What We Do and Why!

Expert system development will capture hard-earned wisdom on how to perform a company’s core business functions. They can replicate that expertise into formal processes and a set of interconnected tools, training, and controls, enabling businesses to own that expertise rather than the expertise just being kept in the brains of key employees. This protects your business from the loss of key employees, but it also allows you to replicate this formal, expert-level process in a coherent system.

expert system development

This means that you can scale your expert system because one person is no longer the bottleneck. It also means you can reduce costs because you reduce the level of expertise required to reliably produce the results you want, and you maybe even be able automate much of it. Because you freeze your previous process as a formal process, you can also optimise it, increasing the speed and value of your output. Finally, because you now have an easily reproducible recipe to focus on, you can control consistency and quality, and even create simple business controls to keep processes running.

In a perfect world, you’d love running your own business; your company would take orders and deliver products while you’re elsewhere. While this fantasy may not come true, using this filter to view your business prompts you to continually look for ways to automate, simplify, and improve operations, rather than adding more people as more work arises.

This is our six-step formula for expert system development.


A deliverable is any result your expert system needs to produce in order to meet the expected or promised output of your system. It’s just a fancy way of saying that you need to be clear about what your system is supposed to produce for its customers. This might include delivering a certain number of physical products to your customers by a specific date, reports or recommendations to your customers on a course of action on how to best address a particular challenge, or other output fulfillment promised by your expert system.

The more complex the expert system development you are dealing with, the more deliverables you will capture in the list. You’ve made all of these deliverables, but you’ve done it in an informal way that one or two key people in your company just “did it.” By specifically defining your deliverables, you’ll take the critical first step toward building a system that can efficiently produce them. After all, how can you produce something that you didn’t consciously identify as you committed to creating?

Without documenting all of these deliverables, how can you convince clients that you did the job and delivered on your promises?


Step Two: Process Lay Out

We lay out the process your company will use to create and deliver on all those deliverables.   Once we have got a rough layout of your steps, it’s time to ask a series of questions to refine your draft process:

  • Which services and deliverables are really important?
  • What results are nice but not essential?
  • What benefits do your employees think your customers want or have asked for but don’t have?
  • How can you eliminate these results, which are really just annoying and unnecessary?
  • How can you shorten the steps and still generate the desired results? And generate an improved result?
  • How to reduce the necessary resources and still generate the desired results? And generate an improved result?
  • How can you speed up this process?
  • How can you automate or model this process, or part of this process?
  • How can you reduce the costs of doing this process without affecting the value of the output?
  • What simple changes or improvements can you make to add value to the product?
  • How can you slightly increase the cost of production, but in such a way as to increase the value of the product so that you can get a price increase for the value that you offer now?
  • Who else in the world has a similar process or tool that you can learn from to help you better design this process?
  • Could you outsource parts of this system? Does it really make sense to do this in the long term?
  • How could the system be made more robust? More stable? Less prone to errors?


Once these have been thought out and these questions have been answered. Based on your questions, we move, add, remove, and play with steps in your system until we deliver a process that promises to be faster, cheaper, better, more impactful, and more scalable.

When we have prepared these improvements, and have laid out the finished process as a complete and accurate recipe for obtaining the desired results. This recipe will be a simple list of each step of the process.(e.g., Step One . . . Step Two . . . Step Three . . . etc.)


Step Three: Determine the optimal level of knowledge for each step

As a business, you want to find ways to relieve your most expensive and experienced employees from doing lower-value jobs. To sustainably scale your business, you need to strive to push most of the work of each expert system development down the value chain so that your experts do less than your expert system. This not only immediately increases your capacity because your expert can be spread over a larger volume of total work, but it also reduces your costs since work done at a lower level is much less expensive.

There is a knowledge hierarchy for each step in your expert system. At the bottom, you have those steps that can be automated, semi-automated, or turned into a template. Then you have those steps that require a person to complete, but not necessarily a skilled person (eg clerical, administrative, unskilled worker, etc.). The next level is for those steps that require a semi-qualified team member (eg paralegal, nurse, apprentice, etc.).  Above that is the skill level, which requires a basic expert to produce these steps.

Finally, the top level of the pyramid is for those steps that require a high-level expert to produce for the business, which in most small businesses is either the owner or one or two key employees.

For example, let’s look at how this hierarchy of expertise works in a law firm to keep these levels clear. Automated, semi-automatic, or template refers to things like a standardised engagement letter that is sent to each new customer, or a library of contract templates they have on the company’s server. Non-qualifying tasks include tasks that an office worker without legal training could perform, such as scheduling meetings, collecting customer data, and collecting historical documents for submission to a lawyer. 

Semi-skilled tasks in this context are likely to refer to those subjects that a paralegal, rather than a real lawyer, can do. Basic expert tasks are those that only a licensed lawyer can perform, although they can be performed by a less expensive and less experienced junior lawyer. The best expert tasks are the processes and functions that require the best legal talent in that firm in that area, things that would likely take years of experience to understand and do correctly.

The purpose of your expert system is to match each step of your expert system development with the appropriate level of expertise.

For most existing companies, the greatest immediate reward for writing an expert system in a given area is the way this exercise shows where they need to move down from the top two levels (Primary Expert and Supreme Expert) to the lower levels. In many cases, a company can quickly increase its capabilities by 30-50% or more simply by reducing many steps of its expert system development to a lower level of the pyramid.

This could include, for example, things like involving administrative staff in scheduling meetings and reminding customers of information they need to get to the company before that meeting, better software that automates or semi-automates billing customers for work performed, or standardisation of the main legal services to allow a less experienced lawyer to perform tasks previously performed by the firm’s senior partners, but which do not require the latter’s in-depth experience.

Most companies have their most expensive top “experts” doing too many steps of their informal expert system. As a result, they struggle with capacity issues and low margins, and are vulnerable to this “expert” getting hurt or otherwise leaving the business, taking with it all the know-how and institutional knowledge that they acquired over the years in this business area.

If you want to expand your business exponentially, you need to reduce your company’s dependence on any expert with formalised expert systems that you have cross-trained and trained your team on.


Step Four: Consistency check

Now that we have a written process in place and have determined what level of knowledge is best for each step, it’s time to refine the process to control consistency. This is just another way of saying that now you need to look for ways to improve quality and reduce variability. Here are a few key thoughts that we use for this:

  • The more you can automate, semi-automate, and template, the easier it is for you to control consistency. All you then need is a rigorous review of your template or automated steps to ensure they are correct. These processes become great “embedded controls” to protect your business.
  • Streamline the process. The fewer steps in any complete process, performed by fewer people, the fewer potential problems.
  • Pay special attention to critical links. Write critical links between tasks and reinforce them.
  • Standardise whenever you can. This will help you speed up the process, increase efficiency, reduce costs, increase impact and improve quality.
  • Create your three “master” documents: master timeline, master checklist, and master budget.
  • We collect institutional knowledge in a structured and searchable place. This includes detailed customer notes not in the minds of your employees, but in searchable text in your CRM. It may also include an organised file of related documents for a particular project or client. If you don’t take action now to capture this essential past story, there will come a day when a key team member will leave your company and you will have to scramble to recreate the institutional knowledge they took with them. Not only will it be financially expensive, but it will also be incredibly stressful and emotionally painful.
  • Train and cross-train your team. We make sure that each key role in your expert system has at least one fully trained understudy. This will give you incredible peace of mind and business depth.


Step Five: Plan the key components of your expert system development, to improve first

It’s more than likely that you don’t currently have any formal expert systems set up, but you do have an informal set of best practices that are in the minds of key members of your team. Now that we have followed steps one through four to formally set up the expert system, it is time to complete the system with the tools, training, and checks needed to improve the expert system. It doesn’t happen all at once, but develops over time.

We start by choosing the piece of your expert system development, the “block”, that you think would have the greatest impact on your business. Picking this block and successfully modeling it in an expert system development will give your staff a visible example of the value and operation of expert systems, and increase their confidence in attacking the next, more complex blocks.

Give this block a title (e.g. New Client Launch, Quality Review Process, Proposal Selection Step, etc.) to make it easier for you and your team to talk about it.

We approach this block from four specific directions:

Critical knowledge to institutionalise: What is the critical know-how about that block of your expert system development that is locked in the heads of one or more key team members? Identify this institutional knowledge and think about how best to capture, store and share it.

Tools to improve and leverage: What tools, patterns, and automation would make this block of your expert system faster, cheaper, and better?

Design and implementation training: What training and cross-training will team members need to successfully use this expert system (or at least this block of your expert system)? How can we formalize or “freeze” this training in easily accessible, updatable and scalable systems?

Controls to Monitor and Ensure Quality: We look at which internal controls, visual, procedural, or built-in, would best help your business ensure that this block of your expert system always performs exceptionally well.


Step Six: Every quarter re-evaluate your expert system to prioritise the next block to improve and refine

Every quarter, we re-examine your expert system development and choose the next “block” to focus on and refine. It’s normal to take three to four quarters to really build a complete expert system for your business. Because these are the processes that produce the most value in your business, you’ll find that your business benefits greatly from the time and attention you invest in each block as you and your team iterate and refine for the long haul.


Get In Touch

For more information about Makarov Systems, De Morgan Cyber Security, De Morgan Web Design, or our development and research areas. Please contact us using our online form.

This website uses cookies. By continuing to use this site, you accept our use of cookies.