The following is about AI in CRM and Customer Success. I had a call with a prospect earlier this week. It started by sharing our cool demo that addresses two major pain points of any CRM and Customer Success solution: (1) Getting data in (2) Finding it easily. And then…I suggested to take it one step further and share our AI for CRM and Customer Success stuff. They immediately replied…”No, No, No! Let’s start with the basics.”

As a wanna-be sales person, I switched to showing what I just called AI for CRM and Customer Success stuff. I think they expected a deep data science discussion about Machine Learning and Natural Language Processing algorithms. But I actually showed them examples of simple and clear coaching instructions, guiding each of their team members to do a better job. The examples included:

 

At that point they said. Ahh. That’s AI? We would definitely be happy to try it out…

What happened?

AI has been hyped as the potential successor of everything. Replacing our jobs and taking over the world.

However most of the current practical implementations of AI are below usable. For example AI chat bots. I recently had a “heated” discussion with a customer service chat bot of one of the largest software companies. I asked it for help and it kept giving me completely wrong answers. Not even close. So I had to ask for human assistance three-four times, until it finally agreed…

Many of the chatbots lack even basic intelligence. They are more like dynamic menu systems, not different than automated phone call menus “For the right answer press 1, else press 2”. I understand the need and the great potential of chat bots. But at this stage of the technology the experience needs to be humble and possibly less personal. Assuming that in many cases the bot will not succeed to provide the right answer.

Recommendation engines are successful

35 percent of what consumers purchase on Amazon and 75 percent of what they watch on Netflix come from product recommendations. See more.

People treat recommendation engines in a much more tolerant way. They need to be right in one or two out of several recommendations that they list. And since the recommendation is optional, even if no recommendation was relevant, it will still not be annoying. This is versus a chat bot that needs to be accurate in at least 80% of the time. Much harder problem to solve.

How is AI doing in the CRM and Customer Success world?

All of the leading CRM and Customer Success providers have announced adding AI to their offering. For some it is still marketing and others have released solutions to market. There are three main categories:

1.    Chat bots – mostly deflecting customers’ cases and prospects’ requests. Successful solutions will find the right balance between the role of the bot and the role of the supporting human. Let the bot do the simple stuff and switch to human when it’s less likely to handle it. The best implementations integrate humans fluidly into the conversation. Consumers should not even notice that a human took over. This obviously goes beyond software and requires the availability of people.

2.    Intelligent insights – Provide insights about people and companies that you are engaged with or trying to. For example: (1) Recent post by one of your contacts (2) M&A announcement by one of your prospects/customers (3) Identifying competitors mentioned in email exchange. Some of these make total sense in demos, but in real life it may take different shape. Many of the insights are practically “shot in the dark”. They don’t have any clear goal and thus could easily become noise. Successful solutions continuously monitor users’ behavior and dynamically adopt their insights to address each user’s specific needs.

3.    Scoring and coaching – Scoring classifies the likelihood of leads to convert, opportunities to win and accounts to retain and expand. The score is important to identify areas that need to be handled. But it is even more important to provide prescriptive instructions to improve the expected outcome e.g. win a deal. Also, in this case successful solutions must be dynamically adoptable based on users’ actions and results e.g. win/loss.

Successful AI solutions strive for data

In the world of Machine Learning and AI for CRM and Customer Success, it’s not only about smart software, data plays a key role. The more an AI solution is used the better it becomes. That’s probably the main thing that differentiates real AI solutions from the rest. Check out Gordon Ritter and Jake Saper of Emergence Capital Coaching Networks. It was a great inspiration for us at Komiko.

Komiko starts its AI for CRM and Customer Success by training the coaching model using two main sources of data signals. The actual results of your past activities i.e. opportunities win/lost along with engagement behavior. Other signals like usage data are added later. The coaching model recommends playbooks that experts review and adjust. It’ll then continuously monitor users’ behavior, engagement and success rate to provide clear coaching instructions and to improve the model.

Happy to have a quick chat about AI for CRM and Customer Success and share more. Click here

Categories: AI-CRM

Leave a Reply

Your email address will not be published. Required fields are marked *