Artificial intelligence is still in its very early stages, but its potential to improve the entire CRM process is becoming clear. Already AI tools are being offered by a few vendors and a scattering of companies are either experimenting with them or starting to apply them.
It’s important to understand that at this stage AI is still a matter of definition buzzword if you will that is only beginning to separate itself from other techniques. The best way to describe today’s AI is “promising”. The potential is obvious, but translating it into results is only just beginning.
General AI in the movie sense is still long ways off. Today’s AI aims to solve specific problems in narrow situations. Within that context, however, AI has the potential to be extremely useful.
Today’s AI in CRM aims to act as an assistant to human sales and marketing staffs. It can advise and point out opportunities and problems, but the judgment and action depend on the human.
Lead scoring, for example, is a fertile field for AI in CRM. The AI system can analyze a wealth of information about prospects and rank them according to their likelihood to purchase. The list can be presented to the sales staff with the most promising prospects first and the sales force can take it from there.
This kind of predictive lead scoring relies on gathering information on the prospect from as many different sources as possible and analyzing it to produce an estimate of likelihood to purchase. Generally, the more information the system is fed and the more sophisticated the algorithms the more accurate the results.
Such systems require considerable training in order to perform well. The training can either be done by experts on AI with some domain knowledge or by sales and marketing staff who understand the field thoroughly.
What this means is that today’s AI seldom works right out of the box. It makes setup and training and the more information it is fed the more accurate it is likely to be.
Making this work effectively requires defining the problem as narrowly as possible. Like computers in general, AI likes certainty and the more certainty it has, the better it performs.
Other areas of AI in CRM are further behind but they are coming along. One attractive is speech recognition. However, this requires solving a cluster of intractable problems.
For example, a truly effective speech AI application needs to understand the customer’s tone of voice and how it impacts what the customer is saying. We’re making progress in this area but in general-purpose speech applications aren’t ready for prime time.
It’s still very early innings for AI in CRM. It is successful in some areas and promising in many others. As time goes on we will undoubtedly see more and more AI features built into CRM systems to help them perform better.