Public Organisations: Automation Budgets
Next: Bugs Bounties
After a prolonged summer break, it’s time to get going again. Let’s move focus to public side now.
How can the public organisations change their operating model by using concepts like digitalisation, modularity, openness and agility?
The following set of posts cover three main areas:
How to go about transforming public service creation?
What types of services are enabled by new methods?
What would the architectures be that enable this?
Citizens’ Automation Budget
Automated decisions would save considerable amounts of money and improve speed and accuracy. But which decisions and operations should be automated?
Why not allow citizens to decide?
It would work so that everyone has a yearly budget that they can invest in digital automation of government – for example 10-20% of paid taxes and some lump sum for adult people who do not pay taxes. If you encounter problems with public services – say decision on your unemployment benefits takes long - you are free to make a proposal that such decisions to be automated. Once enough people join the cause so that your combined funds can pay for the automation, the project proceeds and can ask for quotations through a public bidding platform. The process of selecting winner would be helped by procurement professionals.
Payments are gradual during the implementation and follow completed milestones. Once work is accepted to production, a grace period follows to ensure that it works well. During it feedback is gathered from users and from monitoring team. After final acceptance, remaining funds are released to the developers. Program is open sourced so other teams can re-use bits and pieces in future projects. This brings costs down over time in similar projects.
End result is that governmental services happen in real time, and are always available with greatly reduced costs.
Such automation budget may also be used for other purposes than just automation of laws. Let’s say that I have been scammed when making a transaction on some platform. I might use my budget to implement “scammer alert” - an app that uses machine learning to detect and flag scam campaigns. Typically such a service is implemented with machine learning by clustering actors based on their behaviour (phishing, scamming, virus messages for example) together. Actors who are similar to known scammers would be flagged.
Different initiatives would deal with user interface issues – such as providing access to an existing automation through a voice interface or just fixing usability issues on existing applications. Disabled folks might have a larger personal automation budget so that they can implement needed user interface mechanisms to their services in their preferred way. This because their numbers are lower and you still need the same amount of money to implement technical changes.
The diagram below outlines the process.
Enterprise Automation Budget
The previous automation budget concept works well for issues relating to people. The automation budget can be expanded so that certain percentage of the VAT or other taxes paid by corporations can be used to automate regulatory processes facing them.
The concept can be extended to include international regulations. Trade agreement rules are the same (hopefully) for both parties meaning that in theory there could be one implementation used in all countries that are in the same agreement. Say I ship goods to customers in multiple countries and a various information elements are needed by customs and other public organisations in the recipient country. Ideally there is just a single interface (API) that I integrate with irrespective of the destination as long as all those countries are part of the same trade agreement.
Internal Automation Ideas
A third source for automation ideas would undoubtedly come from within the public service organisations themselves. People working there have inside view into potential inefficiencies.
Most good ideas get ignored in any large organisation – it is the law in hierarchies. No difference whether public or private. Low hanging fruits inside a unit tend to be already implemented. Good ideas cross multiple organisational boundaries.
To understand dynamics, consider how an average traditional organisation works. Work is split into well-defined areas and some manager is responsible within those boundaries. Cross-boundary ideas and co-operation have the risk that overlapping work is detected or some responsibility is moved away to an upper layer or to a different unit thus reducing manager’s little empire. Maintaining status quo is the much-preferred alternative, thank you very much.
A special flavour of public organisations that do not face any competition (being law based monopolies) is that within such units, it tends to be more important who decides, not what it’s all about.
The open idea platforms where funding for projects inside public organisations can directly be gained from citizens is a way for employees to bypass management and have transformative ideas presented and implemented.
Managers will naturally dislike this and they may well have a good hunch whose is behind the proposal, leading to punishment. If you find yourself inside such a “we value feedback very highly and always do our best” public organisation, you can resort to a very old best practice method: be the most vocal critic of your own idea initially until you finally accept that the law is the law and become converted to support it.
Automation Rewards
It’s clear that citizens and enterprises that are hurt by some malfunctioning practice have an interest to fix things through the Citizen/Enterprise Automation Budget. But why not allow 3rd party companies to add their skills to the soup and speed up public delivery for free, or actually paid after the fact by the achieved savings.
This mechanism works so that public services have open interfaces (APIs) and a test or simulation environment for developers.
Anyone – association, private person or enterprise – from anywhere globally is allowed to use them and automate some aspect. This gets tested – first automatically by running a large set of recorded old data in a verification environment – to see its quality. After passing several quality steps, it gets into small scale field use and finally into full production.
Automating the public functions will save quite a bit of money. It is only fair that the people driving these ideas be rewarded in proportion to the actualised saving for some duration? For example, everyone who invested into a useful automation, gets a certain percentage (20%) of measured savings for say first 5 years?
Due to changes in society (ups and downs in economy), it may be difficult to calculate exact amounts, but a ballpark understanding should be possible. For changes that do not seem to change anything, one might conclude that they perhaps made the system faster but did not have other impact.
Automation Flywheel
The automation budget model can ignite a virtuous cycle in society. To understand this better, let’s draw the model s a flywheel.
It is a self-enforcing loop where automation saves costs for tax payers and makes service available every day at any time (24*7). This improves customer experience.
Automation also leads to increased usage as it changes the model from opt-in to opt-out. Opt-in is the current model where you have to know that a specific service is available and go through complex bureaucratic hoops and use the right liturgic expressions before your rights are granted to you. Opt-out is model where citizens are invited to their benefits (positive decision is made automatically for all who are entitled) and need to actively reject it.
The positive impression and improved usage as folks tell their friends about their experiences leads to more people involved in improving services and using their automation budget. This leads to more services being automated.
The growth of automations leads through economics of scale to lower overall costs. Both for implementing services and operating them. The loop starts feeding itself.
Side-note: Solving problems when you have no skills
The automation budget concept discussed in this post is just a special case of a more general way of doing stuff when you do not know how to do stuff.
Basics of it is very simple: you collect a bunch of like-minded people who share your interest and you put together a pot of money. Then you promise to pay for the solution. A variant of this is to hold a competition. In competition model whoever solves this problem or makes the best proposal gets the reward.
You do not have to have the skill, all you need to bring is your problem and invest your time and some money with people in same position as you.
That’s it.
The problem with automations is that they can have errors, how to protect against that? This will be the topic of next post and then we’ll dive into more ways to make decisions about public services.
Next: Bugs Bounties