If you’ve been watching the sales and marketing technology space over the past few years, you probably have seen the rise of the idea of “next best actions,” or guided selling. 

The promise of A.I. is here, and instead of reps having to figure out what to do, technology can monitor a plethora of data points and actually tell them what to do next…

Sounds amazing, but does it actually work?

In this article, we compare the pros and cons of the two types of next best actions solutions, as well as how you can utilize them in harmony to put your team in the best position to succeed.

Intelligent Next Best Actions 

The first type of next best action solution uses intelligence to automatically surface prescriptive next best steps to your team. In theory, this sounds like the dream sales assistant everyone’s always wanted. 

Instead of asking reps to think about what to do, software monitors all the different data points and provides contextual recommendations on the right action to take. 

Example next based action suggestions could include:

  • Prompts to follow up or engage opportunities at risk
  • Suggested forecast or CRM updates
  • Accounts with intent that should be prospected
  • Common connections or discussion topics to introduce in conversation
  • Going wider at accounts 

For most automated solutions, action recommendations are out of the box and do not require any configuration on the user side.

To arrive at these suggested actions, intelligent next best action solutions ingest data points from many different sources in order to power their suggestions. 

These can include any combination of things such as:

  • Sentiment, either verbal or written (Gong, Chorus)
  • Public data points and intent signals (Bombora, Crunchbase)
  • Historical CRM data (average sales cycle and stage length, touch points required to close a deal)
  • Email and calendar activity (last meeting, unanswered email response, next calendar meeting)

Implicitly the presence or lack of these data points can provide a signal to reps to drive the right behaviors and uncover things they might have missed. The benefits are that your team can work smarter with more available insights to make the best decisions. 

The Challenge and Downsides

We must caveat this potential with some things you will likely run into with any solution oriented towards providing automated next best actions.


Like any intelligence-driven solution, the efficacy of intelligence-driven next best actions solutions is only as good as the accuracy of the underlying data. And as soon as accuracy drops below 100% trust in the recommendations starts to erode which eventually results in a lack of engagement from your team. 

When talking to buyers and from personal experience, we have concluded that intelligence suggestions are rarely right even 80% of the time. Unfortunately, there are often false positives and circumstantial information that a computer just can’t properly decipher.

Maybe a prospect doesn’t get back to you, but happens to be your old college roommate and someone you deeply trust. Is that deal really at risk? 

Or maybe you've only engaged a small percentage of people at an account relative to other deals, but you have a buyer who is an ex customer and committed to making it happen.

There’s just way too much nuance for a computer to fully understand the full picture and accurately tell you what to do. 

As a result, there seems to be more value in surfacing the insights you might not be aware of, then pushing the decision on next best actions onto people. In most scenarios, your seller likely knows what they need, but they might not be aware of things happening between the lines and systems when managing so many conversations.

So when evaluating these solutions, it’s probably best to focus on what data can you put at your sellers fingertips vs. expecting a machine to coach someone effectively. 

Unique to Your Process

Another challenge you are likely to face is the ability to customize recommendations to your unique process. CRM and business processes are quite different with every company. The language certain industries and companies use can vary widely. 

As a result, any solution that tries to provide uniform prescriptions can be quite precarious, resulting in lots of false positive recommendations which your team will ignore. 

Data Collection and Time to Value

As mentioned earlier, part of the special sauce of next best action solutions is their ability to ingest historical data and then use them to inform you about the future. Depending on the company, this can take a lot of time to accomplish. And what happens in scenarios where you changed your sales stages or prices 4 months ago and the historical data isn’t relevant to today’s deals?

There’s lots of nuance and getting to an actual value event where you begin to get accurate information to inform future actions can take months. If you are evaluating any solution in the space, you should get crystal clear on how long the data processing and ingestion will take in order to get to a defined value event. 

Point Solutions and Limited Scope

From the time someone comes to your website until they become a customer and ultimately renew, there could be hundreds of little actions that your team must take to ensure success. Most automated next best action solutions are very narrowly focused on one part of the funnel or client engagement. This could be surfacing new opportunities or mitigating deal risk later in the funnel. 

It’s very unlikely you’re going to see any solution that is really good at analysis + actions across the entire customer buying journey and lifecycle,  though I’m sure many software platforms claim to be. So it’s important to get crystal clear on the core competency of any automated next best actions solution, and where their customers see the most value. 

Composable Next Best Actions 

The other next best action type of software is composable, meaning you can define the next best actions for any scenario based on your unique business processes. 

In many ways, these solutions are typically less focused on the intelligence portion of identifying scenarios where action needs to be taken and more focused on actual execution in known scenarios. If you have processes in place that you know work and are simply trying to reduce friction and improve efficiency, composable next best action solutions can be a superpower. 

Other benefits include fast time to value, flexibility to drive processes across any part of the funnel, and the ability to adjust and tweak actions as your process changes or you get more insight into what’s working. 

Two examples of composable next best action solutions are Salesforce Einstein Next Best Action and Troops.

Salesforce Einstein

The Salesforce Einstein platform allows you to present context specific recommendations natively within the Salesforce platform. 


In order to do this, within the Einstein platform your admin must:

  1. Define your recommendations
  2. Build a flow in flow builder 
  3. Use strategy builder to funnel recommendation to users at the right time 
  4. Embed the recommendations on any page 

This can work great if your team is living in Salesforce most of the time. Salesforce Einstein next best actions are also very flexible allowing you to integrate all types of data from the platform. The challenges with Einstein Next Best Action is that it is quite technical to set up for even for a single recommendation, and to be effective, your team must notice it within Salesforce rather being proactively pushed information. 


If your team spends a lot of time in Slack, Microsoft Teams or likes working out of spreadsheet-like interfaces to manage information, Troops might be a good fit for your team. 

With Troops you can define any go to market process where you want someone to take an action, notify them in Slack in real time of the action they need to take, and provide buttons in the message to complete the actions directly from within Slack.


Though adoption and regular usage of Slack or Teams is key here, the fact that these signals are proactively pushed and your teams can seamlessly engage with them across devices make them highly effective. Moreover, because they are delivered through a collaborative interface, teams can work together to complete these tasks and there is often a greater level of accountability by virtue of multiple people engaging.

Interested in bringing next best actions to your team in Slack, MS Teams or spreadsheets with Troops Grid? Get started for free today here

A Blended Approach

As you can see, when it comes to technology helping aid, guide, and drive the right behaviors there are pros and cons to both automated and composable next best action solutions.

Automated solutions stand out in their ability to ingest and process disparate data sources that otherwise might be difficult to see. However, when it comes to driving highly custom actions that are accurate 100% of the time and specific to your unique business process, there is still much left to be desired based on the maturity of A.I. today.

Composable solutions shine when you have defined business processes that you are looking to accelerate and improve greater efficiency around. The fact that they can be deployed immediately and adapt with your  business is also highly compelling.

The reality is that most businesses today can benefit from both and these solutions can often play well stitching them together. The key is to really define the job to be done of each solution so that you can ultimately make sure that each system is oriented to its strengths and the needs of your organization. 

There is still so much innovation happening in this space and we are excited to see where things go as technology advances!


Interested in bringing next best actions to your team in Slack, MS Teams or spreadsheets with Troops Grid? Get started for free today here

Scott Britton

Written by Scott Britton

Subscribe to our blog