Earlier this year, a JP Morgan report flagged something that should get the attention of anyone running an association: the S&P 500 software index has fallen into bear market territory, hitting its most oversold levels since 1990. Valuations have collapsed to where they were in 2022, when the world was bracing for recession.
The economy isn't the cause. Agentic AI is.
Agentic AI refers to AI that takes action rather than just answering questions. Tools like Claude Code allow users to interact with their computers in plain language to pull and analyze data, track expenses, write code, and complete tasks that entire SaaS companies were built to handle. Investors are now pricing in the possibility that if AI can do all of that, the business case for a lot of traditional software starts to get shaky. Even companies that should be benefiting from the AI boom — chip makers, AI labs — are getting caught in the sell-off.
For associations, there are two things to take seriously here. The vendors you depend on for your AMS, your LMS, and your event platforms are operating in an increasingly unstable market. And the organizations actively using AI are already pulling ahead — JP Morgan data shows those companies posting net margins two to three percentage points above their non-AI peers. That gap will keep growing.
None of this requires immediate action. But it does require awareness. Understanding what's happening in the market your vendors operate in is part of running a well-governed organization.
The first thing worth clarifying is that agentic AI doesn't degrade what your existing systems do. Your AMS still processes memberships. Your LMS still manages learning activities. Your event platform still handles registrations. The core functionality isn't going anywhere.
What has changed is the economics of building software. Development has become dramatically faster and cheaper. That means the barriers to entry are lower than they've ever been, and new competitors can come to market more easily than before. It also means that organizations themselves have more options. Rather than relying on a single platform to cover everything, it's now more feasible to pick best-in-class components for specific functions and wire them together. Take a strong e-commerce system, combine it with a solid CRM, and use AI-assisted development to connect them — that kind of approach, which would have been prohibitively complex a few years ago, is now within reach for many organizations.
That's genuinely good news in terms of choice. But it also creates real pressure on the vendors who have historically benefited from associations having few alternatives.
The vendors who respond well to that pressure are the ones investing in AI to modernize their products: cleaning up legacy code, improving user experience, and rethinking what their software can actually do when AI is built into it from the ground up rather than bolted on afterward. The vendors who don't respond well tend to focus on short-term margin, cut back on product investment, and eventually leave their customers behind. The challenge is that from the outside, those two types of vendors can look similar for a while.
If you haven't had a real conversation with your AMS or LMS vendor recently, that's worth changing. A support ticket doesn't count. Neither does sitting through a webinar about their latest feature release. An actual conversation with someone who understands the product roadmap and can speak honestly about where the company is headed — that's what's useful right now.
A few things that tend to come up when associations do this well:
Many discover their current system can do things they assumed it couldn't. Before concluding a capability gap is real, it's worth verifying. Associations often underuse what they've already paid for, and a conversation with the right person at the vendor can surface that quickly.
Many also find that the team they had relationships with a few years ago has turned over significantly. Getting to know who's actually running the product today matters.
And when you do have these conversations, push past the surface level on AI. Most vendors will have something to say about it. The question is whether what they're describing is a real product vision or a marketing response to customer pressure. A chat window added to a dashboard is a starting point. What you actually want to understand is whether they're rethinking the fundamental architecture of their product because of what AI makes possible. What does their system look like in three years? In five? If they can't articulate a clear and specific answer to that question at this point in 2026, that's worth sitting with.
User conferences, when vendors host them, are underrated for this kind of intelligence gathering. Getting in a room with other associations using the same system and talking honestly about what's working and what isn't tends to be more useful than any sales conversation.
There's a more immediate and practical issue that associations should be watching with their vendors, and that's pricing models.
Seat-based pricing — where your software costs scale with the number of users — made sense when humans were doing all the work inside software platforms. As AI agents take over more of those interactions, the logic starts to break down. An AI agent completing tasks inside your AMS doesn't need a seat. If you have five staff members today but fifty AI-assisted workflows running tomorrow, the traditional model doesn't reflect that reality.
Some vendors are already experimenting with usage-based pricing: charging per interaction, per transaction, or per task rather than per user. For associations that run on predictable fixed budgets, that kind of variable cost structure can be harder to plan around. It's a legitimate concern, and vendors who are thoughtful about this transition are building pricing models that account for it.
The vendors who aren't thinking about this tend to hold onto seat-based pricing for as long as possible. In the short term, that might look like a stable cost. Over time, it's often a sign that the company isn't seriously engaging with how AI is changing the way their software gets used.
Here's the thing about all of this market pressure: it doesn't mean associations should be scrambling to replace their tech stack. Major system replacements carry significant cost and risk, and the upside is almost always incremental rather than transformational. A new AMS, even a good one, tends to deliver modest efficiency gains. It doesn't fundamentally change what your association can do.
What AI actually makes possible is more interesting than that. Think about the member experience processes that have been quietly painful for years — the membership application that asks for thirty fields of information, some of which is already publicly available. The committee management workflows that your AMS handles poorly, even though committee work is central to what associations actually do. The conference abstract submission process that's frustrating for applicants and exhausting for the volunteers reviewing them.
These aren't problems that a new AMS solves. They're problems that get solved by rethinking the process itself and using AI to build something better. That work is now more accessible than it's ever been, and the associations investing time there are going to see returns that a system migration simply can't deliver.
That reframe matters when you're thinking about vendors, too. The question isn't whether to leave your current systems — it's whether your vendors are going to help you do the more interesting work, or get in the way of it.
The associations that navigate this period well probably won't look dramatically different from the outside. They'll have the same systems, the same general structure. What will be different is that they understood early what was happening in their vendor market, built the relationships to stay informed, and made decisions with that context rather than without it. That's less exciting than a technology overhaul, but it's what actually keeps an organization positioned well when things shift.