What’s a CoE, anyway?
A CoE (and we will use this all-encompassing term going forward for all such initiatives) is something that is built with certain business and/or technology goals in mind.
Organizations often use the strategy of focusing on few specific technology areas and banking on them for growth by aligning all their resources (man-power, investments etc.) towards it. Focusing on one or a few things always works out better than doing too many things at a time and not getting a return on investment (ROI) on any of them.
Usually, the high-level goals of building a CoE are:
1) Turning vision into action
2) Riding the hype cycle of a specific technology area or a vertical to grow business
3) Using CoE as a vehicle to take the organization to the next level of scale in terms of revenue, market share, etc.
4) Edge the competition out by being better/superior to them in a specific technology area
Before we move on, such CoEs are conceptualized in organizations in the following two scenarios:
1) Organizations that offer product engineering services or product development services set up such CoEs. It is because they want to diversify their offerings for technology stacks and be in line with the changing technology landscape to aid business growth. Product companies may also setup such CoEs to focus on specific technologies that their products are using.
2) Service providers and system integrators (SIs) set up CoEs in organizations (or Enterprises) for consulting and training on that service or platform (for example, Sales Force or SAP). The purpose is to provide a comprehensive guidance regarding use of a specific technology, create organization wide standards regarding these technologies and to spearhead the development in those. These are the people who do not work for a single project or product but provide internal consulting to all teams.
In this article, I present a generic methodology that can be used for building a CoE.
As always, this blueprint is based on my endeavours (or adventures) and experiences of managing and building CoEs. So, this blueprint, in no way, is THE blueprint to be followed. You are smart enough to come up with your own, as long as you know you will succeed!
What’s your CoE?
CoEs could be built for any of the following:
2) A specific vertical or a domain: An organization may want to focus on healthcare as a vertical and HMS (hospital management system) as a specific sub-vertical. Or the key leaders in the organization (you could be one of them) think that there is a good buzz about IIoT (or industrial IoT) as a sub-vertical or area and hence they may want to build a CoE for that.
3) A specific software or product offering: There are multiple kinds of product types available for serving a specific customer need. For example, CRM, ERP, CMS and so on. So, an organization may want to build expertise in such specific software viz. CRM.
Do you have to be a CXO (CTO/CIO/CEO) of some company to build a CoE?
Well, yes and no.
If you are a CXO level person, then you have access to all the resources to make decisions. However, even if you are a project manager or a developer, you can still use this blueprint to build expertise at your individual level with respect to some technology area. The core principles would still hold true.
While building a CoE, you would apply a breadth first; depth later approach since you would want to gain an overall understanding of a technology first, and then go deeper in any specific area.
The CoE Blueprint
Following is the blueprint for building a CoE. It has 4 phases as shown below.
Figure 1 : The CoE Blueprint
The analogy used here is how you start crawling in your childhood, to how you eventually grow in strength and confidence to run fast and win races (not actually, but analogous to achieving your desired goals in life).
In the upcoming sections, for each phase, I will elaborate on:
- the pre-requisites for the phase
- various steps to be taken
- the timeframe needed, and
- the KRAs (the key result areas) to be fulfilled in each phase.
This looks cool, eh?
But the proof of the pudding is in the eating. So, while we discuss the blueprint, let us also apply this model and see how we can build a CoE for one of the hottest and happening technology areas – AI or artificial intelligence.
Let’s call our CoE for AI as AI-CC.
In upcoming sections, at the end of each phase, I will mention about how that phase would pan out for AICC. This will give you a good idea of how this whole conceptual model works in reality. Please look for an AICC section for each phase.
Ready to do some exercise?
Phase 1: Crawl
Before embarking upon the journey of building a CoE, you need to have a high-level idea about a specific technology or vertical in which you would like to build a CoE.
Although it is important to know where you want to go, it is not necessary that you will wind up there. At this stage of building a CoE, you are allowed to go a few steps forward (or crawl for some distance) and then stop and move elsewhere (i.e. pursue other technology area).
Another important aspect is to get a buy-in from key stakeholders about you or your team wishing to spend time on exploring a specific technology area. This aspect will drive the decisions later to funding/investment once the CoE takes some shape (especially the POC-phase).
To choose the technology area, you need to have a clarity about:
- Why is that area important to you, business growth-wise? Is it because everyone is building CoE in that area or it’s because you are convinced that there is a future in it?
- Does that area have business potential and have you validated it? You can refer to the concepts of blue ocean vs. red ocean strategy to understand this aspect better here.
- As an individual, do you think you will benefit from learning about this area as it will add to versatility of your skills?
The simplest way to get a clarity on this is to read up about industry trends, upcoming tech/business areas and so on and form your opinion based on that.
If you are already an avid reader of technology, this should not be too difficult for you.
Crawling into the CoE
Following are the steps you would follow in the crawling process:
1) First and foremost, you do enough market research to know about various possible technology areas or verticals. You would end up reading technical magazines like DotNetCurry (DNC). You would even check out what your competitors (in business) or peers (at an individual level) are learning about.
2) After zeroing in on 3-4 areas, now is the time to read up or understand more about specific technology areas and compare and contrast.
3) Typically, the crawl phase would go on for 4 to 8 weeks.
Once you have done this, you can assess the effectiveness of this phase by referring to following KRAs.
The key result areas for successful completion of this phase should be:
1) Have a reasonable understanding of the spectrum of technology or vertical in which you want to build the CoE.
2) You should be able to have a 10-15 minute elevator conversation with anyone about this technology or vertical after this phase is over. If you can have this impromptu conversation with anyone who understands technology and if that person thinks you are comfortable with a specific technology, then, my friend, you have cleared the 1st phase with distinction!
3) You should have some scenarios identified that can be a good starting point for doing a POC (i.e. Proof of Concept).
Following are some important considerations in finalizing the scope of a POC:
1) The scenarios shouldn’t be too complicated in terms of functional or non-functional aspects as a POC to demonstrate these aspects would eat up a lot of time.
2) They should be relevant for customers/consumers of your organization so you can demonstrate the usefulness of that technology area well enough.
The AICC Crawl
In the context of AICC, we have already completed part of the crawl phase by choosing to build competency in AI. However, to validate the model, you will need to:
1) Get a good understanding about why and how AI is becoming so popular. You will find umpteen articles/blogs about this aspect. Once you do, please take up the 15 min elevator conversation test and see how you perform.
2) Next up, you need to come up with scenarios for building a bot using an AI framework to address a specific business need or requirement. Here is a quick-start link for you.
Phase 2: Walk
Assuming that you have crawled up to here (not literally), you would now have a good understanding of the technology or vertical in which you are building a CoE. It’s one of the key pre-requisites.
But that is not good enough!
What you need now is that the key stakeholders in your organization (basically your bosses) are aligned well with the CoE objectives. So, you need to get a buy-in from them. It is in this phase when you need them to sign off on important things like strategy to build the CoE, corresponding investments needed etc.
The key aspects that you would need to make this phase a success are:
1) Having people with right skills to build a POC or POCs for CoE. Usually you would hire some experts from the market and bank on their knowledge or experience. In addition to this, you identify smart people from within the organization and train them on respective technologies. The reason to follow this two-pronged strategy is that it keeps the costs down and it also helps in building the team from within. Initially you may want to start by building in-house expertise.
2) For specific technology areas or verticals, you would also need to spend money in having licenses for right software, infrastructure and so on. So, if you are going to build a SAP CoE, you would need to have some money allocated to acquire licenses of SAP.
While doing so, it is important to start small in terms of investments – be it people or tools/licenses. It helps to start with a small core seed team of people and then add more members once you start seeing traction.
When you are looking for people within the organization for building a CoE, you need open-minded creative people who don’t mind doing something new every day and throwing it off the next day just because the priorities changed overnight.
You usually wouldn’t want people who like doing fixed set of things every day. They are useful for your CoE when you have many projects coming in.
In this process, you will be doing the following:
1) Since you have a good understanding about what the technology can and cannot do and also what the market needs are, you would spend time in finalizing a business scenario (or scenarios) to implement a POC. The objectives of building a POC are:
- a. The business use case chosen is simple enough to implement in a short period of time (remember, a POC cannot run for months).
- b. POC should not take more than 3 to 4 weeks to implement.
- c. You may not need to have experts to do the POC and you can use internal resources to execute the POC.
- d. You need to validate the POCs by seeking feedback from customers or potential customers and perform course-corrections if any.
2) Typically, it should not take more than 1 to 3 months to complete this phase. The KRAs that you need to accomplish in this phase are as follows.
The KRAs for this phase are:
1) At the end of this phase, you would have a working POC that you can use to demonstrate the use of technology to solve a business problem to stakeholders including internal (CXOs) and external (customers) entities.
2) Another key deliverable of this phase should be a business strategy that elaborates as to how you’d like to use the CoE to grow the business multi-fold. This should include:
- a. What would be the key investments?
- i. People with specific skills: either hire them or train existing ones
- ii. Infrastructure: licenses for specific software or a specific hardware or anything else (e.g. various devices if you are going to build CoE for IoT)
- b. How long would it take before the organization would start seeing the benefit of investment? If you are setting up a CoE for any technology area, you should be patient enough to see these strategy bearing fruits. This period could be anywhere from 1 year to 1.5 years based on the technology area, market conditions and so on. However, it is important to have a common understanding amongst all stakeholders about the timeframe for which they can be patient!
- c. While you are working on a positive outcome, you should also be defining the cut off period for calling the CoE initiative off. Remember, your plan B, is God’s plan A. So, you need to also be ready for handling failures or setbacks.
3) Since you have spent some time in understanding the nuances of the technology well, this is also a good time to get some certifications under your belt so that it adds some credibility to your expertise. Most technology or platforms have some or the other partner program as well. For instance, your organization can become a Gold/Silver certified Microsoft partner by having some certifications and some customer references.
Walk the talk - AICC
You will have to build a POC to complete this phase and you will also have to come up with a business strategy here. It will be practically impossible to elaborate both of them in context of this article. However, I am providing some pointers below.
For building POC, you can check out the various scenarios mentioned earlier and use any of existing AI frameworks (DialogFlow or MS Bot Framework) and build the POC.
With respect to business strategy, you will have to spend time in thinking through this aspect well and build a strategy that includes the following:
1) Mission statement & goals
2) SWOT analysis
3) KPIs to track the success
4) Target customers
5) Competition analysis
6) Sales plan
7) Team composition
8) Operations plan
9) Financial projections
Phase 3: Jog
The key pre-requisites in this phase are as follows:
1) You should already have a team of people available who have some experience or knowledge of the technology. It should be a good mix of expertise and experience. So, you would have some key experts for a specific technology area and then you would train some people from your organization.
2) Since your CoE would be executing the projects in given technology, it is absolutely necessary that you have a plan in place that handles the scaling of operations well enough.
If you have reached this phase, it means that your CoE initiative has started to bear fruits in terms of business growth. Essentially, in this phase, you would have multiple projects running on the CoE offerings for different customers.
The key aspects to note are:
1) With your organization seeing a good traction on a specific technology area, you need to marshal your troops well in terms of creating seed teams and building satellite teams around them.
2) It is possible that you may get projects in specific sub-area of the technology that you have chosen. In that case, you need to align your energies and resources to focus on that specific area. Let’s say that you have started a CoE on Azure and you focused first on PaaS. However, you have customer projects/requirements looking for IaaS offerings. In such a case, you need re-align your teams accordingly to handle that kind of requirement first.
In some variations of CoEs, it is in this phase when projects are handed over to execution units aka business units so that the CoE can focus on enhancing the competencies built and works pan-organization at times.
The success factors for this phase are:
1) You now have a good enough knowledge about the technology since you have executed multiple projects. So, in a way, your CoE is like a well-oiled engine operating at full throttle.
2) While you are executing multiple projects, you are also required to keep on updating the offerings and case studies to ensure that the variety of work you are doing in specific technology area is captured and that can be shared with future prospects.
Jogging in AICC
In this phase, you have multiple AI projects already spanning across different AI platforms like DIalogFlow, Microsoft Bot Framework and so on. Since the core principles of AI platforms remain the same, you should be able to execute a project in some other AI platform as well.
Phase 4: Run
You are a champion now!
Key pre-requisites include:
1) You have significant knowledge and experience in a particular technology area now and you are able to build on that in order to be called as an expert in that area.
2) Your teams have worked on different projects and that has resulted into your CoE being rich in terms of versatility of specific technology areas. However, please do remember that technologies evolve. So, you have to continue sharpening your skills so that you do not fall behind when compared to competition.
After having spent so much effort and time in building a CoE, in this phase, it is expected that you have the required skill and knowledge to be called upon as an expert. So, the CoE that you have built would have to make the following contributions:
1) Come up with best practices for specific technologies and drive them across organization
2) Have whitepapers/blogs written to demonstrate your technical prowess. In some cases, you may even apply for patents etc.
3) Work as internal consultant for the teams that are implementing projects in the same area so that they are more profitable to the organization.
The key result areas that you would be fulfilling in this phase are:
1) You are being identified as a subject matter expert (SME) in a specific technology area.
2) You have a good understanding on the technology and authoritatively talk about best practices for executing projects in a specific technology area.
Running in AICC
For your AICC, you are being looked upon as AI expert and are appearing in various events. If that is happening, then you have achieved the goals.
Some Dos and Don’ts:
Following are some dos and don’ts when it comes to building and running CoEs based on my experience:
1) First and foremost, it is absolutely essential to have a complete buy-in from various stakeholders, right from the Jog phase. Any confusion or misunderstanding over the goals to be achieved could lead to failure of CoE’s effectiveness. Communication plays a vital role in making this happen and it is important to have periodic (at least monthly) meetings with stakeholders where data/details are presented about the CoE progress. This is when you can also discuss about gaps in understanding and have any course-correction done.
2) You need to be patient in the initial phases of CoE when you are building the competencies.
3) Usually CoEs work well when you have a core team assigned to it from day 1 and it stays on like that till the Jog phase. Frequent changes in this team would result into disruptions in the end results as well.
4) A CoE is a front-loaded initiative i.e. you have to first incur costs and then you realize the returns when things start happening for good. So, it is required that all the stakeholders are well aware of this aspect.
5) While you need to be patient, it is required that you are practical as well. You simply cannot endlessly keep on investing. You need to define your cut-off period in terms of realizing the investments on the CoE.
6) While choosing a technology, it is obviously expected that you choose one that has a good market traction. Neither should you be choosing something old or dated nor should you be choosing anything that is way ahead of time.
Candidates for CoE
Based on the current technology landscape, following are some candidates for CoE. Needless to state, this list is not exhaustive .
1) AI/ML: This one has to be on the top of the list.
2) Containerization technologies: Docker/Kubernetes
3) Test automation: keyword driven automation
Do you know of any technologies that are good candidates for CoE. Use the comments section or tweet your thoughts to me @Rahul1000Buddhe
You would have realized by now that building a CoE is not a child’s play. However, with a good focus, commitment and dedication by right stakeholders, it is not very difficult to make it a success either. Depending on how you execute the blueprint, your CoE can be Canter or Cakewalk to Excellence.
The choice is yours to make!
C# and .NET have been around for a very long time, but their constant growth means there’s always more to learn.
We at DotNetCurry are very excited to announce the pre-order of The Absolutely Awesome Book on C# and .NET. This is a concise technical eBook and will be available in PDF, ePub, and mobi.
Organized around concepts, this eBook aims to provide a concise, yet solid foundation in C# and .NET, covering C# 6.0, C# 7.0 and .NET Core. Use these concepts in your next .NET Project or to crack your next .NET Interview.
Click here to Pre-Order this eBook at a Discounted Price!