Exploiting the automation opportunity
Practically every major organisation is on a automation journey, capitalising on the power of Process Automation and integrating Artificial intelligence (AI) with Robotic Process Automation (RPA).
Properly executed, the potential benefits can even surpass the media hype. It really is possible to create sustainable transformation and deliver significant shareholder value. It feels undoubtedly like there is an automation gold rush on the horizon. How to make the most of this new frontier?
As with every gold-rush there are draw-backs, pitfalls and … cowboys.
Companies can only succeed if they know where the gold is, what it looks like, and how to mine it. Done correctly, the risk is low and there are fortunes to be made with the right approach.
In this document we examine the crucial part that AI can play in large scale Automation Programmes. We offer you a 10 point C-suite guide, written from an expert practitioner’s perspective.
Before you get started
A: Understand the automation technologies and services available
There are many technologies available today and the vast array can, at first, be confusing. Many are complementary and many perform the same function; albeit in different ways. The simplest approach is to consider them in groups. The principal groups are: Robotic (e.g. Blue Prism), Desktop (e.g. Openspan) and AI (e.g. Perceptive). We also routinely include products such as analytics and OCR but these are peripheral. To maximise the potential, you will need to understand the role each of these products plays, what their strengths and limitations are and how they can be combined to best effect for your process portfolio. It is worth remembering that these tools are relatively inexpensive but, in combination, the cost may become prohibitive. You also need to consider how this market may shape up in a few years’ time. We often advocate a “champion and challenger” model to guard against the possibility of our clients’ chosen technologies no longer being available or best in class. It is almost certain that you will require a combination of at least two technologies. This is where AI plays a pivotal role.
The AI we are referring to is not cognitive and we don’t expect cognitive technology to play any significant role for many years (We are often asked if we support truly cognitive technology but the truth is it’s not ready yet). The AI we deploy is mainly used to take unstructured, non-digitised information and digitise it, extracting key data as required. It then hands over the inputs (previously unstructured) to a robot or an application in such a way that it can be used. We have worked with many Fortune 100 clients and in every case AI was required to complement the RPA tool(s).
B. Decide how you want to buy services.
Surprisingly the service provider market is relatively immature. Most of the technical product providers will offer short-term help with implementing their products but very few have the resources to do this at scale and prefer to concentrate on developing their products. With a few notable exceptions, the majority of the BPO and Consulting providers have been slow to adopt Process Automation and most are restricted to one technology offering (and a process level approach) which may lead to small pockets of efficiency but will not lead to large scale transformation. Automation is a low-cost service offering, can be transitional in nature and therefore doesn’t lend itself to the commercial models of these organisations.
The choice is between: do-it-yourself, partially outsource (and form a joint automation team) or outsource completely. Running automation in-house has a number of attractions. It is lower cost in the long term, key-skills are retained within the organisation and dependency on external providers is minimised. However, it will take longer to achieve the desired outcomes and many organisations lack the skills, focus and drive to manage large-scale programmes of this nature. If you decide to partially or fully outsource automation services, then you will need to consider the maturity of the services market. It doesn’t make economic sense to outsource to a partner who must train resources from scratch to meet your needs, you may as well do that yourself.
You also need to consider how you want to buy. The standard approach is time and materials but you may wish to pay as you use (access to tools by the minute), pay per transaction or pay for value delivered.
The attraction of the different pricing models becomes more obvious when the portfolio of candidate processes is known. For example, paying by the robot-minute works best with large-scale but highly seasonal work. High volumes of simple transactions, on the other hand, might lean towards a transaction-based pricing model. Note that many of the AI providers will only sell on a value pricing basis, based on a percentage of the overall business case delivered through automation.
Key steps to success
Whatever you decide in terms of technology and service choices, our recommended approach is “think big, move fast and get organised”. This is based on the considerable experience gained by Virtual Operations in supporting our clients’ large-scale automation programmes.
There are 10 essential steps to achieving highly successful process automation programmes:
- Understand your objectives. We have found that the principal drivers dictate the way in which the programme is constructed. Cost reduction is almost invariably a major goal but can be considered as a pre-requisite. Other operational issues are often seen as more important including:
- Enabling legacy systems to interact seamlessly with new technologies (a major step towards creating a digital enterprise)
- Increasing regulatory compliance (and regulator confidence)
- Workforce management (especially coping with peak or seasonal work-loads)
- Reducing the off-shore footprint
- Improved customer experience / satisfaction and follow-up (e.g. combining IVR, Artificial intelligence and robots in call centres)
- All large organisations will have a wide range of processes which could be automated and technologies which can be deployed. Defining your ‘Operational Landscape’ is a critical step in optimising the value from automation. The following link will guide you further. http://virtual-operations.com/key-questions-to-consider/
- Communicate and sell. A wide range of stakeholders will become involved in any programme of significant scale. Regular and timely communication (selling and education) is a vital element in managing the stakeholder group. Business owner buy-in will create a pull for automation. Remember that automation programmes are low risk but will invariably involve disruption. With the holistic approach disruption is anticipated, managed and minimised. Personnel who are to be released or redeployed will require particular attention not least because the detailed process data will be gathered from within this group. External communication is also a major consideration as stock markets look favourably upon well -orchestrated change initiatives.
- Think big and take a holistic approach Aim at least at the functional level & not process by process. http://virtual-operations.com/rpa-and-the-glass-ceiling/
- Build an Automation Centre of Excellence (CoE). Unlike Shared Service Centres, CoE’s in the automation world are relatively small. The purpose they serve is to provide the resources and tools required to ensure a consistent, low risk, leveraged cost to the programme and the disciplines and standards, (strategic focus, governance and prioritisation) that should be applied across the organisation. The CoE will:
- Manage the Target Operating Model design and development (roles, organisations, relationships, governance (working with IT)) of a federated CoE to support and sustain full-scale production;
- Build an automation methodology for use within the CoE;
- Align CoE activity to other areas of activity within the organisation such as IT Roadmap or other process optimisation (eg Lean) and shared services to break through the automation “glass-ceiling”;
- Act as subject matter experts in support of the automation technology;
- Manage skills sourcing and/or training;
- Future-proof the programme by keeping pace with the leading automation technologies.
- Release sufficient resources and budget. Well planned and organised automation-programmes will break-even after approximately 12 months. It is a false economy not to invest in a programme which will have such a rapid and high return.
- Appoint a Head of Automation from within; a focal point and an organiser. This is a senior role and requires a combination of vision, persuasion and project management abilities.
- Train your own people where possible. The most difficult tasks we face are identifying the right opportunities and then mapping them accurately. Once the process has been mapped, the implementation is usually straightforward.
- Run Pilots not Proofs of Concept. All of the leading technologies have now been proven on multiple occasions. If you need reassurance with your own systems and your data then we would recommend that run a pilot which can go into production and deliver value when it succeeds.
- Don’t stop in-flight Automations. In many organisations isolated automation initiatives are already underway. Typically these are tactical, with a single tool (usually Robotic) and are conducted on a process by process approach. Our advice is to let these initiatives continue. They are highly likely to succeed, they present little risk, they will almost certainly underpin automation as a change lever and, most importantly, they can be gathered in when the CoE is operational.
Planned well, from a holistic perspective, the automation gold rush can bring exponential benefit.
In part 2 and 3 of this series we will explore the operational framework required to bring new automation value to life and some tips on how to choose a partner.