Jason Lemkin's team cancelled Notion after seven years. No complaints, no support tickets, no competitor win. Their AI agent absorbed the product's last use case: running staff meetings. Usage went to zero before customer success could intervene.
ASX-listed furniture retailer Nick Scali forced parts of its network offline following a cyber incident, leaving sales and customer service teams working from manual records. The company says orders and deliveries continue, but response times are slower. No evidence of customer data breach yet, AFP and Cyber Security Centre notified.
Atlassian eliminated Loom's Creator Lite role (free viewers inside paid workspaces) and auto-upgraded them to paid Creator seats. A workspace with 10 recorders and 90 watchers used to pay for 10. Now it pays for 100. The free seats were the distribution engine.
Jason Lemkin, SaaStr founder and former EchoSign CEO, tested 30+ APIs for production use. Exactly one vendor followed up to ask if the product worked. The gap between API sales motions and what technical buyers actually need is wider than most vendors think.
SaaStr deployed 20+ AI agents, cut headcount from 20+ humans to 3, grew revenue 47% YoY. The catch: AI multiplies your existing playbook. If your messaging doesn't work at human scale, 10x zero is still zero.
SaaStr CEO Jason Lemkin says Adobe Marketo went down for a day and a half, broke unsubscribe links for two weeks, and missed a 450,000-subscriber send. Then Adobe asked for a 20% renewal increase. Lemkin runs SaaStr with 20+ AI agents and says the company has stopped hiring human SDRs and AEs: legacy SaaS cannot be operated by agents, so it gets replaced.
Jason Lemkin's team moved 10 years of data off Marketo after their AI agent kept hitting API rate limits. The agent recommended the switch. The migration cost $14 in agent time. This is what procurement looks like now.
Jason Lemkin's team trained an AI agent to flip deals to Closed Won, generate invoices, chase payment, and calculate commission. Took 4 real deals to get it right. Now runs end to end with human on the loop, not in it.
Jason Lemkin's SaaStr AI scaled to 20+ production AI agents handling SDR, BDR, and support work while cutting human headcount to just 3. Revenue up 47% YoY. The agents generated $1M+ in direct sales, but they still need traditional CRM infrastructure. Here is why ripping out your sales stack for a Postgres database does not work.
Three B2B acquisitions in 30 days, all around $3B, all buying the same thing: proprietary data to train AI models. Salesforce bought Fin (formerly Intercom) for $3.6B, but the real story is a $100M AI agent line growing 350% YoY. The legacy helpdesk business? Flat. For ANZ sales teams: this validates outcome-based pricing and shows legacy seat-based models are being commoditised.
Jason Lemkin's AI VP of Marketing executed 125 actions in an hour for $13.42, less than California minimum wage. Monthly cost: $94.51. For context, a human VP of Marketing costs around $145 per hour fully loaded. The gap between building AI agents (expensive) and running them (cheap) is reshaping cost models for B2B teams.
A founding customer churned after nearly two decades because Adobe Marketo raised pricing instead of offering a retention discount. The case highlights a common trap: legacy vendors protect NRR metrics by letting at-risk customers walk, rather than discounting to buy time for product improvement. For account managers, the lesson is clear: renewal discounts work when migration friction is high, and losing a logo costs more than the NRR hit.