Kill the GTM stack: Aurasell CEO's case
Jason Eubanks ran 22 tools, paid $3M in fees, and needed 11 ops people to hold it together. Here's what he says the next model looks like.
At SaaStr AI 2026, Aurasell co-founder and CEO Jason Eubanks skipped the AI futurism. He put the exact go-to-market stack he ran at his last company on screen, told the room what it cost, and made a specific argument for why the whole model is breaking. We pulled the key data points from his presentation because they're the kind of numbers that make the abstract real.
Jason Eubanks' GTM stack at Harness ran 22 products, cost over $3M per year in software fees, and required 11 ops people just to maintain it. His argument: adding AI agents to that fragmented stack makes the problem worse, not better. The fix is a unified data layer first, then automation built on top of it.
The 24-30% problem nobody measures
Eubanks opened with a number most revenue leaders don't track: B2B sellers spend 24-30% of their time in front of prospects and customers. That's it. The other 70-plus percent goes to context switching, manual account research, prep, follow-up, and internal overhead like deal reviews and QBR prep.
None of that is selling. All of it is work you're paying full quota-carrying salaries to perform.
His challenge to the room was direct: if you can't say what that percentage is for each of your reps today, go measure it. He treats selling time as a top-line KPI, not an HR metric. That framing matters. When you treat it as a productivity metric rather than a morale metric, you stop managing it with motivation and start managing it with structure.
What the legacy stack actually costs
Eubanks didn't theorize. He put his old numbers on the screen from his time at Harness.
The stack ran 22 products. It cost $3M or more per year in software fees. It required 11 ops team members just to keep it standing. Those 11 people weren't driving revenue. They were stitching together integrations, patching workflow layers that conflicted with each other, and reconciling data across multiple databases. The one output they actually wanted, a single view of the customer journey, stayed permanently out of reach.
The deeper audit came through an exercise he called Project X-Ray. Mid-COVID, his board asked him to cut burn and accept slower growth. So he logged every activity, every tool, and every overlap across the org. The finding that stuck: reps were working inside 10 to 12 products a day to do their jobs, bleeding time to context switching the entire way.
That's the math of tool sprawl when it's concrete. It's not an abstract efficiency concern. It's 11 salaries, $3M in fees, and reps spending three-quarters of their day not selling.
Why adding agents to this makes it worse
Most legacy vendors are responding to AI pressure by bolting agents onto their existing product. Eubanks named why that backfires.
Every niche tool in a typical GTM stack carries its own siloed database. That silo might sync with your CRM at the field level, which looks clean on a data map. But the context -- the actual conversations, activities, and signals -- stays trapped inside each silo. That context is what an agent needs to act intelligently. Without it, the agent is guessing.
When legacy vendors layer agents onto fragmented data, you get agents running on a fraction of the relevant metadata, blind to the full picture. Eubanks calls the result "agentic thrash": low-quality automation at best, and at worst agents that act autonomously and step over each other. Adding more agents to a fragmented stack doesn't fix sprawl. It compounds it, and drives costs up while doing so.
This is the part of the AI sales pitch most vendors are quietly avoiding. The capability of the agent is only as good as the data feeding it. If the data layer is a patchwork of 22 tools, the agents inherit all the gaps.
The architecture Aurasell is betting on
Aurasell's architecture starts from the data, not the agents. Three layers:
A unified data foundation. Structured and unstructured data in one place. The platform ships with 900M contacts and 85M accounts, auto-enriched, with room to extend through custom enrichment.
A conversational context layer. Every conversation, every channel, every signal, feeding one context graph instead of a dozen silos.
An automation layer on top. Some agents come prebuilt and run autonomously. Others you build in natural language. The coverage is described as contact to contract -- the full sales process for both your team and the buyer.
The deployment model is the part worth watching for adoption. You can run Aurasell as your AI-native CRM and migrate off your existing tools. Or you can run it on top of Salesforce or HubSpot as an intelligence layer. That second option lowers the switching cost for teams that can't rip out their CRM this quarter, which is most of them.
What this means for operators building now
The Eubanks presentation is useful even if you never buy Aurasell. The Project X-Ray exercise is reproducible. Log every tool your team touches in a week. Log the overlaps. Count the ops headcount required to maintain the integrations. Then ask whether the productivity gains from each tool justify what it actually costs when you include the maintenance burden and the context-switching tax on your reps.
Most operators who run this exercise find the answer is no for a meaningful portion of their stack. The 22-tool number at Harness isn't an outlier. We've seen similar patterns across founders we work with who are building lean GTM motions and realize their tooling overhead is quietly eating the margin they thought they had.
The AI-native CRM argument Eubanks is making is a bet that the platform layer is about to consolidate the way the productivity suite did. Whether Aurasell wins that race or someone else does, the underlying logic holds: agents built on fragmented data don't solve the problem. They accelerate it.
For operators building inbound through LinkedIn and content, the parallel is worth sitting with. If your distribution stack has the same fragmentation problem -- different tools for scheduling, analytics, engagement tracking, and outreach that don't share context -- you're running a version of the same tax Eubanks described. The 24-30% selling-time number applies to content operators too. Most of the time goes to finding the right posts, writing the right comments, and figuring out what's working. The part that's actually generating pipeline is smaller than it looks.
That's the operational problem Chime is designed around. Not posting more. Not more tools. A tighter loop between the right engagement and the right context, so the time you do spend on LinkedIn is the 24-30% that actually moves the needle, not the overhead that surrounds it.
The GTM stack audit Eubanks ran at Harness is the kind of exercise that gets uncomfortable fast. The patterns across LinkedIn's top creators show a similar consolidation happening on the content side: the operators generating real inbound aren't running more tools, they're running fewer with more intentional distribution.
Frequently asked
GTM stack consolidation means reducing the number of separate tools your revenue team uses to do their job, and replacing fragmented point solutions with a unified platform that shares a single data layer. It matters because tool sprawl forces reps to context-switch constantly, which reduces the time they spend actually selling. Aurasell CEO Jason Eubanks found that reps at Harness were working inside 10 to 12 products a day, and the overall stack required 11 ops people and $3M per year just to maintain. Consolidation reduces that overhead and gives AI agents the unified context they need to work effectively.
