How to successfully Scale Enterprise Automation and AI - The 1st in a series of 9 articles; “An overview”

Alexander Hübel

December 10, 2021

How to successfully Scale Enterprise Automation & AI

The 1st in a series of 9 articles - “An overview”

As many as 70% of digital transformations fall short of their objectives (BCG,2020). Among these digital transformation initiatives, intelligent automation is one of the larger investment areas for companies today and seems to continue to be so in the future (HFS research, 2020). 

How can one make sure they succeed with their automation & AI strategy and become one of 30% percent that fulfills their digital transformation objectives? Based on my experience as head of automation & AI at a large Enterprise, -  driving it from scratch to 1000+ deployed automation & AI solutions, giving millions of hours back to the business and, transforming hundreds of core business processes. I have seen the complexities in managing scaled-up transformations based on multiple technology platforms hands-on. This is the first in a series of 9 articles where I would like to share what I have learned along the way and what I would do differently today. The first article gives an overview of the topics that will be discussed in more detail. The main theme of the series of articles is based on 8 factors I believe are the main reasons companies fail to scale AI & automation.

1. No strategy or clear vision of what you want to accomplish with your Automation & AI initiatives

HFS & KPMG researched why companies fail to scale automation and concluded that "a lack of vision or the non-existence of an enterprise-wide strategy are formidable obstacles that may impede the ability to scale automation" ( KPMG; HFS, 2019). I agree and believe that it is important to avoid the fallacy of starting to implement RPA because someone thinks it's a good idea for whatever reason, missing out on the bigger picture and the vision. In order to have your strategy and vision visible enterprise-wide, It is important to have the right sponsors. Sponsors that genuinely care about your automation & AI project. I also believe you need to identify key influencers amongst the people affected by the automation and AI projects. These are your champions in the various business units that will walk along with you on your automation & AI journey.

2. No established control of the operating model - Who is doing what?  

As part of your strategy and plan, it is important to have a clear idea of who and/or what part of the organization should and are doing. You need to have a standardized way of working, and a way to track business value. Make sure that the right person or people are accountable for the business transformation. Set a structure for your opportunity tracking, your operations, and your platforms. However, there are different views as to what type of operating model is best suited for these types of initiatives. From my experience, a more centralized operating model provides the best opportunities to build the needed capabilities to drive a successful transformation. In any case, no matter what model or methodology you choose or are forced to choose, based on your organizational realities, you will have to manage the democratization theme that will shake up the fundamentals of your operating model.

"You need to have a standardized way of working, and a way to track business value."
My six pillars of the operating model

3. No proper funding or financial model 

Early on, It is important to secure proper investments. Not having the financial freedom to innovate, transform and build capabilities will ultimately lead to a less successful transformation. C-suite should see investments in automation & AI capabilities as long-term strategic investments to ensure the continued competitiveness of the company. This will generate some short-term business value, but over time it will be a multiyear investment, expected to pay off over a 5-10 year period. Moreover, in terms of ROI, there are two aspects. The first one is for quick wins that are idenfitied early on. The payback period for individual solutions could be something like 6 months where you could see numbers on ROI really flourish. These are great investments but only if you consider them individually. This leads to the second aspect of ROI, the structured transformation program, which is where you need to end up. Here you instead might have a five-year-long payback period as it will require significant investments to build the right capabilities and platforms to drive intelligent automation at scale. To achieve this, you need to make sure that the C-suite shares this mindset and continuously invest in this area to succeed. You will fail some of course but as you learn you will also succeed on more things and the long-term business case begin to look really good

Lastly, connected to the operating model is the financial model, which answers the question of "who is paying, for what?". Another crucial design principle to consider. The way you design your financial model will drive certain behaviors, and factors, like ownership and speed will determine your success rate.

4. Inability to create a pipeline with opportunities

A common reason you may become stalled on your automation journey is that you run out of new opportunities to automate. Which will undoubtedly move your whole initiative into a stalemate. Three key reasons behind this can be; One, you did not invest sufficiently in the right focused capabilities fronting your business stakeholders. Two, employees are in fear of being made redundant and therefore are opposed to the movement of automation & AI adoption. Three, you are tracking the wrong metrics resulting in disqualifying great ideas and use cases. This is all fundamental if you want to move from your initial10 bot “pure enthusiasm” initiative to a transformative 1000 solution Hyperautomation initiative.

5. Lack of visionary leadership support and the right body of competence

It is not easy to succeed with all of the above unless you have the right people involved in the scaling process. A few different types of profiles that I believe are necessary in order to scale Automation & AI are; “Visionary leadership”, “Business transformation & Business SMEs”, "Technical/functional automation & AI competence", and “IT & Platform Expertise". If you lack the visionary business leader, you can still do well if you ensure the right competencies in the right areas. However, you are likely doomed to be subscale and less transformative.

four different types of profiles needed to scale your Automation & AI.

6. IT – Friend or Foe

During the many years I have been in this industry, I have heard from several CoE Leads that IT is one of the main blockers towards succeeding with Automation. There is sometimes a built-in resistance in IT towards new digital solutions leveraging existing data systems, especially if those initiatives are driven by other teams. Individuals may feel threatened, existing roadmaps might be redundant, managers may risk their budget and/or team size. Unfortunately, or fortunately, depending on how you see it, automation cannot be done at any scale without engaging the IT teams. However, IT can also be your best friend when scaling. We managed to scale from IT, - otherwise, we would have never been able to drive the same enterprise-wide, funded initiative from outside of IT. In addition, conflicts are predictable, but if you debate every project, solution, and possible integration, you become very slow in your decision-making.

7. You might be doing RPA well but scaling to Hyperautomation is harder

Once you have mastered Robotic process automation, you are likely in need of moving further into the hyperautomation space in order to automate larger and more complex business processes. Hence you start to look at implementing other complementary technologies and capabilities (be it Artificial intelligence/Machine learning, Intelligent document processing, Natural language processing, Low code, Process mining, or others to meet your business needs). Mastering this switch has little to do with technology and instead is more of a leadership question. This becomes challenging once you are in need of prioritizing amongst technologies and placing bets on what to go for next. Key learning is that no matter what technology you onboard, it will take a while until your organization can really master it. Do not over-expect fantastic business value from the first couple of use cases, keep investing to ensure that you arrive at a stage where you soon master the new technology. Then you can start to expect high benefits and efficient scalable implementation.

"Mastering the switch towards hyperautomation has little to do with technology and instead is more of a leadership question"

8. You are missing the glue of holding it all together - The operating system for automation and AI

My final advice for you is to have something that glues all of the above together, or as Forrester (2021) wrote -  how there is a “missing automation fabric” for the whole of business that can integrate to multiple adjacent and complementary automation technologies, process architectures, organizational behaviors to support the goals of human-centered automation and an autonomous enterprise.

automation CoEs might push automation towards the business and business processes, however, the CoE acts mostly in a non-digital way, heavily reliant on documents. Thus, the processes are not managed e2e and, the data is not flowing across the workflow.

"There needs to be a “glue” that holds the initiative together"

When driving automation & AI a large portion of your time and investment will not be spent on building automation. It will instead be spent on engaging business stakeholders, generating opportunities, qualifying opportunities, project and portfolio management of execution, operating solutions, and reporting progress and business value in various dimensions, as well as ensuring IT and compliance stakeholders that you adhere to business processes and regulations. For optimization, there needs to be a “glue” that holds the initiative together. The more platforms you have, the more solutions that are live and the more projects you have to manage. This is where the automation fabric, to steal Forrester’s term, comes in handy. This system manages your automation lifecycle e2e while increasing efficiency in your automation operations. Resulting in the need to digitize your automation & AI initiative or CoE end-2-end.

You´re welcome to read more about the different fallacies in the next article within this series, where I will go into more depth on these subjects and where I, based on my experiences, will share what I have learned along the way, and what I would do differently if I did it again today.

 

Sources: 

KPMG, HFS, (2019). Integrated automation why you’ve been doing it all

Forrester, (2021). The new automation fabric is where digital business happens

BCG, (2020). Flipping the Odds of Digital Transformation Success

HFS research, (2020). Spending on automation and AI business operations worldwide 2016-2023