Computers became popular because of games, Internet – because of porn. A technology becomes fashionable only when it’s easy and exciting enough for the average user. Unfortunately, AI projects up to date were mostly technical and complex for a businessperson. AI is still a sandbox for geeks, not a real-world business tool.
Technically, expert systems are a part of AI. What makes them extraordinary is how easy it is to design a simple real-world expert system and get first tangible results. It’s the fastest step towards “real” AI. We don’t need to wait for the technology to become easier, we just need to start with the easiest piece of it.
Expert systems don’t require hiring tech people or even buying expensive tools. You can design your first expert system with a pen and a napkin. We do that all the time when planning, we just need to go further and make these to-do lists reusable and scalable.
It’s called “knowledge engineering” because we translate our knowledge and expertise into a process to achieve certain result. We just don’t see it as engineering because we have got so used to it.
The most exciting idea that made expert systems popular and effective was separation of knowledge from the software code. It enabled knowledge engineers to focus on acquiring, accumulating, reusing and scaling knowledge and expertise instead of writing code all days long. We no longer need to update software code whenever we extend and update the knowledge base, unlike in conventional software development, where knowledge is hard-coded into algorithms.
Classic expert systems are rule-based, they help experts and less experienced operators to make decisions. But what if we extend the definition of expert systems to all processes that require some level of expertise?
It makes total sense because “What to do next?” is a decision, too. Even if there is just a single option available at the moment (the next step in the to-do list), it’s still an important piece of knowledge.
It enables us to start small, not using any fancy tools or hiring expensive IT people. All that we really need is just a pen and a piece of paper. It’s more than enough for the quick start. Then, as our first expert system brings more and more real-world results, we can gradually convert it into an electronic form and make it dynamic whenever needed. But the most important stage is still the beginning, when we need to make the very first step into the right direction.
We often hear that “Those who organize marketing processes grow faster than those who don’t“. It’s true, but it doesn’t need to be complex and expensive! Complexity usually kills IT projects, not the price.
Making the first step as simple as possible for ourselves, not for others, must become our top priority as marketing managers. If we can’t manage our top-level marketing processes easily, will they be managed better at lower levels?
Making it easier isn’t just an option, it’s the only choice in most cases. We don’t have unlimited budgets and time for fancy technologies, but we still need to evolve faster than our competitors towards what will become new reality tomorrow.
So, you don’t really need to make the choice between expert systems and AI. Expert systems are just the shortest way to AI. In reality, there will be not that much AI at the end because marketing tasks aren’t about intelligence, they are mostly about real-world experience.
If you start to translate your marketing experience into pen-and-paper expert systems right now, you will become much closer to full-scale AI than you can imagine.