Proof of Concept for your CRM

John Eccles, 18 January 2021

CRM systems are well established technology. They have been around for decades. So why would anyone need a proof of concept for a CRM?


What is a Proof of Concept?

A proof of concept (POC) involves an exercise to test the real-world potential of an incomplete idea. It is not about delivering the idea; rather about demonstrating whether the idea is feasible. It should be used in the early stages to show whether a product, feature or system can be developed.

A proof of concept is often confused with other intermediate systems. The following table shows the types of intermediate systems and where the Proof of Concept fits.

POC vs Prototype, Pilot and MVP




A demonstration to confirm that concepts or processes are feasible for real-world application.

To validate functional assumptions, to validate technical feasibility, to identify potential blocking points, to determine the customisation efforts or to detect potential performance issues.


An early version, with a limited number of characteristics, built to validate a concept or process

To trial the concept and to provide subsequent specifications for a real, working system rather than a theoretical one.


The first production version of a concept or process.

To test if a concept or process works as expected, in a real, working system with a limited group of users.

Minimum Viable Product

The essential core system with no frills.

To validate that there is a need for the system and that users will use it in practice.

The most appropriate system depends on the purpose. For a Proof of Concept, the focus is an idea for an aspect or component of a larger system and the purpose is to confirm whether the idea will work.

I found the following diagram helpful. It’s from an article by Kelly Duggan, Proof of concept, prototype, pilot, MVP – what’s in a name?


Application to CRM

New technologies are now available for users of enterprise-level CRMs. Some of these technologies offer significant benefits – if they work in your particular application. The technologies themselves are evolving – they may not deliver value at present – but maybe after the next release….

Here are some examples from Dynamics 365:

1. Scanning of data into the system with text recognition.

AI Builder includes text recognition which can be used to extract words from documents and images. It uses state-of-the-art optical character recognition (OCR) to detect embedded print and handwritten text.

But will it work for you? Will it be accurate enough to be of any use?

We tried scanning cheques with the idea of improving the efficiency of recording payments by cheque. We found that with the current technology the results were un-satisfactory. On the other hand, we found that scanning completed forms can be made to work quite well.

You won’t know unless you try.

2. Sales Insights.

Sales Insights harnesses Artificial Intelligence to empower sales teams to understand data in ways they could not previously. The aim is to augment every selling activity, customer interaction and business decision with intelligence. Insights include predictive lead scoring to monitor the state of your leads, relationship analytics to reveal the health of your opportunities and forecasting to predict your sales next month. Pretty powerful stuff! If it works – if the insights are accurate enough.

We recommend that Sales Insights be tried via a Proof of Concept. The results will depend on the quality of the information available. To get ‘accurate’ insights, you will need ‘enough’ data. How much is enough will vary with the regularity/stability of your sales process. The inherent ambiguity makes Sales Insights a good candidate for a ‘try and see’ approach.

3. Virtual Agent (Chatbot)

Dynamics 365 Virtual Agent for Customer Service combines AI-powered chatbot and insights solutions that enable customer service teams to easily identify and automate common support issues—all without writing a single line of code.

There is great potential to reduce costs by empowering your customer service team to build and update intelligent chatbots that use built-in natural language processing capabilities to engage conversationally with your customers. Productivity can be enhanced with AI-driven insights to help identify and automate emerging or time-consuming customer support issues.

The smart move is to try it – run a POC to check if it works in your situation.

Your customers may love it – or they may hate it. The insights available may indeed lead to efficiencies – but they might lead to customer frustration. A well designed POC will allow you to experiment with it – to ‘tweak the dials’ – to see whether it can deliver the benefits promised.


A Proof of Concept is a means of avoiding the two dangers associated with such new technologies:

  • Waiting too long and missing out on the potentially large gains to be made
  • Jumping in too early and wasting money on a project that fails to meet expectations