Scoreboard

Salesforce will keep Braze's standalone REST API and SDK publicly supported with no deprecation notice through Salesforce's Dreamforce 2027 conference, while steering net-new customers toward Data Cloud integration rather than standalone Braze.

Why Killing a popular API fast triggers churn, so incumbents keep it alive on paper. Where new feature work goes is the more informative measure, and Salesforce's ExactTarget and MuleSoft history shows feature gravity moves to the parent platform while the acquired API ossifies.

Right if: Braze's standalone API is still live with no end-of-life date announced and Salesforce is marketing Braze primarily as a Data Cloud add-on. Wrong if: Salesforce has published a Braze API deprecation or merged it into Marketing Cloud endpoints, or if Braze continues shipping major standalone features at its pre-acquisition pace.

Salesforce to Acquire Marketing Software Provider Braze for $5.7 Billion Read the source story →

US-token-spend growth on AI coding and agent tools will slow, with Ramp's quarterly AI spend data showing the year-over-year growth rate for Q1 2027 lower than the rate it reported for Q4 2026.

Why Ramp's latest data already shows token spend leveling out, and rising interest rates are raising the cost of the neocloud capital that subsidizes cheap inference. When that capital gets dearer, discounting slows and per-agent budgets get scrutinized, which drags the growth rate down even as absolute spend keeps rising.

Right if: Ramp's Q1 2027 AI-spend report shows a lower year-over-year growth rate than its Q4 2026 report. Wrong if: the Q1 2027 growth rate holds flat or climbs.

20VC: Horizontal vs Vertical Agents: Who Wins | Model Routing: Commodity or Increasing in Value | The Future of the Inference Market | Why Founders Should Take Discounts for Tier 1 Firms with Harrison Chase, LangChain Listen to the episode →

Magnite's agentic buying with LG will still require a human to approve each campaign before it goes live when Magnite reports Q3 2026 earnings in early November 2026.

Why The buyer-to-seller translation is hand-built per partner and not trusted to run unattended, agency clients like a Toyota dealer co-op won't accept an agent that can't explain its targeting, and Magnite gains nothing by removing the one safeguard that makes the launch sellable. Removing it would mean budget misallocated at machine speed with no audit trail, which no agency legal team will sign.

Right if: Magnite's agentic CTV buying still runs campaigns through a human approval step as of its Q3 2026 earnings report. Wrong if: Magnite has moved any production campaign to fully unattended agent-to-agent activation with no human approval before go-live.

LG and Magnite launch agentic AI buying for smart TV home screen ads Full Analysis → Read the source story →

By Google Marketing Live in May 2027, Google will ship a self-serve agent-building capability inside its ads stack that lets advertisers assemble multi-step automated workflows without custom code, validating the claim that the agent layer commoditizes within 12 months of today.

Why Google and Meta both already sell agentic automation (Advantage+, automated bidding) and are racing to keep activation inside their walls. Exposing agent-assembly tooling is the natural next step and directly serves their incentive to keep advertisers building on-platform rather than in an independent stack. Google Marketing Live is the canonical venue where Google announces ads-product capabilities to the market.

Right if: Google announces or ships a no-code agent-workflow builder inside Google Ads or its Marketing Platform by then. Wrong if: the most advanced agent tooling Google offers advertisers by that date still requires custom engineering to chain steps together.

Tinuiti CEO: Data Foundations, Not Agents, Are the AI Differentiator Read the source story →

Instinct will not publicly release an audited or third-party-verified figure for either its transaction volume or its user count before Invest Like the Best's next Noah Shinn appearance or a Sequoia/Benchmark portfolio disclosure, whichever comes first, through 2027-04-30.

Why Every impressive figure in this episode is founder-reported: "over $1 billion transacting," "10% day-over-day," "very small user base." These three don't sit comfortably together (10% daily growth compounds to roughly 30x a month, which can't run for long on any base large enough to move $1 billion), which is exactly why an audited breakdown would deflate at least one of them. A private, invite-only company raising at a $10 billion valuation has no reason to expose the denominators, and every reason to keep transaction volume (a gross, flattering number) separate from revenue and user count. Companies disclose numbers that flatter them and stay silent on numbers that don't; here the silence protects the headline. The opposite outcome (voluntary third-party verification) would only happen if a number got stronger, and founders who have strong numbers usually cite the source the first time.

Right if: no independent or audited transaction-volume or active-user figure for Instinct is published by then. Wrong if: Instinct, its investors, or a third party publishes a verified figure for either metric.

Noah Shinn - Building Instinct: The Personal Agent - [Invest Like the Best, EP.493] Listen to the episode →