CRM software & AI: how deep will it go?

AI has been, and still is, a hot topic for anyone in Tech to write about. The shakeup for the public, namely the release of the early demo of ChatGPT on the 30th of November 2022, is now way past us. By now, many people vaguely understand what is or is not AI, and most importantly: that it is not solely Generative AI. Within this article I want to discuss what AI entails within Customer Relationship Management, as it’s sometimes hard to see what AI looks like in practice. I want to build it up gradually: discussing first what AI is in my opinion, what CRM software is, the combination of the two, what its advantages and problems are, the potential still behind it, the outlook of CRM AI and what its main drivers will be for the future.

AI in general

When it come to the definition of Artificial Intelligence, I especially like this blog . Written by Ellie Gabel and posted in 2023, she describes different complexity levels of AI, varying in memory and self-awareness. I’d recommend reading the blog if you want to get more into this. The part I want to focus is the part about the different branches of Artificial Intelligence:

1. Machine learning - A common application of this AI is to predict demand. It uses historical demand coupled with different variables like weather, time of day, etc. Based on that, it predicts future demand. It is trying to form connections within its trainings data and make logical predictions based on that. But why not do that ourselves? Simple: these programs can handle much more data than the average human ever could and is able to quantify relationships between variables. Very impressive overall and it’s a super important part of AI for industries like retail, which deal with large amounts of data.
2. Deep learning - Commonly associated with the term ‘neural network’. This AI is designed along the principle of trying to replicate neurons within the human brain. It is used a lot for image recognition, popular examples include applications within the medical field like identifying tumors in pictures. There are some super exciting opportunities for deep learning, which have the potential to improve our lives so much.
3. Natural Language Processing - Training algorithms to recognize, process, and respond to human speech. Whether this is verbal or written: these algorithms are seen everywhere. Examples include filtering of language, understanding large amounts of customer reviews on product pages, and chatbots. This field has always been highly relevant within CRM and will continue to play a big part.

The blog also describes robotics, fuzzy logic, and expert systems. These are super interesting and relevant parts within the field of AI, but are less relevant to this article, hence I will not describe them here. Feel free to visit the blog to read up more on them!

CRM in general

Customer Relationship Management is there to enhance a customer’s experience when interacting with the company. It consists of sales, service processes, forecasting, and analyzing trends and behaviors within the customer base. When we say CRM, we usually refer to the systems that sales agents use for customer relationship management. Within my job as the Solution Chief Revenue Officer for SAPs MEE region, I am leading all the CX dedicated resources of SAP that are customer facing. Meaning: I do not only sell CRM Software with my Team, but I am also a broad user of CRM. There are a multitude of companies developing and facilitating CRMs to customers: these include big names like SAP, Salesforce, Oracle, and Microsoft and a big bunch of start-ups and scale ups aiming to enter this part of the software industry as well. The companies that use their software originate within all kinds of industries and are of varying sizes. Obviously, you would not find this software in your average brick and mortar store, but the moment a company has a sales department, the opportunity to implement a CRM opens up. The need for a CRM especially exists when a sole seller manages many different accounts, and/or needs to tend to many different individual customer wishes, and/or has a wide portfolio of products to offer customers.

There is thus a proper use case for CRMs: from managing large amounts of customers, to tending to many individual wishes, and having a wide portfolio to offer. However, there are also drawbacks to using CRMs. These include:

1. Proper training & adoption - Employees need to be skillful in the use of the CRM and it needs to be widely adopted within the department to ensure proper functionality. After all, they are the people using, and hopefully benefiting from, the software. It wouldn’t work properly for the company if only a part of the departments adopts the software or knows how to use it.
2. Technical support - Software is not a one size fits all solution. Different business needs require different adoptions of the product, and companies need to stay on top of updating their software. Depending on company size, the technical support is either done in-house, by an external third-party, or by the CRM provider.
3. Supervision - The CRM gives employees a lot of power and a lot of responsibility. Often, this needs to be managed by sales managers to ensure that individuals comply with internal rules and performance guidelines, to ensure that the company maintains its reputation.

What I can clearly say from my user role is: most inefficiencies and issues we face in day-to-day CRM usage actually originates from integration into other systems. Be it back-office solutions like ERP, BW, and the management of non-standard (contractual) situations. With other words: the most beautiful CRM solution is worth nothing, if it is not integrated properly!

AI x CRM

Now, it’s clear that humans are an essential part within CRMs. They interact with the customers, input important data, have interactions, and are ultimately responsible for the accounts they tend to. CRMs are tools to interpret and manage customer bases more efficiently by them. CRMs mean nothing without the data being properly managed and decided upon by humans.

There are important use-cases for AI within CRMs. However, it would be unwise to imagine that a relationship with a customer can be properly tended to only by AI. In interacting with customers, so many subtle signals are given off and interacting with customers often requires cultural insights, next to plain logic and being able to respond to sudden opportunities. No AI will ever be completely able to replace humans. In the context of CRMs, we need to see AI as a tool, to make a sales employee’s life easier and their work more efficient.

We see AI used within CRM systems in several ways. Within SAPs Sales Cloud V2 system, sales employees can be reminded to check up on a customer if they haven’t interacted much lately, and they can see calculated probabilities of success for different accounts for example! Features like these make the lives of sales agents a lot easier by allowing them to prioritize specific accounts and staying on top of them.

Next to that, the lives of sales managers is also made easier. The AI in the SCV2 system allows them to check performance, not only on national or worldwide level, but also per region, location, and per individual! Even cooler is the functionality within the system that can run predictions on pipeline growth and expected deals. Together with probability of deal success, the AI can calculate how much should be going into the pipeline, allowing them to work more closely with other departments like production and logistics, to dedicate enough resources to the appropriate locations or functional areas.

The future of AI x CRM

How AI in general will develop is a big question mark, the same is true for AI specifically within CRMs. I believe it would be unwise to just take an expert’s opinion (including mine in case you consider me an expert ??) and base your actions on that. No one will ever truly know beforehand how fast or slow technology will develop within either the short- or long term. If someone does state a very clear conclusion on it however, run!

So… was the title of this article clickbait then? In a way, yes. I would not dare to give a single answer to the question. There are varying depths of AI and its usage application within CRMs can pivot into a lot of paths. We can, however, think about the drivers behind AI within CRMs. What is the AI dependent on and what can be possible obstacles for its future development?

There are several drivers within AI and CRMs. Historically, an important driver was the computer power to host the application of AI. This remains an important factor within the development, as Moore’s Law is still a hot topic. With advancements like quantum computing, we will have to see how much computer power actually opens up. I think that with recent geopolitical issues between Taiwan/China/Europe/USA a serious issue opens up: computer chips are grossly overdependent on resources that are finite and that are held by a limited number of countries or organizations. The future will tell whether we will have enough resources available to produce the needed computer chips to dedicate to even better AI systems.

Next to this is the relevant legislation. Already there are several organizations and countries pushing for different kinds of legislation around AI. Already, the EU has introduced the AI Act. This is not necessarily a bad development for the maturing of AI, I believe it merely requires us to stay within certain boundaries and rules of the game. From a business perspective it may seem limiting, but the EU legislation has a clear goal in mind: to protect their citizens. Let’s be realistic; AI can be used for both good and bad. Limiting negative impacts of the use of AI will be incredibly important and the legislation will also determine how much impact we can make with the positive sides of AI. Only the future will know what the results are from the development of AI, the impact of legislation, and the scandals that may happen when there are companies that break legislation.

It’s not just legislation, however. There is also the social factor: how accepting will we be of using AI systems and how accessible will the appropriate data be? As for now, a lot of people are quite comfortable with sharing a lot of data regarding their internet browsing and personal demographics. This may change in the future as awareness rises regarding potential risks and the increasing focus on personal privacy. The introduction of ‘the right to be forgotten’ by the EU has certainly set the tone: people do have the right to remove their online presence. How that will develop in the future is unknown, that will depend on customer wishes and priorities.

So… the final answer is ‘it depends’. That is both the easiest and most honest answer. What do I expect personally? I see the future of AI & CRM as exciting. Various systems already existing will see great performance improvements over time, as we further learn from past business cases and adopt these into current systems. New systems will be introduced; systems focused around further making sales agent’s and manager’s lives easier. Focused on automatic agendas for customers, likelihoods for success, better suggestions for customer interactions, etc. I would love to see fully automatic systems for the sales agent’s job, with them still having a final say in things and being the face of the company. After all, the AI is just like the CRM; a support tool to make your life easier.

Conclusions

To conclude, AI and CRM systems are already heavily intertwined. Various forms of AI like machine learning and Natural Language Processing are already present and form an important part of our functionalities offered to customers. The future is exciting, but we will have to see how different areas will impact the development of AI, and AI within CRM specifically. Factors like computer power, legislation, and social acceptance will play an increasingly important role within the development of AI within CRM. The future is exciting, but a lot is still to be seen, that is for sure. With SAP being at the frontline of making useful business cases out of AI, a lot is still possible and under construction!