Coaching is an integral part of any sales process. It’s a way to ensure that sales teams create a great customer experience by providing (or finding out) all necessary information in a way that resonates with the customer while progressing the Opportunity through the various stages. With this in mind, what if you could set your team up for success with an interactive way to practice pitches and provide relevant feedback at every stage of the deal, all in line with company specifics?
In this post, we’ll explore how to get started with Agentforce for Sales, from how it optimizes the seller experience, to how it can save your sales managers time while increasing their team’s efficiency.
Agentforce for Sales
Agentforce is front and center in the ecosystem these days, with Salesforce empowering customers to build their own tailored experiences through the use of out-of-the-box features and generative AI capabilities. While Agentforce Service Agent was the first of its kind to become generally available, the sales offering is also generally available as of today!
Agentforce for Sales has been created specifically to supercharge sales processes, from the beginning of the funnel through to Lead nurturing, or throughout Opportunity progression, be it a brand new deal or renewal. While in this post we will delve into the coaching capabilities, also included in the same offering is the ability to nurture leads around the clock, which you can learn more about here.
Agentforce comes to the rescue within your team’s sales process by coaching sellers to advance a specific deal in no time – through personalized yet objective feedback, at every stage of their Opportunities. This way, everyone on the sales team can benefit from their very own dedicated coach, whenever it’s needed, to get ready for any upcoming customer or prospect conversations – all directly within their flow of work.
The generative AI features come into play to support all of this in a secure and scalable manner, as the agent’s responses can be grounded in Sales Cloud, or any external data you need it to consider, while the Einstein Trust Layer ensures your data remains yours every step of the way. This really does sound like a match made in heaven, doesn’t it?
Organizations that purchase the Agentforce SKU can enjoy this new functionality with their usage aligning to the new consumption-based pricing model, while companies choosing to make use of Salesforce’s Foundations offering are entitled to one thousand free conversations for reps to use with their dedicated coach.

Turn on Supporting Features and Finish Configuration
For Salesforce Admins, getting started with Agentforce for Sales will be a breeze. Once available, a dedicated Setup page will appear, as well as a Feature Set in the new Sales Cloud setup experience named Sales Cloud Go, which is where the journey starts.
There are four easy steps to go through, starting with enabling supporting features: Einstein Generative AI and Copilot, as well as Data Cloud. In the context of coaching, Data Cloud is paramount to store files and to ground the LLM responses with information from unstructured data, through the use of retrieval augmented generation (RAG).
You will also have the option to navigate to the Opportunity record page and add the new component – you can also add this later on depending on your preference. I will cover this part when it’s all ready for the agent to be exposed to users.

Following this initial step, steps two and three are all about the new Agent user, and of course assigning your internal users permission to access or customize it.
Note that the Agentforce Sales Coach permission set is solely meant for the dedicated Sales Coach Agent user you will be creating, while you can choose between the Use Sales Coach Agent permission set or the Manage Sales Coach Agent one for your team. Since these are in the end permission sets, depending on your own setup, you could add them to existing Permission Set Groups as well.

Onboard Your Team’s Personalized Coach
The exciting part of the configuration starts at the last step – at this point, you will be prompted to navigate to Agent Builder and complete the one-time guided setup of your coach. Salesforce aims to make this setup experience as easy as possible to allow customers to try out the coach in just a few minutes. In Agent Builder, Topics reflect the jobs to be done, and each yields one or more associated Actions.
At the topic level, there are also clearly defined instructions, so that the Agent knows exactly when it is appropriate to use each of them depending on the context. Finally, Actions represent the outputs of the agent or what the agent comes back with. In this case, its job is to provide feedback to the sales rep – hence, four out-of-the-box Prompt Templates are readily available for each standard Opportunity Stage.

As each business is different, and as, ultimately, Salesforce is a highly customizable suite of products, it should come as no surprise that you can go ahead and adapt this Agent to suit your needs and implementation. This includes but is not limited to the Instructions, Actions, and of course, the language and style. Remember that Actions can use Prompt Templates, but also flows and Apex in the event that you would like to have even more control over the output of the coach.
Finetune and Test the Prompt Templates
Readily available Prompt Templates power the out-of-the-box Agent action, but perhaps your organization has bespoke data on the Account and Opportunity objects, as well as their own Products and other relevant information to be considered.
As you may have already guessed, Prompt Templates can be adapted to show what your team needs to see, as well as any data that should be considered that’s specific to your Salesforce implementation.
The Prompt Template Workspace is where you can tweak the instructions, the format of the feedback using simple syntax markdown, for example, and of course, ground the Prompt Template with any relevant information it should consider while providing feedback.
Added resources can be Opportunity fields, Related Lists, flows or Apex, and of course, those previously created Retrievers which allow you to tap into the unstructured company-specific data that resides in Data Cloud.

Once you’re happy with the instructions, make sure you thoroughly test the output by introducing a transcript as well as an Opportunity record. In addition, if you have already invested time and effort in using Model Builder, you can absolutely switch to another custom LLM and re-test to analyze the quality of the results before deciding which one is most suitable for your use case.

Practice Makes the Perfect Pitch
Agentforce for Sales includes an AI-powered coach available to each user in their flow of work, and respectively on the Salesforce Opportunity record. Once you’re happy with the configuration and test results above, it’s time to add the new out-of-the-box Lightning Web Component on the Opportunity record page – if you haven’t already – so that your users can start harnessing the power of Agentforce!

Once available, the user can access the coaching experience by clicking a button and, based on the stage the Opportunity is in, choosing the appropriate action. Keep in mind that users will still be able to practice their pitch even if they choose an action meant for another stage. The feedback will still be specific to the stage the Opportunity was expected to be in based on the instructions determined in the Prompt Template (and at Topic-level) in Agent Builder.
However, the Opportunity in one of the stages covered by the actions will use the intended action for that particular stage, and guidelines will be provided in line with what the user is expected to cover from a best-practice perspective.
The entire experience is browser-based, with no need for external tools or exiting the Opportunity page at all – access to the microphone and camera will be requested, but only the microphone is mandatory. For example, the option to practice the pitch is available out-of-the-box at the discovery stage, and users will have five minutes to talk out loud while the Agent is listening to the pitch.

Following the pitch, Sales Coach will process what was said and – in just a few seconds – provide personalized feedback in the format established in the Prompt Template, including the information you decided to ground the response in. If no changes have been made to the Prompt Template just yet, the resulting feedback will be similar to the example below, containing a deal summary based on Opportunity information, the actual feedback, potential areas to focus on, and of course, next steps.

Another great way to practice future customer conversations with the personalized coach is role-playing. Unlike the pitch practice above, this time the Agent will respond to you with both audio and text, creating a realistic conversation in response to the instructions it receives. Role play is a great option for the negotiation stage, as overcoming objections and discussing discounts could very well be what the sales rep and the customer discuss prior to an offer being made.
Once again, the maximum duration is five minutes, and the experience, which remains in the browser, is exposed in a modal on top of the Opportunity record.

The actual interaction will work exactly as you would imagine two colleagues role-playing as part of a workshop. The user should click the Speak button, then click again when they are ready for the coach to respond as if it were a customer. Since the role play is recommended during the Negotiation stage, the sales rep will have to say no or provide an alternative to the customer asking for a discount – as in a real conversation.

Ending the role play will result in the coach processing the transcript and preparing meaningful feedback for the sales rep. While fairly familiar in structure compared with the output from the pitch practicing option, out-of-the-box feedback will include additional points, such as the overall impression of the call and the key strengths identified during the conversation.

All Agent conversations will be listed on the Activity Timeline on the Opportunity as a Task, offering sales managers a one-stop shop when checking all of the interactions a sales rep had relating to a specific Opportunity.
Coaching can be linked directly to revenue outcomes, for example, in a win/loss analysis comparing Opportunities with and without coaching. As the Task is directly linked to the Opportunity, it’s simply a matter of creating a Salesforce report to uncover metrics related to coach usage across deals.

More to Come Soon
The team at Salesforce is already considering what valuable enhancements to make to Agentforce for Sales. We know from the product lead that a dashboard for sales managers/sales operations – on the usage of the agent and impact on deal performance – as well as expanding the use beyond deal coaching to other areas in the CRM (the ability to use the coaching mechanism while working on a Lead, Account, etc.) are top the list. So stay tuned!
Final Thoughts
With real-time, personalized coaching for each sales rep as part of the familiar Opportunity progression in Salesforce, Agentforce for Sales could become the ultimate enablement tool, both for new hires and more seasoned members of the team. By practicing their pitches and engaging in realistic role-plays, sales reps can be more confident going into real customer conversations, with just-in-time feedback based on company best practices, sales processes, and of course,all the data available in the CRM.
The post Give Every Rep a Dedicated Coach With Agentforce for Sales appeared first on Salesforce Ben.