The race to dominate go-to-market automation has become so frenzied that simply tracking who launched what, and when, now constitutes its own full-time beat. AI-powered agents for sales, marketing, and revenue operations are everywhere—announced in press releases, whispered about in Slack channels, and occasionally... completely imagined.
Sorting the real from the rumored has become surprisingly difficult.
Take the curious case of "Kami," an open-source GTM platform that industry chatter suggested was launching this year. Except it wasn't. Doesn't exist, actually—at least not in the form described. More on that mystery later. First, the platforms that are genuinely reshaping how companies think about revenue operations.
When Enterprise Moves Fast
ZoomInfo, the B2B data giant, went live with GTM.AI on June 1. Not another dashboard. Not another sales tool. Instead, the company built what it calls a "headless context layer"—infrastructure designed to sit underneath other AI agents and keep them honest. The thesis: autonomous agents are only as good as the data they're trained on, and if that data is stale or fabricated, well, you've just automated garbage.
"Ground every AI agent in verified GTM data," the company says. It's a defensive play as much as an offensive one, positioning ZoomInfo as the connective tissue in what's becoming a fragmented ecosystem of specialized agents.
Highspot had jumped in a few weeks earlier with its own GTM Agent, unveiled on May 4. The pitch there leans more ambitious—consolidating strategy, execution, and performance into a single AI-driven system. Whether that's genuinely achievable or just aspirational packaging remains an open question, one that early adopters will answer over the next year.
Then there's the more dramatic move: Glyphic's reinvention as Airspeed on May 20. The rebrand wasn't cosmetic. It signaled a fundamental rethinking of what the product does—shifting from conversation intelligence to what the company now describes as an "agent-native GTM execution platform." The language matters. "Agent-native" suggests workflows designed from scratch for AI, not bolted onto legacy systems.
Seismic rolled out AI enhancements earlier in the year, back in February. 6sense demonstrated lead-scoring capabilities at a B2B marketing summit in late June. The cadence is relentless. If you're a major player in this space and you haven't articulated your AI agent strategy by now, you're reading as behind—whether you actually are or not.
The Builders in the Shadows
While the enterprise vendors compete on integration and polish, a parallel movement has been unfolding in GitHub repositories and developer forums. Open-source GTM tooling doesn't generate the same headlines, but it's gaining traction among technical founders who'd rather assemble their own stack than pay recurring enterprise fees.
Autogtm, for instance, bills itself as an open-source AI GTM engine. Daily lead discovery, automated scoring, hookups to outbound tools like Instantly. The GitHub repo shows active commits as of midsummer, though documentation remains sparse compared to what you'd expect from a commercial product.
A developer named Federico De Ponte released something called opengtm on PyPI in April—version 0.1.0, MIT-licensed, modular. Discovery, research, qualification, messaging, even "AEO health checks." It's less a platform than a construction kit, aimed at teams willing to write some code in exchange for not paying subscription fees.
Perhaps the most telling open-source project: GTM-Bench, which arrived in June with an accompanying academic paper. It's a benchmark suite specifically for evaluating agentic GTM tasks. The subtext is pointed—marketing claims about AI agent capabilities have gotten ahead of actual performance, and someone needed to build a way to measure what these systems can really do.
Reddit threads from early summer capture the scrappier end of this spectrum. Developers sharing "GTM cofounder" skill packs for Claude, Cursor, and other AI assistants. Free, rough, customizable. Not products in the traditional sense, more like recipes that technical users can adapt.
The Mystery of Kami
Which brings us back to the platform that isn't.
Whispers about "Kami" as an open-source GTM automation tool have circulated for weeks, but the evidence trail goes cold fast. Multiple companies use the name, none in GTM. Kamiapp.com? That's a K-12 education platform, serving tens of millions of teachers and students—its recent updates involve classroom AI features, not sales workflows. HeyKami markets itself as "your AI cofounder," but the landing page offers little detail and no product repository.
Kamiwaza AI did announce a 1.0 release in May, but it's an orchestration tool for regulated industries—healthcare, finance—not go-to-market automation. Other entities named Kami include a senior care platform, a creator economy service, even an open-source design system for professional documents. One GitHub repo, tw93/Kami, is actively maintained but focused on document formatting, not revenue operations.
No Product Hunt listing. No matching GitHub repository connecting "Kami" with "GTM automation platform." The confusion likely stems from name collision in an AI landscape that's both crowded and moving too fast for its own clarity.
It's a useful reminder that not every rumor materializes, even in a sector where new launches seem to arrive weekly.
What Actually Matters Now
The legitimate platforms represent genuinely different bets. ZoomInfo is wagering that the future looks like multiple specialized agents sharing a common data foundation. Highspot is betting on consolidation—one system to rule them all. Airspeed is designing for a world where agents aren't assistants but primary actors in GTM workflows.
The open-source camp offers something else entirely: control and flexibility, with assembly required. For founders comfortable stitching together tools, projects like opengtm provide building blocks without monthly invoices. For teams that need enterprise support and polished interfaces, the commercial platforms make more sense. Neither approach is inherently superior; the right choice depends on technical capacity and risk tolerance.
Forrester published research in late April arguing that traditional go-to-market models are collapsing under AI pressure, calling it the "GTM Singularity." Maybe. But the immediate challenge isn't existential—it's tactical. When multiple vendors claim agent capabilities, distinguishing real automation from rebranded workflows becomes essential due diligence.
The platforms are proliferating. What remains unclear is which architecture will actually simplify go-to-market operations rather than just adding another layer of complexity. Centralized enterprise systems? Open-source assembly? Some hybrid that hasn't fully emerged yet?
The market will decide, though probably not quickly or cleanly. For now, the only certainty is that the go-to-market stack isn't getting simpler anytime soon.
