A week or so ago I mused on how chatbots and AI technology may be in the near future for local government service delivery, communications and customer service. From that post some interesting discussion arose on Twitter and it’s had me thinking more. So here are five (not terribly imaginative) potential chatbot prototypes the public sector could pursue as proof of concept and to learn from.

Oh – and if you fancy collaborating on an alpha of any of these, or exploring the need for AI from an existing online process, get in touch as I’d love to be involved.

The five ideas I’ve jotted down here are all fairly light touch services – the sort I’d put at the bottom of the pyramid of user needs, meaning the interaction is usually low risk but high volume. They are the sort of transactions which are already likely to be being pushed as ‘digital by design’ delivery and by encouraging those that can to make their journey online it frees up human capacity for those services with more complicated or sensitive needs (like social care).

This is deliberate – some will say it’s low hanging fruit but for a prototype with an emerging technology I think these types of service present involved enough processes but low enough risk to proceed to pilot (should that become an option based on success of prototype). There’s a bit of a theme – making a report, booking an appointment, retrieving information – but it’s likely that these types of customer journey are fairly high volume (and potentially relatively high cost).

My experience across the public sector tells me these would be the right testing ground for chatbots and that a number of councils (and possibly other organisations) could be ready to start discovery on these during 2017/2018, should they choose to.

So, five places the public sector could start with chatbot and AI technology…

Reporting a pothole

Let’s start with the classic. A high volume customer journey for most top tier councils, which often spills to other layers because no normal person knows or cares which council looks after the roads, just fix the damn things.

Most councils will have tried to ‘channel shift’ this process by now – offering an online form probably, maybe even adding in options to get updated by email / text when something happens on the report you made. The objective for the organisation here is to lower the cost of the transaction by moving to the cheaper digital channel, while fulfilling the user need for convenience and efficiency of reporting. The quality of reporting mechanisms varies wildly across the sector, as does the willingness or ability to integrate external systems like Fix My Street (sadly given it’s 2017 and this integration should be a given by now).

Of course, making the process of reporting easier is great from a user perspective – let them What’sApp the location, a picture, some details about how they’d like to be contacted when something happens to the pothole.  But it may not be that attractive for some councils who simply don’t have the means to act on all the reports given current budgets – easier reporting sets an expectation something will happen. Is it better to have more reports, satisfy user’s with the reporting process, but greater dissatisfaction with the results (or lack of)? Another quandary for another day perhaps.

Other highways services ripe for chatbots? Your good old winter service of course – say goodbye to hours of staff time spent answering gritting enquiries and busting myths by allowing AI to take that strain during a ‘seasonal weather event’.

Housing repairs

Ok, I stole this one from Neil Tamplin, one of the people I discussed chatbots with on Twitter following my post. He’s also been thinking about it, and wrote this post about how AI could emerge in the UK housing sector. He even helpfully mocked up how a chatbot conversation may go when booking a repair to a property.

And he expands out further – could a chatbot talk someone through a simple repair, or advise them of any additional reports or support they are entitled to?

Neil’s post is well worth a read and a think about – it’s not a sector I have much experience of so I defer to Neil (and others) on the opportunities and challenges here.

Getting in touch with a councillor or finding a decision

Facilitating democracy, while not a cost saving or high volume customer journey, is one which local government should morally be obliged to support. Encouraging participation and making it easy to find your representatives, contact them or check up on decisions should all be something councils are looking to support.

While many elected representatives are still tech averse and more in tune with the dangers rather than the opportunity of the online world this won’t always be the case (I’m an optimist, get over it) and certainly isn’t true of all their constituents. So, a first step is maybe information retrieval – who is my councillor, how do I contact them (in a way they are open to)?

Or, knowing the council were debating or deciding on something that matters to you but working a 9-5 so not being able to attend or follow the meeting as it happened, or have the time or fortitude to wade through the minutes allow people to use Messenger to have a chatbot inform them of the decision, maybe even help them understand it and what happens next.

Missed bin collections

Another classic from the local government selection – reporting a missed bin collection, checking on when your bin should be put out, checking which colour bin it is this week because you’re in a stand-off on the street with no-one seeming sure where you’re at in the kerbside cycle.

Another one that normal people don’t know and don’t care who is responsible for, and are generally annoyed by the time they report it because they’ve been inconvenienced / are sold on the ‘lazy public sector’ lines / are waist deep in waste.

There’s loads of ways this service could make use of tech, some of which are already being used or tried out by councils. From signing up to email or text reminders that

Booking a GP appointment

Ok, so this one like some of the others suggested here is really an interation of an existing digital process. Booking and cancelling of appointments, renewing prescriptions, or checking up on parts of your medical record – doing this online has already been rolled out but why make someone go to a website if they could just use a messenger app (yeah, I know – because of a ton of privacy and identity reasons but we’re not so far along as to count ourselves out with those yet – ethics probably the next post in this accidental series).

Perhaps AI might even be able to pick up some strain on getting people to better self-care by offering advice, or getting them to present at the right point in the system by suggesting they see a pharmacist or practice nurse for some ailments or routine appointments. Putting aside the potential for AI to miss something more serious (but can’t humans do that too) there is an opportunity to support efficiency and joined up wellbeing with AI – I maybe just haven’t exhausted my thinking on it here.


Challenges? Oh, yes, there’ll be those of course. Hat tip again to Neil Tamplin who pointed me at this tweet / slide which sums up the core ones quite concisely. Again, my experience in local government suggests to me that the clean data may well be the first sticky ground to overcome here, as it has been in earlier customer journey work. But if we don’t start somewhere – and with these low risk, high volume, light touch service prototypes – how will the public sector prepare itself to unlock the AI potential for the wicked problems, like social care?

Even with those challenges this feels like something ripe for a LocalGov Digital hack day, or similar. If you’d like to collaborate on discovery of alpha of any of these ideas or chatbots / AI for another public service then get in touch – I would love to be involved.


I’m a freelance digital content strategist offering a range of services across transformation, leadership, digital delivery and training – take a look at the ways we could work together and get in touch.