Programmable Revenue·Issue 02·June 30, 2026

The Proactivity Issue

The weekly research magazine of the graph8 and CIENCE research lab. Real papers, real measurements, no invented data, a human editor on the gate.

Pages this issue

Articles, all sourced

Weekly

AI newsroom, human-approved

Open

Free to read and cite

The letter

From the editor.

Thomas Cornelius · editor of record · June 30, 2026

Welcome to Issue 02.

Issue 01 was about the instrument: you cannot manage a level you cannot name. This issue is about the verb. Naming the level tells you where you stand. Proactivity is what you do next.

Three pieces, one argument. The feature reads a new paper on proactive outbound dialogue and finds that the agents that close are not the ones that talk more, they are the ones that probe the buyer’s hidden concern before they pitch. AI Research Corner reads a B2B uplift model that scores the lift, not the lead, and reminds us that most of what a pipeline counts as a win would have happened anyway. Strategy Corner takes the largest macro bet of the decade, the AI capital build, and asks what disciplined pipeline spend looks like when the dividend is still a probability and not a fact.

A number ties the issue together by accident. The uplift paper reports a 2.7x increase in incremental revenue per account. The macro model calibrates a 2.7x jump in AI-sector productivity. They measure different things and we do not pretend otherwise. But both say the same plain thing: the return is real, it is uneven, and you only earn it if you spend where the lift actually is.

The issue also carries three spotlights, agent-researched from public sources. Ron Gabrisko built the revenue engine that took Databricks past a reported $4B run-rate. Rakhi Voria has stood up sales-development and commercial orgs across four companies. Tailscale turned a publicly-reasoned free plan into its acquisition engine. These ship read-only this issue: we observed each public funnel, led with what it does well, and held the scored Pressure dimensions (speed-to-lead, the nine-day cadence, the composite score) for the disclosed mystery-shop, which fires only with sign-off and a five-business-day right of reply.

One section is still held. Trending Now waits on a clean telemetry window. We would rather hold a section than fake one. That is the standing deal, this week and every week.

The newsroom is a roster of AI agents. The gate is human. Every figure carries a source record, every number traces to a locator or shows its arithmetic, and anything we could not verify was cut.

We run the founder call every Tuesday, and a new issue lands the same morning. See you next Tuesday.

Contents 48 pages · 7 articles

p. 3 The verb is proactivity, and you only earn its return where the lift actually is.

Issue 01 named the level. Issue 02 is about the verb. Three papers, read together, say the same hard thing: the return on acting first is real, uneven, and only collected by spending where the incremental lift lives, not where the raw number is biggest.

From the Field · Agent: Research Desk, Agent: Math Desk, Agent: Psychology Desk, Agent: GTM Desk, Thomas Cornelius · 7 min

Literature p. 7 Probe the concern: a small model went from 26.2 to 48.5 percent acceptance by surfacing what the buyer would not say.

In a food-delivery outbound-call simulator, a small open model was conditioned on the buyer's hidden concerns and taught to surface them before steering. Simulated acceptance on the merchant task rose from 26.2 to 48.5 percent, within 2.3 points of the best flagship and far above a standard reward-shaping baseline at 30.5 percent. The lesson is about message, not model: probing beats pitching. All evidence is simulation-only and not yet peer-reviewed.

Feature · Agent: Research Desk, Agent: Math Desk, Agent: Psychology Desk, Agent: GTM Desk, Thomas Cornelius · 13 min

Literature p. 15 Score the lift, not the likelihood: a B2B uplift model tripled revenue per account by skipping the sure things.

Most accounts a propensity model ranks at the top would have bought anyway. A new preprint, VALOR, ranks accounts by how much a touch actually changes the outcome instead, beat its baseline by roughly 17 to 20 percent on Qini, and in one four-month field test moved incremental revenue per account from $445 to $1,185. We walk the tables, the arithmetic, and the limits.

AI Research Corner · Agent: Research Desk, Agent: Math Desk, Agent: Psychology Desk, Agent: GTM Desk, Thomas Cornelius · 14 min

Literature p. 24 Ron Gabrisko built a multi-billion sales engine on one idea: let the customer sell

Databricks reports a $4B revenue run-rate and 650+ customers paying over a million a year. Its CRO scaled the org from sub-$1M to that number across a decade by pricing on consumption, hiring sellers who can run the pilot, and naming the reference customer as the primary sales tool.

Leader Spotlight · Agent: Profiler Desk, Agent: Pressure Test Desk, Agent: GTM Desk, Thomas Cornelius · 9 min

Field p. 29 Rakhi Voria builds the room before she fills the pipe

The VP, Commercial Sales at Flock Safety has spent a decade standing up sales-development and commercial orgs at scale. Read-only, this pass: a look at the craft, and a first observation of the public funnel her segment feeds.

Leader Spotlight · Agent: Profiler Desk, Agent: Pressure Test Desk, Agent: GTM Desk, Thomas Cornelius · 9 min

Field p. 34 Tailscale: the free plan is the funnel.

Tailscale gives away a real product (zero dollars, up to 6 users, unlimited devices) and publishes the economics that let it. The mechanism is product-led acquisition fused to radical pricing transparency: the acquisition engine and the funnel are the same thing. We read the public funnel only. The enterprise hand-off and the score are pending a gated outbound pass.

Company Spotlight · Agent: Profiler Desk, Agent: Pressure Test Desk, Agent: GTM Desk, Thomas Cornelius · 10 min

Field p. 41 Spend where the lift is: a $381B AI bet, a 5-to-58-point GDP range, and one discipline for your budget.

The five biggest U.S. tech firms spent $381B on AI infrastructure in 2025 and guide to $755B in 2026. A new NBER paper reads that ramp as a rational bet and backs out a coin-flip-wide payoff: 5 to 58 points of GDP by 2030. The pipeline lesson is spend discipline: name the bet, hold out, measure the lift.

Strategy Corner · Agent: Research Desk, Agent: Math Desk, Agent: Psychology Desk, Agent: GTM Desk, Thomas Cornelius · 14 min

Literature

Colophon

End of Issue 02.

Produced by the Tenbound newsroom agents (Research Desk, Math Desk, Psychology Desk, GTM Desk, Profiler Desk, Pressure Test Desk) under the editorial gate of Thomas Cornelius. No invented data. Public sources only. Licensed images only.

The methods

Image credits Cinematic still: a row of black domino tiles on dark polished stone, the first already toppling and glowing in the brand gradient while the rest stand in shadow.: Tenbound (Illustration, AI-assisted (Higgsfield nano-banana-pro), cinematic conceptual 3D render, original concept, no photographic reference)Cinematic still: a dark faceted vessel on a plinth, its lid lifted just ajar so a blade of brand-gradient light escapes the seam from within.: Tenbound (Illustration, AI-assisted (Higgsfield nano-banana-pro), cinematic conceptual 3D render, original concept, no photographic reference)Cinematic still: an antique brass balance scale on dark velvet, the beam tipping toward a single disc glowing in the brand gradient.: Tenbound (Illustration, AI-assisted (Higgsfield nano-banana-pro), cinematic conceptual 3D render, original concept, no photographic reference)Engraved portrait of Ron Gabrisko, Chief Revenue Officer of Databricks: Tenbound (Illustration, AI-assisted (Higgsfield), Spotlight engraving from reference photo)Engraved portrait of Rakhi Voria, VP, Commercial Sales at Flock Safety: Tenbound (Illustration, AI-assisted (Higgsfield), Spotlight engraving from reference photo)Engraved illustration of the Tailscale product mark and wordmark: Tenbound (Illustration, AI-assisted (Higgsfield), Spotlight engraving, Tailscale wordmark)Cinematic still: stacks of matte betting chips on a dark felt table, one central column glowing in the brand gradient and one lit chip set apart, held back.: Tenbound (Illustration, AI-assisted (Higgsfield nano-banana-pro), cinematic conceptual 3D render, original concept, no photographic reference) Cite this issue Tenbound (2026). Programmable Revenue, Issue 02: The Proactivity Issue. tenbound.com/programmable-revenue.

Every issue in the archive