Establishing control of your operating model

Alexander Hübel

February 4, 2022

Establishing control of your operating model

-       The 4th in a series of 10 articles - How To Successfully Scale Automation and AI in an Enterprise

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 ensure 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 fourth in a series of 10 articles where I want to share what I learned along the way and what I would do differently if I did it again today. This fourth article continues on the theme of setting your operating model. In this article, I deep dive into what I believe are vital parts of the operating model.

My six pillars of the operating model

In the last article, I did a deep dive on the topic of having a centralized vs. a federated or decentralized operating model. The ones of you that read that article understood that I lean towards a more centralized model in general, however, taking the pros and cons of each model into consideration.

No matter what model you choose, the areas outlined in this article need to be done correctly and be well-aligned for your initiative to succeed. When it comes to "establishing control of your operating model," I am referring to the question of defining; "who is doing what" in the automation journey. My answer to this question lies in the six pillars pictured below, and I will describe each pillar in more detail. Again, there is no right or wrong when setting the model, but different models will enable various levels of scalability. The below picture aligns with my thinking around the operating model:

Illustration - Operating model

1.Methodology and value tracking

I believe that a somewhat standardized way of working and tracking value is vital. If you cannot track the value, you will be questioned on "whether we should continue to invest in this" and "why" - which you need to answer. This can be hard to accomplish. Try to answer this question, for example; can you measure your automation improvements and link them down to improvements in the profit and loss statement? I assume you can't, and that's fine. However, do expect that question from parts of the c-suite. No matter what, you will have to articulate what value you plan to bring and what value you have delivered. For this to be possible, you first need to ensure that you collect the baseline data for each opportunity. Secondly, you need to quantify business cases from value and cost perspectives. Thirdly, you need to follow up on these business cases to the highest degree possible for each use case (or group of use cases grouped to solve a particular problem). Hopefully, you don't only do this on paper but also match it with actual transactional data from your processes and solutions.

You need to have proof points, and you need to have a mechanism to follow up. You need to have a defined company-wide way of working that includes benefit tracking and realization. It is also essential that the way of working includes defined governance structures on the strategic, tactical, and operational level as well as a methodology for the end-to-end workflow of automation & AI. Here, it is vital that the methodology includes the identifying, prioritizing, building, deploying, and operating of automation & AI solutions and their corresponding IT platforms.

Illustration - operating model

2. Business transformation & value realization

The highest level of the operating model is the business transformation. The business transformation needs to be owned by the business and its executives - a Centre of Excellence (CoE) cannot be responsible for transforming a line of business. However, don't expect that each line of business will figure out how to transform themselves by leveraging automation & AI. Instead, the CoE needs to support and act as trusted consultants to the executive team, CxO, business unit heads, and others driving the business transformation. The CoE needs to push and proactively listen to the business, understand their priorities while challenging them and their execution. All this while managing the complexities of working with new technologies, figuring out how to measure and track the value, and addressing turf issues with IT. I believe that one reason why automation and AI initiatives fail is because they do not manage to stay relevant enough on the executive level.

Illustration-operating model

3.Opportunity identification / discoveries

The next layer in the operating model is where and how to find opportunities to automate. This is where the automation & AI CoE and the business marry and come together as one team. The fallacy of believing sustainable automation can be delivered by the IT department alone has been written about by many (McKinsey, 2019; Deloitte, 2019, EY 2017). It does not matter if you are the business, the center of excellence, or IT; you need to work as one team. Another article in this series will dig deeper into the topic of optimizing your pipeline, but at the" Large company," this is how we did it;

We filled the automation & AI CoE with transformational people, often with management consultancy backgrounds or with significant experience from the line of business. We then equipped them with a growing portfolio of automation & AI products & capabilities that could solve problems within the business. We also invested in building a pipeline for each stakeholder (refer to an earlier article about Champions). This investment in building the pipeline was not a one-time effort; instead, it was continued investment in ensuring there were always many use cases to be developed. A scalable delivery, defined governance structures, and ways to measure and communicate value were vital.

I expect that many companies underinvest in meeting the senior business stakeholders with relevant transformational automation & AI competence to solve their business problems. Treating this as a standard IT project, asking someone to define requirements will not fly. Neither does it work paying a consultancy lots of money to do the job for you; once they leave, you are back to square one. Instead, you need to build relevant transformational capabilities that can talk to senior business leaders, understand their pain points and figure out how automation & AI can solve them and how you can create business value. You can complement that capability by consultancies or by leveraging junior or inexpensive resources to work on the detailed process mapping, volume gathering, and solution designing.

It is critical that the central team is a strong muscle keeping the transformation aspects together. This is the oil in the machinery. Suppose the central team is not connected through the enterprise. In that case, you will likely not succeed in transforming at scale but be much more dependent on individual sponsors in specific business units and functions to move the needle. Building a continuous pipeline for a wide array of hyperautomation technologies, which is key for automation & AI transformation to succeed, will require this solid central focus. If not, your potential centralized investments in delivery and platform capabilities will not have enough business to sustain.

Illustration- operating model

4. Delivery

The next level is delivery which is about building the solutions. I assume most companies today have some level of DevOps; hence the lines between delivery and operations will not be static. Building automation & AI solutions are similar to any project about managing stakeholder/customer expectations, timelines, cost, and quality. Some are straightforward like RPA projects, some are much more experimental like ML projects, and you will need different types of technical skills to master them. Here it would be best if you had the competence to build and how to make it scalable. This competence is available in the market and can be bought; it will not need the same strong connection to the specific business as the transformation layer does. The cost will vary depending on how you can leverage different geographies to source talent. Let me outline a few success factors that will set you apart from most organizations enabling you to scale from 1->10->100->1000s of solutions/delivered projects over a few years.

  1. Hire strong tactical competence to manage the portfolio of projects. Traditional portfolio, program, and project management skills combined with a few years of specific automation & AI competence.
  2. Ensure you implement a financial model that allows you to scale (like a business-funded model where business lines pay for each project). I will go further into the financial model in my next article in this series. 
  3. Build at least a basis of capabilities in house to ensure you secure needed competence long term (In the large enterprise, we decided to target approx. 50 % of inhouse delivery capability)
  4. Balance demand and supply by leveraging partnerships with one or more vendors. This gives you the flexibility to scale up and down depending on the pipeline without risking a bench of internal resources
  5. Thoroughly link the delivery to business transformation needs by implementing the proper measurements and governance structures
  6. Continuously measure and follow up to ensure you can drive continuous improvements in the delivery process (otherwise, you risk your projects becoming too expensive!)
  7. Try to streamline but also realize that each technology is a bit different. You will need different capabilities and methods when delivering ML vs. RPA type of projects
  8. Plan for ways to organize yourself to have the ability to over time deliver on larger, more complex projects requiring to combine echnologies into "Hyperautomation" solutions. When scaling, you need rigor, defined, and standardized processes to ensure re-use, cost competitiveness, and compliance in the delivery area. With the correct tactical management, that will come as you scale – you for sure don't need to set everything before you start.

Delivery can be centralized or federated. If you federate, you might move the delivery closer to the actual day-to-day operations, which can be good for many reasons. However, that will not enable you to fulfill most of the critical points above, requiring scale. Significantly leveraging a wider technology portfolio and not only simple RPA will need a more centralized way of working. Having a central team requires that the team works very closely, hand in hand, with the business.

Illustration- Operating model


The fifth level is your operations which I believe also makes sense to centralize. Managing bots and algorithms might sound simple in theory, but learnings from enterprises show that it is complex and costly. Many customers that I talk to in my current role state that costs increase almost linearly with new implementations. This is not sustainable, and likely after your first couple of hundred bots, executives will stop seeing this as an innovation program and more as any other operational entity. At that point in time, continuously increased run costs will not be accepted; instead, there will likely be targets for reducing costs.

Here the job is about creating a model that enables an ever-increasing volume of solutions to go live without increasing the costs of managing these solutions linearly. This will, over time, require managing automation & AI operations in an equally professional way as other major operations. Eating your own medicine automating the processes and procedures of managing automation is a must!

This requires strong operation leaders who can balance different pressures on cost, business expectations, stability, and security in close coordination with business stakeholders, transformation leaders, and delivery managers.

llustration- Operating model

6. Platforms

Finally, the pillar you really should centralize is the one regarding your platforms. You need a centralized platform strategy to avoid having different functions and businesses buying their own platforms ending up likely buying similar platforms in multiple other places, which will not drive any opportunities for synergies.

Managing automation platforms is close to managing other IT applications. This piece has two major stakeholders; 1. The business and the CoE parts engaging a business; 2. IT. There are two upcoming articles in this series about Hyperautomation portfolio strategy and the trick of managing IT as a stakeholder when scaling automation & AI. The person who leads automation & AI platforms needs to manage and balance both of these aspects, as they likely will be contradictory to each other. Business stakeholders and the rest of the CoE will expect functionality to solve the business problems to be delivered at a reasonable cost. The job of the Platform leader is to enable that, and that is why automation & AI exists. Bringing the functionality is about understanding the business pain points and how the right technology can solve those. The sequence of onboarding a wider set of platforms will differ business by business, depending on their specific pain points. Some will live long on only RPA, while others will have to rely on a broader toolset to scale earlier on.

The cost side is essential to keep track of, as onboarding a wider range of platforms will undeniably create increased cost and complexity in the backend regarding upgrades not only of these platforms but also as the wider enterprise stack that these platforms operate on is upgraded or changed. The cost piece can be removed from the equation from a business perspective by centralizing those costs into a wider central IT budget or fully charged out to create increased transparency and a consumption-based internal price model.

The second stakeholder is IT. IT will demand control of security, compliance, architecture, cost, and stability. They will most likely value those aspects higher than any business value generated by transforming business with automation & AI. The larger and more complex organization you operate in, the trickier this will be. Therefore, it is important to get buy in from IT to be able to operate the AI and automation platforms on top of the enterprise systems. Selecting a platform leader that can influence the enterprise and IT architecture principles could be one way to go. Operating the automation & AI platforms together with IT is a vital part for scalability and value generation to the business.

The platform role also includes strategic aspects such as brokering the right scalable deals with the product vendors. Pay as you grow models or enterprise deals? I have no personal preference; it's all about the situation and timing. I have learned that you will likely not scale as quickly as you expect when negotiating these enterprise deals with the vendors; hence, you risk paying too much for quite some time until you have the scale you need. On the other hand, such a deal could speed up a lot, especially if you can take platform costs centrally. A good strategy could be to explore enterprise deals once you have mastered the technology. 

To conclude, in parallel with setting the enterprise-wide strategy, it is essential to decide who does what. Establishing control of an operating model for automation & AI across the company is, in my opinion, based on six pillars. I believe you need to have a somewhat standardized way of working, and you need to have a way to track value. Then you need to make sure that the proper mandate and competence are accountable for the business transformation. After, you need to set a structure for your opportunity identification, discovery, development, operations, and your platforms. The operating model also needs to be adapted for scale and ensure that factors like financial model and funding play to the advantage of the planned transformation. The topic of the following article digs deeper into this topic and how to nail your financial model.


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

McKinsey, (2019) Automation at scale:The benefits for payers

Deloitte, (2019) Understanding the challenge of implementing your virtual workforce

EY, (2017) 10 common pitfalls to avoid when delivering Robotic Process Automation projects

HFS research, (2020) Spending onautomation and AI business operations worldwide 2016-2023