bots ? CRM

In 2019, the percentage of companies using CRM systems rose to 74%, while the percentage of websites using chatbots is, based on my talks with customers, still below 10%. So, barring some kind of weird anti-trend, most companies that use chatbots probably use CRM systems, but not many companies that use CRM systems also use chatbots.

 However, I am confident that the two will end up being used together in the future - we aren’t at this level of perfect synergy quite yet, but there are already several reasons why bots love CRM (and vice versa). At Solvemate, we’ve successfully partnered with FreshWorks, Zendesk, RingCentral, and SalesForce (with more CRMs always being added!) to bring these early synergies to our joint customers. Although each product has an invaluable offering on its own, the synergies they create by working together are so valuable that having bots and CRM systems working seamlessly together across vendors is just a matter of time.

What does integration look like today?

 Why do I think a perfect synergy exists? Because many parts of the CRM system provide a natural counterpoint to chatbot functionality. A couple of the most important ones involve the flow of data between the two systems, and already exist today:

Handovers

Chatbots automate answers to common questions. But what happens when a question requires a human answer? Well, a seamless handover from chatbot to human requires both chatbot and CRM system to work together. With the two working in tandem, both types of questions - the easy-to-automate ones and the ones that require a human touch - are handled. A ticket handover (creating a case that is answered asynchronously) works pretty well, but when it comes to live chat handovers, some CRM systems still do not offer the technical foundation for looping in an agent in real-time. 

 Personalisation

Chatbots display information to their users. But how can a chatbot tell its users apart to show, for example, personalised information regarding an individual’s payment plan? With access to the customer-level data contained in a CRM system, a chatbot such as Solvemate can already personalise every interaction and serve up the most relevant information. But the data fetching options from CRMs are limited for a couple reasons. Firstly, not all CRM systems support easy access to the data of their customers. Secondly, not all customers of CRMs have the necessary data for a chatbot available (which would ideally be a summary of the customer data which could be fetched from the database, e.g. past orders, shipping status, newsletter subscription, etc.). 

Automating routine requests

Chatbots can easily propose informative solutions to repetitive questions, but what about questions that require repetitive actions to be taken? For example, imagine customers uploading a handwritten form, which then needs to be entered into another system. If an agent has to process this kind of task thousands of times, why wouldn’t they want to automate it? Enter automation. With integration to an RPA (robotic process automation) queue - which, in turn, would ideally happen by integrating with a CRM system - a chatbot can not only identify what process needs to happen - it can actually order the process to take place. The CRM integration would even allow you to see which tasks are being repeated the most often (and so are good candidates for automation). For more details, you can listen to my (German) podcast “AI Unplugged”, Episode 14 with Walter Obermeier from UI Path.  

 What will the future look like?

 Even with the “basics” above, the synergy between bots and CRM systems is clear. But of course, we are always looking to the future, and thinking about how we can truly bring integrations to the next level. 

Here is my overview of what I see in the future of chatbots within the next 5 years

Live chat handovers

Although live chat handovers are of course already possible, certain bots cannot integrate into certain CRM widgets (yet). In the perfect integration, all bot or CRM interactions will be contained within the CRM chat widget, which will allow handovers to occur without any break in user flow. Although each chat vendor currently offers their own chat widget, in the long-term, bots should live within the chat-widget of the CRM system.

Routing

On top of handovers being seamless, the routing logic for producing queues would be integrated into the native queuing logic of the CRM. Today, we can already route tickets in the right queue, but the APIs for that are rudimentary. In the future, chatbot vendors will be able to directly “plug in” to the native routing logic of the CRM and use all their features to make sure that the right service agent gets the right ticket. 

Content Management

Many companies, before starting with a chatbot, start with a Help Center or “FAQs”. The ideal integration would be able to seamlessly import existing questions into a “chatbot friendly” format (i.e. very short and to-the-point). Whether this “short version” of the Help Center article is maintained manually or with AI trained to summarise work, there will be no need to maintain information in two separate places. 

Integrated dashboard

The analytics dashboard would be contained inside the web app of the CRM. Instead of having to navigate to different tools or screens, users of the CRM system would be able to access the bots from within their familiar CRM system. The chatbot analytics dashboard could be integrated with a simple API pull that then generates charts inside the CRM’s dashboard. 

Algorithm data

The clustering algorithm, which suggests new training tasks for the bot trainer, will have access to all conversations and even more customer data (all pulled from the CRM system). For example, the algorithm could look at all past tickets to suggest which solutions are most important. Incoming tickets could even be flagged, either by the algorithm or by customer service agents, if they cover content that is missing from the bot. This flag could then alert the bot trainer to add the content to the bot - or perhaps the bot could even add the missing information automatically.

---

In short, the ideal bot product must be connected with many parts of the CRM system. Today, the two are already connected, and because of this we are able to take advantage of a host of synergies. But there’s always a next step to take - and, in this case, the future of bots and CRM integrations looks very bright indeed. Why wouldn’t it? After all, bots ? CRM!