Even before the age of austerity, the public sector has often had to find ways to do more with less. And now, this remit is more crucial than ever. In an era when public funding has been cut sharply, many have begun looking at more innovative ways to solve their problems. Some have engaged in setting up shared-services, some have outsourced work, and others have developed capability in their internal teams to in-source previously outsourced functions in an attempt to improve savings. Then there are those that have looked to emerging technologies for ways to reduce costs.
The challenge with many of these changes in direction is that the ‘end game’ is hard to envision and to document with accuracy, where the unknown ‘unknowns’ pose a risk and often result in costly problems arising further down the line.
Is Enfield Council a pioneer of the public sector… and who is ‘Amelia’?
There are visionary things afoot in the Council of the London Borough of Enfield. It is reported that this north London local authority is to be the first to employ artificial intelligence in a key role within its workplace – its call centre.
The deployment of Amelia, a “cognitive agent”, comes as part of a working partnership between the council and New York-based technology company IPsoft. It is their hope that this advanced system will help with the certification and authentication of licence and permit applications, as well as responding directly to residents’ queries.
According to the Financial Times: “Amelia uses natural language processing to understand context and emotion in requests, and machine-learning techniques to improve its response when it encounters new problems.”
Enfield Council’s Director of Finance and Corporate Resources and Customer Services, James Rolfe, stated that the system was developed so that “customers shouldn’t even notice they’re dealing with Amelia”.
The hope is that if Amelia is proven to handle calls in the right way and in a consistent manner, the burden currently placed upon existing Enfield Council administration teams would be reduced, freeing them up for more complex, value-adding tasks.
The life of a transactional ‘bot’
If Amelia proves itself capable, this could encourage further research into IA integration within the workplace – and not just in the public sector. Many look to large Enterprise Resource Planning (ERP) implementations and more to help reduce the burden on their admin teams and improve workflow, but the challenge is that both vendors and clients can sometimes underestimate exactly what needs to be done to give their clients what they need.
In using bots for routine and transactional tasks, the risk of purchasing a system that doesn’t do the job is reduced, as there is already a working model upon which to base the bot’s behaviour – the day-to-day work of existing staff.
Better still, with a bot providing support, all the answers it gives are likely to be ‘controlled’, ensuring a consistency of service while allowing for more detailed monitoring to take place by humans – as all responses stand a greater chance of being coded and recorded for later analysis. The insights gleaned from this data can help to develop the bot and improve service levels, but it could also be used to highlight shortcomings with the systems it’s working with and within. Armed with this information, you may have a clearer idea of how your systems would better interface with your bot, and your employees.
AI in the private sector
While full automation is a while away yet, these types of initiatives are crucial stepping stones to test the effectiveness of AI in supporting service-based organisations. Already the private sector is piloting this technology to support the completion of routine and tedious tasks. In the US, NextLaw Labs, wholly owned by the global law firm Dentons, has recently invested in ROSS Intelligence, a tool based on the same technology as Watson, IBM’s AI supercomputer. The intention, in this context, is that lawyers and other employees can ask it legal questions and get specific answers. The intelligence behind this system will learn and evolve over time as it handles greater volumes of queries.
Similarly, Riverview Law in the UK has teamed up with Liverpool University as part of a Knowledge Transfer Agreement to develop a suite of tools that will help to automate legal decision-making, and to reduce the administration needs associated with managed legal services work, which is often commoditised.
According to Riverview Law, the system will be able to “create, automate, maintain and evolve complex end-to-end workflows and processes for all areas of legal, compliance, risk and related activity… The platform manages all activity from instruction, triage, case management, document creation and storage to alert triggers, workflow creation and forensic audit.”
Outside of law, the technology that Amelia has been built upon has already been used by organisations such as Baker Hughes, the oil services group. In their case, Amelia helps to provide policy guidance to mortgage brokers for a high street bank, as well as managing invoice queries.
The rapid adoption of these systems has many fearing the impact that automation may take on jobs. Research and advisory firm, Forrester, predicts that these advances could lead to a net loss of 9.1 million US jobs in less than ten years.
That said, both IPSoft and Enfield have both been careful to explain that their programme would be used to complement existing services without posing a risk to existing jobs.
“It’s about freeing up our people so they’re not dealing with routine stuff that Amelia can deal with swiftly, so they’re free to deal with more complex problems,” Mr Rolfe said.
With such advanced applications of this technology being apparently possible, it would seem that getting a ‘bot’ such as Amelia to handle residents’ queries would, in the grand scheme of things, be just a small step. But there are wider implications for the use of artificial intelligence on such a grand scale.
Good for Enfield
While many claim to be forward-thinking innovators, and some actively exhibit progressive behaviours towards the implementation of technologies like this, it’s rare to see it in action. Enfield Council has taken a bold first step into utilising a technology that others in the public sector space are either unaware of, or talk very little about. Arguably it’s initiatives like these that will help to shape the face of public sector services in the future where much smaller, ‘lean teams’ will be required to rely on technology to respond to the needs of the masses. Other councils would do well to follow their lead.
If more were to implement this kind of technology, particularly in light of recent political, and potentially economic, turmoil, the reality is that at some point savings are likely to stop coming from efficiencies, and instead start coming from notional salaries as artificial intelligence renders some council job roles redundant.
7 steps to trialling a ‘bot’ initiative for yourselves
1. A ‘bot’ to do ‘what’?
As part of your service reshaping, you will already have identified areas of transactional processing where you are attempting to reduce time and cost spent. For example, call centre handling has long been an area of high manual labour cost on a per-transaction basis. Services along these lines can be evaluated to ‘experiment’ with.
2. The ‘art of the possible’
Leaps in technology are often considered positive, but it is important to keep an eye on exactly what it is you are trying to achieve with the help of a pilot initiative of some kind. If this type of initiative hasn’t been approached before in your organisation it would be sensible to put together a ‘scout’ team to explore how other organisations are using such technologies and in what context. This helps you understand ‘the art of the possible’ and starts to inform your thinking as to whether such technologies should be considered for your own services and. if so, in what context.
3. Early market engagement
Having first explored the ‘possible’, next put together your outline requirements of the service you are trying to reshape. These should include some outline business objectives (shaping) of what you would like to achieve, the current status quo of how the process works now, the existing resources and baseline costs to operate the existing service. These should ideally be aligned to the organisational objectives so that everyone in your organisation understands in what context such an initiative, if implemented, would help the organisation to achieve its wider business outcomes. Then approach the vendor market on an exploratory basis to ‘sound out’ their thoughts as to whether they believe your objectives are achievable, or whether you have left your senses behind!
4. Business case
Business case. Having had the early market engagement, you will be able to start to firm up your business case to determine whether such an initiative is realistic and is likely to achieve your goals. What are the quantified business benefits you expect to achieve and by when, will they justify the costs involved, and do you have stakeholder buy-in? Bear in mind that any agile technology prototyping comes with risks, so you’ll need to outline clearly how you intend to mitigate those risks.
5. Requirements shaping
If your business case gets approved, then you will be able to start to firm up your requirements. It’s key to keep at the forefront of your mind the business objectives you are seeking to achieve, how your organisation will benefit and on what basis.
6. Procurement process
It’s now time to formally engage with the market to see how their own solutions might help you achieve the business objectives you had set for yourself. Your vendors need to undertake the appropriate pre- and post-contractual due diligence exercises to ensure that you can achieve the outcomes stated in your business case. Note that some individuals in your team could get carried away with the excitement of new technologies and particularly innovative solutions – such as AI/robotics, so it’s really important that throughout the procurement process you don’t lose sight of the business benefits you are seeking to achieve on a quantified basis. This will keep your feet firmly on the ground so that when you see something ‘shiny’, you ask ‘So what? How will this achieve X, Y or Z objective?’.
7. Post-contract performance management
Once you go ahead and purchase the pilot initiative, keep monitoring the success of the pilot and whether it is achieving the business objectives you had set out for yourself. If it isn’t, but you honestly believe it’s a timing or alignment problem, then keep reshaping the implementation process and/or have internal discussions about adding in new objectives and retiring those that you now consider should not be continue with.
Note: If you are in the public sector, there are specific EU regulations you need to comply with (for now), particularly for exploring the ‘art of the possible’ and ‘early market engagement’ steps above to ensure a fair and equitable supplier selection process should you enter into a formal procurement process later. If you would like to discuss this, please contact us directly.
The ‘rise of the machines’ can be a worrying time for staff on the front line who are currently carrying out tasks that may well become automated in future times. It is important, therefore, to calm any fears about job losses while also keeping in mind the realistic evolution of such technology to expand its reach and remit to take on greater responsibilities.
The world of Isaac Asimov’s iRobot may not quite be here yet in body, but it’s certainly here in spirit. Artificial Intelligence within a wide range of systems is becoming more and more prevalent. Tesla’s self-drive cars are a good example of science fiction meeting reality. But while this technology may be a move in the right direction, a recent alleged fatality during testing shows that much is still to be done in this area to make it as safe as it can be.
We all have to accept that the future in ‘transactional processing’ will become more and more automated and the pace is accelerating. The question is, are you ready?