How Reps Use AI Sales Agent
We analyzed interactions from B2B sales reps using our Sales Companion agent. In this blog, we group these queries into six key categories: product discovery, competitive positioning, content generation, and deal support. If you’re building an AI sales strategy, use these insights to focus your agent's features, reduce non selling time, and speed up adoption.
Key Takeaways
Almost half (44.3%) of all queries are about products and services. This shows that AI agents must provide accurate, easy-to-find product information.
Even though AI can answer questions directly, reps often (15.9%) still look for specific documents like decks and battlecards. Finding the right files is just as important as answering the question.
About 12.8% of prompts are for content creation. Reps want agents to help draft messages and sales materials, turning the platform from a simple search tool into an active deal assistant.
Competitive queries make up 12% of the total, but they are crucial for winning deals.
About 12.5% of prompts involve administrative tasks like pricing and deal support. Using AI here cuts down on admin work and helps reduce errors in proposals.
Power users (2.6%) who refine their prompts and provide feedback are essential. They share the insights needed to drive adoption and improve the agent.
About the data
Revenue enablement programs must fit each organization’s sales process. To understand these findings, it’s helpful to know the context of the data collected.
B2B Enterprises in North America
Mature revenue operations with more than 50 sales reps
Multiple offerings in a highly competitive market
Sales motions that include new logo sales and upsells/cross-sells
Product (Offering) Information (44.3% of all prompts)
When selling multiple products in a competitive market, you need more than just general content. Buyers want to know how your offerings solve specific business problems. When selling to existing customers, reps must also answer questions about past purchases to build trust. Our data reflects this: nearly half of all seller prompts are requests for specific offering information. Using AI to find this data is much more efficient than traditional search methods.
For instance, in a traditional setup, a rep's path to finding information for “How does [the AI assistant] help [a specific business unit]?” looks as follows:
Filter the content for the business unit tag/label.
Add another filter for the AI tag/label.
Focus on assets that are externally sharable like product demos, data sheets, and more.
Scroll to the section they are looking for and synthesize the response.
And this sequence works only if the underlying content is properly tagged and labeled.
With AI, the rep is guaranteed not only a much quicker outcome but also a more accurate one. Furthermore, there is no burden on the rep to decide which tags to select and which labels to filter on. AI removes the dependency on the rep and equips her with the right answer.
At Enablix, we talk about helping reps get the job done. In this segment, this is evident from a class of prompts that is related to the offering but not directly about the offering. For example, “[product name] point of contact” is different from “how does [product name] work?”. Both are about the offering, but the context is different.
About Product/Offering (27.7%)
Representative prompts
How does [the AI assistant] help [a specific business unit]?
What is included in [the entry level suite]?
Does [our platform] track [a specific workflow] in service?
What is the difference between [product A] and [product A Pro]?
Does [third party vendor] integrate with [our platform]?
Does the new [module] work with older [hardware]?
Customer was told our [product] is not compatible with [a rival platform]
Will the integration override the capacity rules the [department] already set?
What fields push to [a competitor system]?
A line item list of everything included in [a suite] for [a region]
Product Domain or Company Information (16.6%)
Representative prompts
[product name] point of contact
latest headcount for [a region]
[internal program name] upcoming events
What is [our company]'s AI strategy?
Find the collateral (15.9% of all prompts)
In the pre-AI world, sales enablement platforms were architected to get the rep to the content asset. The goal was to help reps find the right pitch deck, one-pager, battle card, and so on. But, as we discuss in the previous blog, AI simplifies the process, allowing reps to reach their goal directly.
If a rep’s question is “do we support SSO?”, the answer is Yes or No. It is not a data sheet on security.
We see that shift in our data, as the majority of the prompts are geared toward getting the job done. However, the pursuit of content assets from a library has not completely evaporated. Reps still need to find collateral, and they use AI to get that job done. We believe the primary driver for reps continuing to need content assets is that they need to share something with their prospects and customers. Every representative prompt confirms this reality.
Representative prompts
Do we have a one-pager on [a product] for [a customer segment]?
I need a document on how to change [a feature setting]
Show me everything the marketing team has published recently
Any case study for [a vertical] using [our product]?
Rank the strongest customer stories, broken out by industry
Do we have a testimonial video for [a product]?
I cannot find the steps to export [a report] to a spreadsheet
Creating content (12.8% of all prompts)
Tools like ChatGPT and Claude have proved how easy it is to create content with just a prompt. Our data shows that reps expect this same simplicity from their sales companion. They want the agent to find information and help them create relevant content for prospects.
Not all content that a rep creates using AI is a document or a presentation. As you will see, the majority of the content created is in the form of messaging that is then included in an email, a text message, or even a proposal or deck that the rep is working on.
This can be a function of what is approved by the organization for the sales companion agent. If the sales companion agent can create decks and pricing documents, then you see an equally concentrated use of the agent to create sales materials.
Message refinement and follow up (10.1%)
Representative prompts
Rewrite this so it is safe to put in front of a customer.
Cut this back to a few bullets and the key numbers.
Give me the same thing as an HTML page I can send
Write an initial outreach email to a [buyer title] I met last week
Help me with a subject line for a warm client one day after [an event]
Draft a quick follow up email about [a product feature]
Write an email to customers who lapsed, I am trying to reactivate them
Reword this so it does not come across as a sales pitch
Create a battlecard for [a product area] between us and [competitor]
Marketing campaigns (2.7%)
Representative prompts
These ads are going on [a social platform], what are some caption ideas?
I need to create a customer facing flyer using our brand guidelines
What would you put on a landing page for prospects?
Competitive intelligence (12.0% of all prompts)
Objections and competitive handling questions are excellent signs that a buyer is qualified. In sales, the absolute worst-case scenario isn't pushback—it's apathy and silence.
Therefore, a decent concentration of questions is targeted in this area. A few patterns to note:
One might point out that only 12% of prompts focus on this category, whereas almost half of the questions concentrate on details about the offering and the company. We think that is an excellent signal because it is representative of an enterprise sales funnel. A B2B organization is bound to get a lot of product/offering-related prompts because there is a larger audience base interested in that information. Conversely, only qualified and interested buyers seek competitive intelligence and pose objections.
Though making up only 12%, these are the most critical prompts, and enablement teams cannot afford to fumble them.
They are for qualified buyers so the stakes are high.
They influence a deal more than other types of questions. In terms of deal attribution, they carry significantly more weight than many of the offering-related questions.
Representative prompts
How do I position against [the main competitor]?
How does [competitor] win against us?
Where does [competitor] leave their customers exposed on [a technical area]?
Which newer rivals should I know about, and how do I position against them?
Show what is included in [competitor]'s all-in-one bundle and what is missing
How is [the competitor] positioning AI for our industry?
Does [the main competitor] have proof points in AI?
Pipeline Support (12.5% of all prompts)
The goal of revenue enablement is to reduce non-selling time. Admin tasks consume a large portion of a rep's day, and AI can significantly reduce this burden. Automating these tasks helps close deals faster and minimizes proposal errors.
As agents become more powerful, reps increasingly use them to complete these admin tasks. These prompts, though representing a small percentage, are vital to the rep’s success. Automating these tasks using agents helps win deals faster with less effort and reduces errors in pricing and proposals that the enterprise sales industry notoriously suffers from.
These prompts also reflect the broad roadmap of what other capabilities reps are seeking and which other tasks they prefer the agent to perform. At Enablix, we are constantly getting feedback from users who identify such use cases and want them supported inside the sales companion itself.
What questions did reps ask?
1. Half of the questions are exactly what you would guess: “How do we help CMOs in reducing costs?” (product information from content) and “Where is this pitch deck I need?” (to discover the content).
2. The part that was hard to predict: 12% was downstream work, work where a rep wants the Sales Companion to create a deal-specific document (one-pager, pricing guide) inside itself. But there is a pattern hidden underneath: reps get to create deal-specific content once the other half works reliably. When the “How do we help CMOs in reducing costs?” question returns the right answer, then the Sales Companion earns the trust to build the one-pager that gets sent to the CMO for that specific deal.
Sales Companion is always only as good as how well your content is organized underneath; it gates everything above it.
Pricing, quotes, and proposals (3.6%)
Representative prompts
Putting together a proposal for a small single location customer, I need supporting material
Turn this proposal text into a single page branded document
Customer asked why [a free product] has to be activated if it costs nothing
Partner programs, industry, and regulation (3.6%)
Representative prompts
I need all partner enrollment links for [a product line]
Is [our product] eligible for [a partner] co-funding?
What are we doing about [a state regulation] taking effect in [a month]?
How does contracting work for [a partner specific product]?
Questions to check whether the partner is selling [our product] correctly
Internal process (1.6%)
Representative prompts
How do I request [an internal service]?
Who is the product marketer for [a product line]?
Documentation or best practices for setting up [a configuration]
Am I allowed to say "[a specific marketing phrase]"?
Which one-pagers are already approved and ready to send to customers?
Deal and account-specific coaching (3.7%)
Representative prompts
My customer has [product A] but not [product B], how do I start that conversation?
Meeting with a [multi location] group considering adding [a product]
A prospect on [a competitor] says they are satisfied, what do I do?
This customer lost their temper with me and accused me of overselling
How do I talk about [an AI product] with a [senior title] at a small customer?
I am meeting with the [department] lead, what language should I use?
I need the demo script for [our core product]
Walk me through building a plan for one of my accounts
Run a demo on [a product] with me and help me dial in my flow
Write an outline for a renewal review with an existing customer
Meeting a [senior buyer title], what three discovery questions should I ask?
Give me a thirty second pitch, key benefits and discovery questions
Help me counter a talk track when they raise [a renewal objection]
Write a quick cold call talk track for [a product] when speaking to [a role]
Working the agent itself (2.6% of all prompts)
Power users are those who go beyond asking basic questions; they actively refine prompts and look for ways to maximize the agent's value. These users are essential evangelists for driving adoption and improving the agent.
Identifying these users is important as these users are often users who use the agent most frequently and are already getting value out of it. These power users can serve as evangelists to increase overall adoption and provide valuable feedback.
Prompt building (1.9%)
Representative prompts
Ask me up to ten questions, one at a time, to help me build a stronger prompt
Create a reusable prompt from this session
Create a learning path for me to understand [product] versus [two rivals]
What does my user profile actually contain?
Do not execute this prompt yet, just draft it
Troubleshooting and answer quality pushback (0.7%)
Representative prompts
Why did you not recommend [a relevant product] at the beginning?
The clip you pointed me to is dead, is there a newer one with current figures?
Critique your own reasoning and tell me where it is weak
The formatting is broken when I print this