've been looking at that as well. Say, what we're trying to do is really... Now we've dipped our toe into the Azure development and we've established the data platform and we... Essentially, the trade capture was not a prototype, because it was actually a finished product, but the first foray into Pemberton using PowerApps platform. What we want to do is, I think, extend that use to different teams, but also use it as a way of basically applying technology to the business processes that we've identified across the firm.
So, in parallel to what we were doing during the trade capture development, we were essentially documenting, capturing the business processes across the firm, which that exercise is now being more or less complete. What we now needs to do is put some technology underneath those business processes and really start exploiting efficiency opportunities that come with, as you say, eliminating spreadsheets and applying some technology to calculations that have been done by hand.
There are certain number of calculations that are done, certain areas that are very applicable to having technology applied to them and would save, potentially weeks... Certainly man-days, man-weeks of time and certainly instances within the year but yeah, I think what we've done is a very good start and we will extend the use to various processes. The next project that we have lined up is actually more complicated, but it's part of the same trade process, if you like.
So, what we've done with the trade capture is start in the middle, but there's obviously processes prior and after. So, basically what we're addressing in the next project hopefully is a portion of the process prior to trade capture. So, it's actually the deal or instrument setup process that we'll be looking at next and that one's a little more complicated because it involves more business teams. As I say, something around using the IM channel to coordinate the actions between those teams would be very useful in a real time way or near real time. So, yeah, we're just going to extend its use.
Matthew Stibbe:
Plenty on the agenda there, fantastic. Has anybody got anything else that they would like to add or cover?
Harry Boyne:
Good.