Implementing a Council Wide Casework Management System – 13th December 2019 Weeknote

Welcome to the first weeknote for the implementing a Council wide casework management system project. This system will be used by all services and help them to deal with enquiries, complaints, freedom of information (FOI) and subject access requests (SARs) quickly and easily so that Citizens get responses in a timely manner.

This project breaks a bit from tradition (as far as my weeknotes go); We aren’t building a thing, we are procuring and configuring a thing. This poses it’s own challenges as we are working in a way that is unfamiliar to HackIT (waterfall delivery). I, personally, have not delivered a waterfall project in, oohhhh, about 7 years.

The difference between Agile and Waterfall in GanttPRO Blog


This is proving a shock to the system for me and I am struggling to help navigate the team through the rough waters of full upfront specs and trying to think of everything immediately.

My comfort zone is learning by doing, testing and feeding back. The big bang approach of waterfall deliveries makes me (and my colleagues) more than a little nervous so we’ll be working hard to mitigate that nervousness as best we can with our provider whilst still moving forward with the procurement and implementation.

Out of the comfort zone progress

Although this is the first weeknote, the work to assess and procure a system has been ongoing for some time. Three solutions were assessed against the needs of the Council (via G-Cloud) and iCasework (from Civica) was selected as the system that best meets the Council’s current needs.

In the past 2 weeks, the core project team (myself, Helen Mcfarlane, Simon Gray and Mandy Woodhull) have been working hard alongside Bruce Devile and all the service representatives to populate the information requested (including branding, templates and a business structure). We’ve managed to do this in a first draft then final version style as elections and annual leave have left some of the key players we need unable to commit the required time to feed in.

The first draft is good enough for Civica to complete the initial configuration of the service so that we can have our overview workshop next week. Booking this in has involved a lot of diary interrogation, emails, meetings and prioritising to make sure the right people are looking at the right things at the right time.

Next week we will be in a better position to assess how much configuration the system will need to work for the Council and get our ducks in a row before the Christmas break.

Improving the Repairs Hub (side hustle style)

Just a quick update from the Repairs Hub land to say we have been beavering away in the background and making improvements as a side hustle.

In the old world, when agents working in the repairs contact centre add contact details to a repair, they have to either retype details or go back to the tenant details page in CRM, copy the name, past it into the repairs hub and then go back to tenant details again and copy the telephone number. This takes about 20 seconds per job*

flow for RCC agents
We got rid of the red bit!

In the new world, agents now see the tenants names and contact numbers just above the fields for contact name and number whilst in the raising a repair form.

Repair form with contact details

This will save them from retyping or having to copy and paste from a different window/tab.

Thank you Igor from Unboxed, Bukky from Hackney’s Dev team, Rashmi who found time for Bukky to help and the RCC leadership for their unstoppable enthusiasm to iterate and improve.

I’m hoping to side hustle another improvement come new year so, watch this space 😀

*As timed by me when I shadowed calls for an afternoon and noticed what a faff it was, not an official figure

Improving the Repairs hub – 8th November Weeknotes

This is the final weeknote of our intensive 4 week project to address the most pressing issues facing Repairs hub users (last weeks note is here).

The avid readers amongst you will know that colleagues from HackIT, the repairs contact centre and Unboxed have been working together to improve the Repairs Hub application based on feedback that’s been gathered from users since March 2019.

This last week has been all about code code code. Igor from Unboxed joined us for 5 days in which we have been writing, testing, approving and deploying code regularly.

kitten fist bumping

Monday saw us move hosting from Unboxed’s Heroku to Hackney’s. Thanks to the wider Hackney and Unboxed dev teams for supporting with this.

Tuesday saw the team develop and get flags for new builds onto dev and Hotjar onto the NCC staging environment (thanks to Emma Lewis from the NCC CRM team for progressing this so fast). Bukky also put some revised error messages to live.

Wednesday was busy; Hotjar went live on NCC CRM and we verified feedback was coming through as expected (it was, hooray). Address and contact alerts went onto dev for approval and Bukky put more error messages live.

Thursday was a day of approvals for Barnes. New build flags, contact alerts and address alerts were all approved and deployed to live. Barnes, Bukky and Igor also agreed what the minimum viable product for caller notes was in the time remaining (show the name of the agent adding and the date the comment was added).

Friday has been our final code push to show agent name and time/date of comment on the front end get approved and deployed to live. The final error messages went to live and we also gave our final show & tell. Thanks to all who attended and fed back to us on how you think we’ve done

happy child walking

All in all, each one of the four weeks has been a success. We’ve addressed the highest priority feedback and made improvements for users and residents. Our colleagues from the repairs contact centre have enjoyed working more agile and we’ve built really good working relationships. We achieved one of our stretch goals in our final sprint and have gathered even more insight form users to build a more holistic view of what is important to focus on next.

We’ve been brutally honest about what we can achieve now and what we’ve ear marked as the next thing. The job isn’t done yet, and we know it, so watch this space!

Improving the Repairs Hub 1st November Weeknotes

It’s week 3 of our 4 week push to improve the repairs hub for agents using it every day in the Repairs Contact Centre (RCC). If you want to see what we’ve been up to in the previous weeks, our weeknotes are publicly available to view.

This week, the focus has been squarely on identifying and preparing for the work Igor from Unboxed will be completing next week. He’s going to be doing lots of front end work to:

  • Flag when a dwelling is a new build
  • Clearly display the legal disrepair status of a dwelling
  • Flag if there is any cautionary information and what that is
  • Make error messages clear and informative

We are also going to try and show the tenants name and contact number to agents when they are raising a ticket and include details of who added a note and when (these are our stretch goals).

The preparation

Bukky and Richard worked together diagnose why some properties in Universal Housing are correctly alerting as under legal disrepair but were not alerting in the repairs hub. They eventually tracked this down to mismatching reference numbers. Bukky updated the query (to include a reference number we didn’t know we should be using) and, hey presto, all of the alerts now show in Universal Housing and Repairs hub! Hoorah!

Go team!

Currently, address and contact alerts come through as one piece of information so Bukky needs to make a further change to deliver these as distinct entities. He also needs to create mapping between the abbreviations we currently see and the actual alert message. We can then display plain English alerts and not coded messages to agents.

Chris Lynham reviewed our new error messages and confirmed that the information included in the error is enough for the support team to progress issues with and is also plain English for users. He advised on the correct next steps to take about errors too so we know we’re asking users to take the correct action once they’ve seen an error. We also checked them against GDS’ error messages in the design system. We can’t use their styling but we have made sure to follow their advice on what is a good error and what is not. I think we’ve nailed it and new our error messages are ready to go (as soon as Bukky has updated them)!

We also prepared HotJar to be added to the NCC CRM this week. Although it’s a few weeks until we have any developer time to do this, HotJar is all set up with the new feedback survey so will be ready for RCC agents to start putting feedback into once the tracking code is on the NCC CRM pages.

The pickle (or bit of a mess)

This week, we’ve managed to lose our way around notes. There are a plethora of questions that need answering; Which notes should we see in the repairs hub? Where do they come from in Universal Housing? What type of notes do we need to add via the repairs hub? How do we do that? What do we need to change about the notes we already have? What privacy considerations do we need to make with regard to notes?

Of the questions we have, we managed to answer none and, instead, generated more. It’s a pretty disappointing result as decoding notes was a real focus for this week and it feels like we’ve gone backwards. We’re hoping that when Barnes returns form his leave next week, we’ll be able to get the direction and focus we need around notes to make sense of them.

The postponement

We’ve had to swallow a bitter pill this week. We had high hopes of moving to Google Single Sign On (SSO) for the repairs hub but we’ve been struck by technical issues which we are unlikely to clear before our 4 weeks is up. The first is that, for security reasons, all service environments (live, dev and staging) need to be on a * domain. Our environments aren’t. We would also need to do this for the Neighbourhood Contact Centre CRM (NCC CRM) system to make the change really valuable. This isn’t a huge deal to resolve but it is extra work we didn’t expect. Secondly, we discovered that credentials agents use for logging in to NCC CRM might power much more background than we thought. We’re investigating this now to see what login details are used for apart from authentication; We’re hoping nothing as that means switching to Google SSO for both services is doable, just not within our time frame. If the answer is ‘some other stuff’ we have a much more difficult thing to untangle. However, Google SSO is our chosen solution for identity and authentication so we won’t be letting it drop!

The perspective

Bukky and Emma spent two sessions shadowing calls with RCC agents this week. Thanks so much to the team over in Robert House for making us so welcome! It was amazing and very valuable to see, first hand how the systems were used together and get insight into how a 20 second ‘annoyance’ per call can add up to minutes and hours of time over the course of days and weeks. Well be adding a few more ‘quick wins’ to our backlog, that’s for sure!

We also had some RCC agents joining us in stand up this week to see just how we are communicating to get things ‘done’ in this 4 weeks. It was great to see some fresh faces and get their extra point of view on the services and our approach. Thanks to all those who took 15 minutes out to attend.

Our last show & tell is on Friday 8th November at 12:00. If you’re interested in coming, drop us a line!

Improving the Repairs Hub 25th October Weeknotes

Welcome back to the Improving the Repairs Hub weeknotes! We’re half way through our intensive 4 week project aiming to improve the Repairs Hub for repairs call centre agents and managers. You can see  last weeks note if you want to get up to speed.

In case you aren’t familiar with the repairs hub, here’s a bit of history; The repairs hub launched in October 2018 for use by the leaks hub. Work to extend the service to Repairs Contact Centre (RCC) agents began March 2019. The RCC have now been using the service long enough for us to have good feedback that we can base further iteration one. This project aims to address as much of the feedback as possible over the course of 4 weeks.

The People

We are delighted that Igor from Unboxed will be back with us for the week starting 4th November. This is the last of our 4 week project so we’ll be working hard to make sure he has everything he needs as soon as he joins us.

We also welcomed Richard from the application support (app support) team here at HackIT. Involving colleagues from the app support team is critical to the repairs hub work as all the data we serve is provided by 3rd party software. Guess who looks after that 3rd party software…That’s right, app support. Now Richard is with us, we’re looking forward to a mutual knowledge exchange and ensuring that the work we are doing is done in the right way. Talking of what we’ve been doing…

The Work

This sprint we have been focusing on planning out how we will resolve connection errors, getting new build data via the API and fixing an issue where agent names were vanishing from tickets.

It’s not possible to never have a connection error, ever. What we can do, though, is make sure that the information about the error gets to the people who can fix it as soon as possible.

We have seven controllers in the repairs hub. Each of these is responsible for connecting with third party software to retrieve or add data. In order to diagnose connection errors, we have checked in on the logs for the controllers. The appointments controller is the culprit for most of the connection errors so that’s where we started. We have identified and documented each type of error the controller can return. We are now going to work with the app support team to design an error message we show to users which will give the information needed to investigate the issue and signpost them to the correct team to contact about the error.

We discovered that some errors sent by our 3rd party providers were being lost in the ether so we’ve added code to correct that. From next week, we will be able to go to our providers with their own errors and ask them to resolve it. How good is that?

Thanks to many examples provided by Nicola, we hunted down and fixed an issue which caused reporting agent names to vanish from tickets once they were completed. Bukky found some attributes were missing in the code so we added those in and, voila, everyone’s names are back. If a name is still missing, let us know!

Richard has been kept on his toes hunting for legal disrepair data in Universal Housing. With the help of Nicola and Jeanette, he’s done it! Next week, Bukky should be able to get that data coming through to the repairs hub.

Finally, we had our show and tell this week! Thanks to everyone who came. It’s really great to hear the passion in the room to want to keep moving the repairs hub forwards. We are right there with you (and we’ve got the snacks).