AI Proposals: How to Create Ones That Follow Your Sales Playbook

Your sales team is probably already using AI to draft email responses, research prospects, summarize discovery calls, or prepare for their next meeting.

And if it's helping them get through all of that work faster, why wouldn't they use it to create sales proposals too?

It seems like an obvious next step, but when we talk to sales leaders, we find that many are still skeptical about handing over something as important as a sales proposal to AI. The thought of losing control over what reps are promising, which pricing they're quoting, or what commercial terms they're sending out to buyers makes them understandably nervous.

And rightly so. Once a proposal lands in a prospect's inbox, there's only so much you can do to take it back. These are real concerns for growing sales teams, and ones that can't simply be brushed aside in the rush to adopt AI.

So, how do you take advantage of AI proposal generation without losing control of what goes out to buyers?

Here's where we've landed at Qwilr: let AI read the deal, and let your rules decide what the buyer sees. AI is great at pulling a buyer's goals out of a 45-minute call. It shouldn't be the one deciding your discount.

If you're looking to bring AI into your proposal process (or improve how your team is already using it), this guide will walk you through how to create AI-personalized proposals that follow your sales playbook.

But first...

Should you use AI to create sales proposals?

The answer is, it depends on how much control you keep over what reaches the buyer.

AI proposal generation has come a long way, and there's no shortage of ways to get started. You could use Gemini to draft proposals in Google Docs, or use ChatGPT to personalize content around a buyer's needs, or explore dedicated proposal software with AI built into the workflow.

And there are good reasons to consider bringing AI into your sales process.

According to Gartner's April 2026 forecast, sales organizations using AI-driven enablement are predicted to achieve 40% faster sales-stage velocity by 2029 compared with those relying on traditional approaches.

That's huge, but before you give 30 or 50 reps the freedom to generate proposals using AI, there are a few questions worth asking.

• Does AI have access to your latest messaging and approved content?
• Can it apply the right pricing and commercial terms?
• And how much control do you have over what ultimately reaches the buyer?

The good news is that these are challenges you can address by bringing together AI, reliable deal data, and predefined sales rules.

We've spent a fair bit of time exploring this at Qwilr, and if you're still on the fence, the next few steps might just change your mind.

5 steps to make AI proposal personalization work for your sales team

The following five steps will help you put the right foundations in place and make AI proposal personalization work across your sales team.

Step 1: Decide what AI should personalize and what needs to stay controlled

Before bringing AI into your proposal process, it's worth starting with a few questions:

• Which parts of your proposal should change based on the buyer?
• Which parts need to follow your company's approved sales playbook?
• Where can AI use information from discovery calls and CRM records to tailor the proposal?
• Which decisions can be automated using predefined rules, and which still require a rep's judgment or approval?

These questions help you decide how much of the proposal creation process you can confidently automate, where personalization can genuinely add value, and what your team needs to have in place before rolling AI out more widely.

A useful starting point is to map out your proposal components and decide how AI should contribute to each one.

As you can see, AI doesn't need to generate every part of your proposal. Some elements can be populated from existing systems, others can change according to predefined conditions, and some will benefit from AI extracting information that would otherwise take your reps time to find.

You can begin with a few repeatable proposal components and gradually expand the process as you become more comfortable with what AI can handle.

Step 2: Give AI the right buyer and deal information

Think about how much information your reps gather before they even get to the proposal stage. Between CRM records, discovery calls, meeting notes, and follow-up conversations, there's usually plenty to work with. The challenge is making sure the details that matter find their way into the proposal.

Start by looking at where that information currently lives and what AI would need access to.

• CRM records: Company information, deal size, buyer requirements, and other details captured throughout the sales cycle.
• Discovery calls and meeting notes: The buyer's challenges, priorities, objections, and the outcomes they're hoping to achieve.
• Approved sales content: Product information, customer stories, and supporting materials that can inform the final proposal.

Of course, having access to all this information is only useful if you know what to do with it.

Say a prospect mentions during discovery that they need a new solution implemented before the next quarter. That detail might be buried in a 45-minute call recording, but it's exactly the kind of context that should influence which implementation information appears in their proposal.

This is something we've been working on with Qwilr's Smart Proposal Engine, which brings CRM information and context from sales conversations into proposal creation. Reps select the deal from your CRM, whether that's Salesforce, HubSpot, Dynamics, Pipedrive or Zoho CRM, and add call transcripts or meeting notes, including Gong calls.

Here's how it works.

Before connecting AI to your own data sources, though, it's worth doing a little housekeeping. Are your CRM fields consistently filled out? Are discovery notes capturing the details that influence the buyer's decision? And is your approved content current?

After all, the output can only be as useful as the information you're giving it.

Step 3: Turn your sales playbook into rules AI can follow

Your sales playbook probably has plenty of guidance on how reps should position your product, recommend solutions, quote pricing, and handle commercial terms. But how much of that guidance is built into the process of creating a proposal?

Let's say your business offers three service packages, with different pricing and implementation requirements depending on the customer's size and location. That's already quite a few combinations for your reps to keep track of, especially when they're working on multiple deals at once.

The next step is to turn those recurring decisions into rules your proposal workflow can apply.

Here's what that might look like:

Of course, there will always be exceptions.

A buyer might request a custom configuration, negotiate a discount outside the standard pricing structure, or require different contractual terms. It's worth deciding upfront which of these situations should be flagged for review rather than automatically included in the proposal.

This is where Qwilr's Conditional Content comes in.

As part of the Smart Proposal Engine workflow, it allows you to define conditions that determine which content appears, right down to individual text elements, images, and contractual clauses. Pricing configurations can also follow predefined rules, helping ensure that the quote reflects the options selected for the deal.

And while rules decide what appears, AI Context shapes how AI writes. You add your brand voice, key messaging, and sales playbook once in your account settings, and Qwilr's AI applies them every time your team uses it, no matter who's prompting.

The value of having approved information flow into proposals is something ShowingTime has experienced firsthand. The company previously relied on manually created PowerPoint proposals and individual reviews for larger deals.

By connecting Qwilr with Salesforce, the team could pull customer information and pricing into its proposals, reducing errors and the need for manual review.

If you're wondering where to begin, take a look at the proposal templates your team already uses. You might find that several versions exist simply because of differences in industry, package, or region.

Those variations are a useful starting point for identifying the rules you can build into your proposal process.

And as those rules become part of your workflow, you can accommodate more buyer requirements without maintaining a separate proposal template for every possible combination.

Step 4: Make it easy for reps to follow the process

If you've worked with sales teams long enough, you'll know that getting everyone to change how they work can be a challenge, especially when experienced reps have their own tried-and-tested ways of putting proposals together.

One way to get ahead of this is to test your AI proposal workflow with a small group of reps before introducing it to the wider team. Ideally, include a mix of experienced sellers and newer reps, and have them create proposals for real opportunities using the new process.

As you test the workflow, pay attention to what happens along the way. Are reps still switching between tools? Do they repeatedly need to fill in details that already exist elsewhere? Are certain inputs confusing or unnecessary?

And sometimes, getting reps to use a new process comes down to how little you're asking them to do.

Take ExtensisHR, for example. Its sales team previously worked with lengthy pitch decks that could run to 30 or 40 pages. With Qwilr, creating a proposal became a matter of filling in around nine fields, with the process taking minutes rather than hours.

It's a useful reminder that the experience for the rep matters just as much as what you've configured behind the scenes. The fewer unnecessary steps they have to navigate, the easier it becomes to get them on board, especially with AI, which can be intimidating.

That being said, even a well-designed workflow won't convince everyone overnight. Your more experienced sellers may still need encouragement from their managers, while others might have suggestions for making the process easier.

Give them room to share that feedback, especially during the early stages, and use it to simplify the workflow before rolling it out more widely. After all, the easier you make the approved process, the more likely reps are to use it when they have a proposal to send.

Step 5: Test your AI proposals before you scale the process

By this stage, you've done most of the groundwork. But before your entire sales team starts creating AI-personalized proposals, it's worth checking how well the process holds up across different types of deals.

And I wouldn't limit that to whether the proposal looks good or reads well.

Try putting a few realistic scenarios through your workflow.

Start with a simple deal where all the required information is available, then move on to one with a more complex product configuration, or a buyer requesting something outside your standard offering.

Pay particular attention to three areas:

For example, imagine a rep preparing a proposal for an enterprise buyer who wants an implementation timeline shorter than your standard offering. But that doesn't mean the proposal should automatically promise the requested timeline.

Your team still needs to decide how to review and approve nonstandard commitments.

In Qwilr's Smart Proposal Engine, you can configure the workflow around the information reps need to provide and the conditions that determine how the proposal is assembled. It's worth testing those configurations with different deal scenarios, particularly when pricing, legal terms, or product requirements vary.

Once you've identified the gaps, make the necessary adjustments to your input fields, source information, or configured rules, and test the affected scenarios again.

You don't need to anticipate every possible exception before getting started. But you do want to be confident that your most common proposal scenarios work as intended, and that reps know when a deal needs additional review.

That's what gives you a more reliable foundation for expanding AI proposal creation across the team.

And there you have it! Five steps to help you bring AI into your proposal process while keeping your sales playbook firmly in place.

But before we wrap up, there's one more way AI can help your sales team, particularly once those proposals have landed in your buyers' inboxes. And it's worth exploring what happens next.

What else can AI do once your proposal has been sent?

Your reps might have several proposals sitting with buyers at any given time. Some have been opened multiple times, others have gone quiet, and a few might be attracting attention from stakeholders who weren't involved in the original conversation.

If you're using Qwilr, you'll already have access to this information through proposal analytics, where you can track who's viewing your proposals, which sections they're spending time on, and when buyers are showing signs of interest.

But what if your team is already heavily invested in AI and wants to bring that same proposal intelligence into the tools they use every day?

That's where Qwilr MCP comes in. It connects Qwilr with AI assistants such as ChatGPT and Claude, so you can ask questions about proposal activity, retrieve information across deals, and take supported actions without having to leave your AI workspace. And since Qwilr is now listed in the Claude and ChatGPT directories, connecting only takes a few clicks.

For example, you could ask:

• "Which proposals have received the most engagement this week?"
• "Show me the deals where buyers have gone quiet."
• "Which sections are buyers spending the most time reviewing?"
• "Who has viewed our proposal for [company name], and when?"

Our CEO, Mark Tanner, explains how this works in practice:

As you can see, you can take it further by connecting Qwilr with other tools in your workflow. Imagine asking your AI assistant to pull together a weekly summary of proposal engagement and send it to your team's Slack channel, highlighting the deals that might need attention.

MCP also supports creating new proposals from existing Qwilr templates using information from your CRM, so teams can bring more of the proposal process into the AI tools they're already comfortable working with. (Looking for ideas? Here are 7 proposal prompts to try with the Qwilr MCP.)

It's another way to make AI part of your everyday sales workflow, while getting more value from the proposal information Qwilr already captures.

Ready to put AI to work in your proposal process?

There's plenty to be excited about when it comes to AI proposals, particularly as the technology continues to evolve and more of the work involved in creating and managing proposals can be automated.

And whether your team is already experimenting with AI or still figuring out where to begin, you don't have to change everything overnight. Start with the information and processes you already have, identify where AI can make the biggest difference, and build from there.

If you're curious to see what's possible, explore Qwilr's Smart Proposal Engine and Qwilr MCP. And if you'd like to see how either could fit into your sales process, our team would be happy to show you around. Book a demo.

(The Smart Proposal Engine is part of our Scale plan, and automating proposal creation may use paid credits.)