In late 2025, Gene Lee, Co-Founder of Ramp, posted a candid LinkedIn announcement that landed across the GTM operator community: Ramp had shut down their Outbound Automation Team (OATs), the AI SDR program they had built and run for five years.

The post is worth reading carefully because it captures the honest read on where automated outbound is heading. Quoting Lee directly:

"But our environment has changed quickly. Today, there are dozens of automated outbound solutions, all operating on similar data which our competitors also have access to. We now have ~500 people in Sales who are world-class storytellers and problem-solvers. We have multiple flagship products to sell which means our selling motions are more complex. And the world is experiencing cold-outbound fatigue."

This is one of the most honest public statements about the state of fully-automated outbound from a company that ran it well for five years and accounted for "30% of all pipeline generated" through it. The plateau is real.

The interesting question is not whether Lee's diagnosis is right. It is. The interesting question is what Ramp built in its place, and whether their answer is the complete answer for the rest of the market.

What Ramp built instead: Growth Engineering

Ramp's stated direction is to invest in Growth Engineering — applying the same caliber of engineering talent they put into customer-facing products to internal GTM infrastructure. Lee describes the priorities:

"We're focusing on hard problems that won't be solved at scale by buying a solution: group selling and orchestration, multi-channel incentives and experimentation, CRM foundations and integrations, online and centralized customer data, a consolidated UI/UX for sellers... And we invest the same caliber of talent that we would place on teams building products for our customers into Growth Engineering."

Their Software Engineer, Product Growth role at Ramp confirms the shape. The job description reads: "Scale our customer success, sales, and marketing functions through automation and AI assistance... Own the customer journey from discovering and applying to Ramp through onboarding, activating onto our product suite, and expanding usage over time."

This is a coherent bet. Growth Engineering at Ramp's scale means embedding engineers into the GTM motion to build first-party data infrastructure, orchestration tooling, and AI-assistance that raises the floor and ceiling for all 500 sellers. Brendan Short summarized it well in the comments: the new function takes the orchestration and tooling layer that OATs had in a silo and generalizes it across SDR + AE + AM + CS.

The bet is that the long-term moat in GTM comes from owning more and higher-quality first-party data than competitors, and using that data to make every seller more effective. As Lee put it: "Slop in, slop out as far as AI and automation are concerned."

For Ramp specifically, this is defensible. They have one of the most data-rich GTM motions in B2B. Card-level transaction data on 50,000+ customers. Spend patterns. Vendor relationships. The first-party data layer is genuinely a competitive asset they can compound.

Where Ramp's bet is incomplete

Here is where I want to challenge the model.

The Growth Engineering bet focuses on personalization, deliverability, and orchestration — the things engineering can systematize. It optimizes the inputs to the sales motion. But the activation layer that actually closes senior buyers in 2026 is not primarily a personalization or deliverability problem. It is a relationship credibility and trust problem.

The engineering approach to sales gets the inputs right. It gets the message hyper-targeted, the timing optimized, the channel-mix balanced, the seller equipped with synthesized context. What it cannot do is manufacture the trust that makes a senior buyer take the meeting in the first place.

At the senior-buyer altitude — VP Sales, CRO, CFO, CIO — the data points that move the conversation are not better personalization. They are: who else at a peer company chose this vendor, what specifically did they get out of it, who on the buyer's own network can vouch for the founder or AE making the ask. Even a simple customer name-drop ("the CRO at [peer company] told me they cut their pipeline coverage ratio by 30% with this") or a relevant case-study reference can move a deal further than the most beautifully personalized AI-generated outbound sequence.

The relational layer compounds returns on the same first-party data. Two GTM teams with identical signal stacks will produce different pipeline outcomes if one has activated a relationship layer on top and the other has not. The team with activated relationships closes the meeting that the team without never gets booked.

This is the part the Growth Engineering bet misses. Or more precisely, it is the part Growth Engineering is not designed to solve.

The 2027 stack: Growth Engineering plus Relationship Orchestration

If Lee is right that the long-term moat in GTM comes from owning more and higher-quality first-party data, then the question for the rest of the market is what sits on top of that data layer.

The 2027 stack I expect to see in mature GTM teams will have three layers, not two:

The signal layer: intent data, first-party telemetry, account scoring. This is where 6sense, ZoomInfo, Apollo, and the existing intent infrastructure live. Ramp's first-party transaction data sits here too.

The orchestration layer: workflow automation, AI-augmented sellers, deliverability and personalization tooling. This is where Outreach, Salesloft, Apollo's sequencer, and Ramp's emerging Growth Engineering function live. This is the layer that activates the signal layer into outreach action.

The relationship layer: warm-intro orchestration across the four super-connector groups (your team, your customers, your investors and board, your advisors and partners). This is the layer that converts the activated signals into booked meetings at senior-buyer altitude. This is the layer Growth Engineering cannot build, because the underlying asset is relational, not technical.

Teams that have only the first two layers will produce high-volume, well-targeted outbound that converts adequately at the lower-altitude buyer levels and stalls at senior-buyer altitudes. This is roughly what the current state of the art produces. It is also what Ramp's announcement reflects: the OATs program was world-class on layers 1 and 2, and the return on it plateaued anyway.

Teams that add the third layer — the relationship layer — close senior buyers at the rate the signals suggest they should. This is the next architectural step.

From the trenches

The pattern I keep seeing across our customer base: companies that have invested deeply in their signal infrastructure (6sense, ZoomInfo, Apollo, or homegrown equivalents) and their orchestration layer (Outreach, Salesloft, Gong) still see their senior-altitude pipeline structurally constrained when they have not built the relationship layer on top.

A specific shape: a SaaS company with a healthy intent-data signal that fires on the right ICP accounts. The AE sequence into the named champion at the senior altitude. The reply rate is the new normal of 2026 — under 3 percent. The AE does everything the playbook says. The signals do not convert at the senior-buyer level.

What changes outcomes: the customer-name-drop or board-member-vouch that the engineering layer cannot manufacture. The CRO at a peer company who introduces the vendor by name. The board member who says "you should look at this." The advisor who has a prior relationship with the prospect's CISO. Each one of these turns the same first-party signal into a booked meeting that the most beautifully orchestrated sequence could not produce.

Even at Ramp's scale and data sophistication, I would bet the senior-buyer altitude has the same structural ceiling. Their 500 world-class storytellers are reaching prospects with better content than anyone else. The question is whether prospects open the email at all, and whether senior buyers take the meeting. That gate is increasingly relational, not informational.

Where Boomerang fits

Boomerang does not replace Growth Engineering. We do not compete with the first-party data layer. We do not compete with orchestration tools.

We are the relationship layer on top.

For every priority account, Boomerang surfaces the warm paths that exist across the four super-connector groups. When a signal fires from your existing stack (6sense, ZoomInfo, Apollo, internal intent data, including the kind of first-party data Ramp is investing in), we surface the strongest warm path into the account, draft the intro request in the connector's voice, and close the loop when the meeting books.

The architecture pattern I expect to see in 2027 winners: signal infrastructure + orchestration + relationship orchestration. Ramp's bet covers two of three exceptionally well. The third is what the rest of the market will increasingly compete on.

What to do this quarter

Three operational moves for GTM leaders thinking about this:

Audit your senior-buyer-altitude pipeline conversion rate against your lower-altitude conversion rate. If senior-altitude is structurally lower (it usually is by 2 to 4x), you have a relationship-activation gap that orchestration alone will not close.

Map your existing relational coverage across your four super-connector groups against your top 100 target accounts. Most teams discover they have warm paths into 40 to 60 percent of their priority accounts that they have never operationalized.

Build the operational motion that converts surfaced relational coverage into intro requests, then booked meetings. The motion is not technical. It is a cadence and an asking discipline. The teams that build it produce a meaningfully different shape of pipeline than the teams that do not.

For the broader read on why the AI SDR motion plateaued, see The AI SDR Backlash and Cold Email Reply Rates Fell 70 Percent. For the four-group framework of relational coverage, see The Customer Referral Engine, The Investor Warm-Up Play, and The Partner Co-Sell Play.

For the broader architecture of relationship-led pipeline, see our warm introduction software page.

Lee's OATs announcement is one of the most honest public statements about the state of automated outbound from a credible operator. The diagnosis is right. The path forward Ramp has chosen — Growth Engineering — is necessary but not sufficient for most teams. The companies that will win the next architectural cycle are the ones that combine the engineering layer with the relationship layer on top.

Signals plus relationships equals outcome. Engineering alone is half the answer.

Source: Gene Lee, Co-Founder, Ramp. LinkedIn post, late 2025. Reproduced excerpts under fair use with attribution. See also Ramp's Software Engineer, Product Growth role for additional context on the Growth Engineering direction.

Shankar Ganapathy is the co-founder of Boomerang, the operational layer for relationship-led pipeline. Before founding Boomerang, he led product in the account planning signals space.