Home » Business Topics » Data Strategist

Artificial Intelligence strategy for SMEs: A pragmatic approach



Is an Artificial Intelligence (AI) strategy feasible for SMEs? (Small to Medium Enterprises)?

In this post, we present a pragmatic approach for SMEs to consider AI (at low cost)

To provide some context – What is AI? – We use the following definitions

Machine Learning means learning from data

Deep learning is representation learning or automatic feature detection i.e.  learning the structure of the data in addition to learning from the data itself

Artificial Intelligence are machines that can reason – currently mostly based on deep learning

We are working with Narrow AI (in context of specific domains) – as opposed to general AI (Terminator!)

So, with that in mind – let’s start with saying ..

  1. SMEs not early adopters of technology (including AI)
  2. AI is expensive in terms of computation, skills, data, delayed payoff etc
  3. And of course we have COVID 19 to contend with!

Most people start with something like chatbots – which is fine – but misses a larger purpose of what AI can do.

Another starting point is the message ‘own your own data’ – but most companies are already aware of data and there are existing regulatory issues around data (ex GDPR) which they must comply with

So, with that background, here are some pragmatic strategies for SMEs to implement AI


Understand what AI can and cannot do (including capabilities of machine learning / deep learning etc)

This stage needs some study but its not impossible to learn AI from a functional (non- technical/ application) perspective. You would know what algorithms do, what problems do they solve and what do they need to solve these problems. You can then apply these t your business problems

Think Process Automation instead of Artificial Intelligence    

Once you understand the possibilities – think automation – instead of AI  I once posted about a Gedankenexperiment / thought experiment where I suggested that you should take a three step approach First, remove all humans – then automate – finally bring back the humans where you will have a different and a more streamlined process. This argument for implementing digital transformation is valid for SMEs also. It forces you to think of business problems in context of AI

Look for use of AI in existing products – especially in high cost areas

AI is everywhere. Once you understand functionally what can be achieved, you can look at your existing products and think about how you can use AI there. Focus on high cost domains such as Cybersecurity, Legal and Human Resources.

Explore Cloud platforms

Explore existing Cloud platforms. Azure, GCP and AWS have all excellent implementations of AI – and these platforms make it easier to deploy AI. On the Enterprise side, Azure especially is mature.

Explore AI for analytics

Every business uses some form of Analytics. Try to use AI for these functions. The payoff in terms of reducing costs could be a lot higher.

Implement machine learning

Considering my definitions above, in most cases, you will be implementing a machine learning algorithm and that’s perfectly fine. More broadly, don’t get seduced by the hype!


In this article, we discussed some pragmatic strategies for use of Artificial Intelligence for SMEs.

To conclude, first invest in pragmatic learning on the opportunities and limitations of AI (when used in the broadest sense including Machine Learning)

Once you understand the possibilities, implement for your business case either using existing products or the Cloud

Image Source: OECD