In an August 5 disclosure, Salesforce said President and Chief Engineering & Customer Success Officer Srini Tallapragada will step down effective August 6 and become Special Advisor to the CEO through August 2027, ending a 14-year run. Marc Benioff promoted Chief Revenue Officer Miguel Milano — who ran Salesforce’s European business for a decade before a stint at Celonis — to Chief Operating Officer, and named recent Microsoft hire Rohan Kumar Chief Platform and Engineering Officer, with Alexa Vignone taking the revenue chief seat. It is the most significant reshuffle of Benioff’s direct reports in years, and it lands three weeks before earnings.

After sliding on the August 5 management news, Salesforce traded around $192 on August 7 (roughly $190.60–$194.70 on the day, up about 2.6% from the prior $186.77 close), for a market cap near $157 billion. That leaves the stock up roughly 16% over 30 days and about 31% above its June 26 low of $146.32, but still down around 30% year-to-date. The tape is behaving as though investors read the reorg as decisive rather than destabilizing — a bet the August 26 print will have to justify.

Salesforce reports fiscal second-quarter results after the close on August 26. It is the print that has to convert a busy month — the DoD IL5 authorization, the VA Missionforce ramp, and steady Agentforce KPIs — into current remaining performance obligation (cRPO) acceleration. Now it carries an added layer: it is the first quarter under Milano as COO and Kumar over platform and engineering. Good Agentforce metrics that still don’t inflect cRPO would likely hand back the summer’s bounce.

Salesforce announced IL5 authorization for Agentforce 360 through Missionforce National Security, letting the Department of War run autonomous agents on controlled-unclassified workloads — defense logistics, recruit onboarding and training, military-family administration, and turning complex data into real-time command insight — plus a low-to-high model to extend onto classified intelligence networks. Army Human Resources Command goes first, standing up 24/7 AI support for 9.2 million soldiers, veterans, and families. It moves federal AI past chatbots into accountable, context-aware autonomous execution — the highest-assurance validation Agentforce has yet received.

Following the summer general availability of the Shopper, Buyer, and Merchant agents — and native selling inside ChatGPT, Google Search, and the Gemini app — Salesforce’s Modern Point of Sale, iOS-only at launch, adds Android and Windows in August ahead of the holiday peak. With AI influencing roughly 20% of global online sales (about $262 billion) last holiday season, Salesforce is racing to own the agent that carries a shopper from discovery through checkout and into service — the consumer-facing complement to its enterprise and public-sector agentic push.

HubSpot posted Q2 revenue of $911.7 million (up 20% as reported, 17% in constant currency) and non-GAAP EPS of $3.26 against $3.02 expected, with subscription revenue of $894 million and operating cash flow of $222.8 million. Customers topped 306,000 (up 14%), but net new adds of about 7,000 fell short of the usual 9,000–10,000, and management guided as though Q2’s headwinds persist through year-end. Shares fell roughly 20% after hours — the clearest mid-market read yet on whether AI features can re-accelerate net revenue retention, or merely defend it.

After Morgan Stanley cut Workday to Underweight in July ($185 to $145 target) and CLSA piled on with an Underperform and a $92 target, the stock remains down roughly 29% year-to-date and about 42% below its 52-week high, with no fresh catalyst until it reports. The “persistent HCM deceleration” thesis has hardened into the market’s default skepticism toward legacy per-seat SaaS — the same factor bet weighing on Salesforce, just without a VA or DoD headline to offset it.

Gartner’s 2026 forecast holds: enterprise software spending rises 14.7% to more than $1.4 trillion, generative AI the primary driver, with 40% of enterprise applications expected to embed task-specific AI agents by year-end — up from under 5% a year ago. Global spend on AI-powered applications could reach $2.52 trillion in 2026, a 44% jump. The pie is expanding fast; the question every August print answers is which vendors actually capture it — and, increasingly, whether they capture it as seats or as consumption.

A survey of 300 SaaS CEOs found 97% intend to move off pure per-seat pricing within two years, with hybrid models — a base seat fee blended with usage- or outcome-based credits for AI features — now the most common structure being adopted. It is the pricing-model evidence underneath this summer’s “SaaSpocalypse” framing: the seat is not disappearing, but it is no longer the only unit anyone prices against — and HubSpot’s soft net-adds this week show what happens when the market fears a vendor is on the wrong side of that shift.