👋 Hi, it’s Kyle Poyar and welcome to Growth Unhinged, my weekly newsletter exploring the hidden playbooks behind the fastest-growing startups.

I’m running the second annual State of B2B Monetization survey to see how tech companies are adapting their pricing models in the age of AI. If you’re passionate about this topic, too, please consider taking the 5-minute survey. It’ll make us all smarter. The results will be shared back in this newsletter in May.

Your next customer might be an AI agent, whether you’re ready or not.

This could sound a bit like science fiction. Some of the early experiments, including Walmart’s agentic commerce partnership with OpenAI, haven’t panned out (they’re now testing a new model where Walmart’s AI agent gets embedded directly within ChatGPT).

But this isn’t as far off as it might seem.

AI already recommends which products to buy. It helps buyers negotiate with sellers. If you're building an app with an AI agent like Claude Code, you're primed to delegate at least some decision making to a trusted agent. And now AI agents are getting their own credit cards!

Some signs that agentic commerce is coming, even for B2B products:

Prospects are showing up ready to buy, spending less time on your website and more time in AI answer engines. Zero-click purchases feel like an inevitability, likely beginning with developer tools and commodity products. The question is: will you be ready when AI agents start buying?

Humans like simple, easy to understand pricing.

We get overwhelmed by having too many choices and would prefer perhaps three or four packages. Product jargon goes over our head, and we like clear solutions that address what we’re trying to accomplish with the product. Predictable, flat rate pricing reduces our anxiety about buying. Oh, and we love feeling like we got a deal.

Tech companies responded by keeping pricing relatively opaque. Only 50% of B2B tech companies publish their pricing online, according to my 2025 State of B2B Monetization report. Even among those who do have transparent pricing, enterprise plans tend to be gatekept behind a “Contact sales” or “Custom quote” CTA.

Pricing pages were never really optimized for buying anyway; they’re mostly marketing content to encourage prospects to take the next step.

AI agents don’t want to contact sales. Best case scenario, they’ll guess at what they think pricing looks like based on what they can find online. Worst case scenario: your “contact sales” CTA means you’re dropped from the agent’s shortlist of products altogether.

Agents are no longer tricked by behavioral pricing moves like charm pricing ($9.99) or the deal effect, either. Their evaluation will likely be structured and rational based on a set of decision making criteria.

This means that AI agents need content and details to make an evaluation. In fact, the more detail, the better. AI agents have endless compute power to navigate complex pricing and find a global optimum.

What AI agent buyers might like when it comes to pricing:

Answer engine optimization (AEO) has become the new SEO. Pricing needs to follow suit, and your pricing needs its own AEO makeover.

Here are eight things you can do right now to prepare for AI agent buying.

You might start with a simple: “How much does [your product] cost?”

Then go a step further including questions like “How does [your product] pricing compare to [closest competitor]?” or “Should I buy the [Standard] or [Enterprise] edition, and why?” or “Is it worth upgrading from the free tier to the paid plan, and why/why not?”

This is your starting point, and it’s what your human buyers already see today.

Technical products might already be seeing agentic purchases, particularly developer tools where Claude Code (or insert your AI coding tool of choice) automatically selects things like your database, payment stack, or analytics tooling.

Enterprise-focused products probably see a different form of this. AI procurement tools might review your RFP response, help your prospects negotiate based on pricing benchmarks, or redirect employees to the best product based on which tools have already been approved internally.

Measure AI influence on existing customer purchasing so you have a baseline for what’s happening right now, and you have an early warning when it inevitably changes. I’d use different metrics for different stages of the buying journey:

You might not be ready for fully transparent pricing, and that’s still OK. In that case, expose rougher pricing guidelines like “pricing is based on user licenses” or “pricing starts at $10,000 per year”. Or introduce an interactive cost estimator with rough pricing paired with expected ROI. This gives the AI agent bounds to reason within.

AI agents ideally want to understand the exact product capabilities and costs, structured in a way that’s machine-readable (ex: it has headers, meta descriptions, tables, precise product language). This documentation should be easy to find – ideally it’s linked directly from the pricing page.

You could take this a step further with a markdown version of the pricing page, as Vercel does. Pricing markdown files seem to be gaining steam with companies like Resend claiming they help agents become less confused about how pricing works.

AI answer engines like to outsource their thinking to others who do the research and collating for them. These engines often plagiarize grab things in chunks, which fits nicely into an FAQ format.

FAQ sections to consider:

LLMs want clarity on exact product capabilities to make sure the product fits their buying specs. For more complex products, this level of granularity might not easily fit within a pricing page or FAQ section.

GitLab solves for this with dedicated landing pages for both their Premium (“Why Premium?”) and Ultimate (“Why Ultimate?”) tiers. This includes a richer table with intricate product details, organized under clear categories.

While this post mostly focused on your own pricing documentation, LLMs won’t treat your website as the sole source of truth. They’re looking to find consensus across all available surface areas. Recent research from AirOps found that third-party sources account for 85% of brand mentions in AI search.

You’ll need to audit how third-party sites talk about your pricing including review sites (ex: G2, Capterra), Reddit, blogs, social media and analyst reports. Some of this content you control and edit yourself. For other sources you’ll likely need to do outreach to correct inaccuracies. Some companies are even going so far as sending cease and desist letters if not completed, although that feels a bit extreme to me!

You could imagine AI agents bringing their own purchase criteria, too. They might prefer the products that they’re able to use most effectively: which have the best APIs, which expose the most data, which have pre-built skills the agent can pull from?

Vercel does this by giving away product expertise in the form of open agent skills, reusable capabilities for AI agents. The skills make AI agents better at using Vercel (and better in general). It’s not much of a leap to imagine AI agents defaulting to tools they’re already trained to use, similar to how a Salesforce Admin might be predisposed to choosing Salesforce.

Agentic purchasing could eventually look like digital ad pricing, as Metronome’s Scott Woody predicts: programmatic buying, infinite customization, a direct link between price and performance, tools to stick within defined budget parameters.

We’re still (very) early when it comes to autonomous agentic purchases for B2B products. This theoretical future can be hard to wrap your head around, particularly for those who are accustomed to six-figure deals with custom, negotiated pricing.

I’d argue that your customers increasingly trust AI agents in the buying process because (a) AI is useful, (b) it feels objective, and (c) many customers don’t want to talk to a sales rep if they can avoid it. For now, AI agents act like an influencer on the buying committee; they’re informing which products get evaluated and how these products compare.

If you believe, like I do, that this trend will continue, now is the time to prepare. Track AI influence around different phases of the buying journey. Codify your existing packages and pricing. Create some external pricing documentation. And define clear rules and operating principles around discounting.

When AI starts buying, will you be in the consideration set?

Other things to click on: