The Wire

Real-time updates from ANZ tech sales

2 months ago
News

AI agents bypass Marketo, Outreach, Salesloft: why automation tools face existential risk

## The Tools Agents Don't Need SaaStr ran an experiment: ask Claude, OpenAI, and Gemini which APIs work best for agentic workflows. Stripe topped the list. No surprise there. What caught attention: when asked about marketing automation and sales engagement, all three models said the same thing. Marketo, Outreach, and Salesloft have no use in an agent-driven workflow. The reasoning: an agent crafts and sends better emails itself. It does not need a sequence builder, a template library, or a cadence tool. These platforms exist because humans cannot manually send thousands of personalized emails or track hundreds of follow-ups. Agents have no such constraint. They generate each message in real time, pull context from the CRM, and execute the cadence natively. The entire productivity layer disappears. ## What This Means for Sales Stacks SaaStr's own numbers show the shift. Salesforce spend went from $12k to $22k annually. Seats dropped from 10 to 2 plus one agent. Token consumption is up because agents run constantly. Meanwhile, the Marketo equivalent got cut entirely. Their AI VP of Marketing writes campaigns, segments audiences, sends emails, and measures results without touching a marketing automation platform. The pattern extends across categories: **Marketing automation:** Marketo, HubSpot enterprise, Eloqua. Built to let marketers template communications at scale. Agents generate each email fresh with full context. **Sales engagement:** Outreach, Salesloft. Built to run sequences because SDRs cannot track 400 cadences manually. Agents run cadences natively and generate each touch based on actual prospect behavior. **Conversation intelligence:** Gong, Chorus. Built to extract insights from call transcripts for humans to act on. Agents ingest transcripts directly and act without a dashboard layer. **Project management:** Atlassian (Jira, Confluence), Monday, Asana. Built for human coordination. Agents have memory and context windows. They do not need Kanban boards or wikis. Atlassian, the $4.4 billion Sydney-headquartered company, sits in an interesting position. Its tools are built for human coordination in software development and IT service management. Strong market position, but fundamentally designed around biological constraints agents do not have. ## The API-Native Advantage The reason some platforms work for agents while others do not comes down to architecture. Stripe exposes clean APIs with structured data and predictable workflows. Legacy B2B tools like Marketo, Outreach, and Atlassian were built around human-driven processes: configuration, manual handoffs, UI-centric usage. That makes them less agent-ready. Agents favor API-native workflows over platforms that require a human to click through screens. When the product itself is the workaround for human limitations, agents bypass the product entirely. ## The Ratio That Matters Gartner data shows vendor consolidation taking 30-50% of new AI spend. The first cuts: tools that exist purely as productivity layers for humans. Even if agents only handle 30% of these workflows by end of 2026, that is 30% of the customer base with no native need for the product category. This is not about agents using the same tools faster. This is about entire categories becoming redundant because the constraint they solved no longer exists. The sales engagement platform was a workaround for humans who could not personalize at scale. The agent does not need the workaround. ## ANZ Context For Australian sales teams evaluating AI tools: watch what gets consolidated first. Marketing automation and sales engagement platforms are high-risk renewal categories. The comp model for SDRs and AEs may shift as outbound volume becomes less about human touches and more about agent-generated quality. Atlassian remains the standout ANZ company in this conversation. Founded in Sydney, major global enterprise footprint, but the core products are still built around human coordination. The question for any ANZ tech employer in the collaboration or productivity space: does your product exist because humans have constraints, or because the problem itself requires human judgment? The agents have an answer. Some categories stay. Some become features inside broader AI systems. Some just get cut.

SaaStr
2 months ago
News

Startup Year axed: $15M program drew eight students, spent $80k

## Program cut after massive miss on targets The federal government's Startup Year program is dead. Eight students signed up. The budget was $15.4 million over four years, designed to support 2,000 loans per year. Total funding released: $80,000. Department of Education officials confirmed the numbers at Senate Estimates on Friday. The program was discontinued in May's federal budget. ## What the program promised Startup Year offered HELP-style loans for students and recent graduates joining university-run accelerator programs. Participants could access up to two loans of $11,800 each, repayable through the tax system like HECS debt. Labor first floated the idea in 2016, revived it in Anthony Albanese's 2021 budget reply, then allocated $15.4M after forming government in 2022. Years of consultation and legislative changes followed. The gap between plan (2,000 loans per year) and reality (eight total) suggests the problem was not awareness or legislation. Universities were supposed to deliver accredited accelerator programs. Either they did not build them, students did not want them, or both. ## Why this matters for sales hiring Startup accelerators are a common talent pipeline for early-stage companies building SDR and AE teams. When accelerators do not launch or attract participants, that pipeline stays dry. The university angle matters too. Campus recruiting and graduate programs are standard plays for building sales teams in ANZ tech. If university-backed startup programs cannot get traction, companies relying on that channel need other sourcing strategies. Worth noting: $15M is not massive funding by startup standards, but it is real budget. The 0.4% uptake rate (eight participants versus 2,000 target) is a data point for anyone building programs that assume students will trade future tax liability for current opportunity. No word yet on whether the eight participants who did sign up will keep their funding or face clawback.

SmartCompany
2 months ago
News

Nine writes off $49M Pedestrian investment, hands it to Vinyl for nothing

## The Numbers Nine Entertainment exited Pedestrian Group for nominal consideration. Total investment: $49 million across two acquisitions (60% stake for $9.3M in 2015, remaining 40% for $39.3M in 2018). Peak valuation: $100 million. Current valuation: effectively zero. Vinyl Group (ASX:VNL) is acquiring 100% of the youth media brand with no cash, debt, scrip, or ongoing royalties. Nine's ASX filing confirms the transaction followed unsolicited buyer enquiries, not a planned divestment. ## What Happened Pedestrian, founded in 2005, became part of Nine's digital media strategy during the traditional media pivot to online audiences. The business expanded through licensing deals (Business Insider, Lifehacker, Kotaku, Gizmodo, Vice, Refinery 29) and launched Pedestrian Television on 9Now as a youth-focused VOD channel. The cracks showed in 2022 when Business Insider's licence ended. A pivot to web3 content (The Chainsaw) failed as crypto imploded. By mid-2024, CEO Matt Rowley departed alongside multiple redundancies. Licensing deals for Lifehacker, Kotaku, Gizmodo, Vice, and Refinery 29 ended. Only PopSugar remained. Nine, which has collapsed from a $4 billion valuation to $1.18 billion in seven years, is offloading assets across the board: Domain sold, ACM regional print divested, radio stations and regional TV network also gone. ## Why It Matters This is a textbook case study in failed digital media investment. Nine bet $49 million on youth audience growth and digital advertising revenue. The hypothesis: build reach, monetise through programmatic and branded content deals. The reality: licensing costs, restructuring cycles, and a shrinking digital ad market. For sales teams in media or publishing, the lesson is clear: audience does not always equal revenue. Pedestrian had brand recognition and traffic, but could not convert that into sustainable margins. The web3 pivot showed desperation, not strategy. Vinyl is picking up a known brand for nothing, which either means they see capital-efficient upside in the audience, or they are betting on cost-cutting and cross-platform synergies with their existing publishing portfolio. Either way, Nine walked away from the entire investment rather than continue funding losses. Pedestrian co-founder Chris Wirasinha now runs adtech startup Linkby, which raised a $23M Series B in 2024. He got out early.

Startup Daily
2 months ago
News

Goterra enters administration after failing to raise growth capital

## Goterra enters administration after failing to raise growth capital Canberra-based ag-tech startup Goterra has entered voluntary administration after failing to secure the investment needed to scale operations. Teneo's Daniel Walley and Martin Ford were appointed administrators on Wednesday. The company is pursuing a going concern sale. ### The business model Founded in 2016 by former sheep farmer Olympia Yarger, Goterra uses modular, container-based systems housing black soldier fly larvae to process food waste on-site. The output: insect protein and soil conditioner. Customers include farms, restaurants, hotels, supermarkets, hospitals, and local councils across Australia. The pitch was decentralised waste processing with lower emissions and costs than landfill. The company raised an $8 million Series A co-led by Grok Ventures (Mike Cannon-Brookes' VC fund) and Tenacious Ventures. ### What went wrong In a statement to SmartCompany, a Goterra representative said the company has "working technology, contracted customers, and operating licences that are genuinely difficult to replicate." "This was not a product failure or a market failure," they said. "We ran out of runway while pursuing the investment we needed to scale." Worth noting: the gap between contracted customers and securing growth capital. That suggests the business was selling but could not close the funding needed to expand fast enough. ### What it means For commercial teams in climate-tech and ag-tech: contracted customers do not guarantee your next round. Goterra had technology, licences, and revenue. Still entered administration. Administrators are looking for a buyer to keep the business operating. If they find one, this becomes an acqui-hire story. If not, the contracted customers and operating licences get unwound. No public information on current sales team size or commercial leadership. The job was likely operations-heavy given the on-site processing model.

SmartCompany
2 months ago
News

Aurasell CEO: His old GTM stack cost $3M, 22 tools, 11 ops people

## The Stack Audit No One Wants to Run Jason Eubanks, CEO of Aurasell, skipped the AI vision slides at SaaStr AI 2026. He showed his old GTM stack from Harness instead: 22 products, $3M per year in software fees, 11 ops people to keep it standing. Those 11 were not driving revenue. They were stitching integrations, patching workflows, reconciling data across silos. Reps worked inside 10 to 12 products daily. They spent 24 to 30% of their time actually selling. The rest went to context switching, manual research, prep, follow-up, internal busywork. You are paying quota-carrying salaries for that. Eubanks calls it Project X-Ray: the audit he ran mid-COVID when his board asked him to cut burn. The finding that stuck was tool sprawl killing productivity, not headcount. ## Why Bolting Agents Onto Legacy Tools Fails Every niche tool brings its own database. That silo might sync with your CRM at the field level, but the context stays trapped. Conversations, activities, signals: all siloed. Agents need that context to act intelligently. Without it, they guess. Legacy vendors bolting agents onto fragmented data get what Eubanks calls "agentic thrash": low-quality automation, agents stepping over each other, costs going up. Aurasell's play: one unified data layer first, then agents. Structured and unstructured data in one place. 900M contacts, 85M accounts, auto-enriched. Conversational context across every channel feeding one graph, not a dozen silos. Automation layer on top, some agents prebuilt, others you build in natural language. The deployment is smart: run it as your AI-native CRM and migrate off legacy tools, or lay it on top of Salesforce or HubSpot with bidirectional sync. Rip and replace is optional. ## The Proof: $2.7M Closed in 41 Days Aurasell showed a new rep's first 41 days, ending in a $2.7M closed deal. Day one: territory already built and prioritized by ICP. No spreadsheets, no other tools. AI columns ran custom research at scale. Which accounts hired a new CRO this year? One prompt, pulled from reputable sources. Contacts pulled and ranked by propensity to engage, auto-enriched with email and phone. Sequences built by prompt, unique messages for every account and persona off the custom research. Cold call blocks surfaced with context attached: recent events, discovery questions, talk tracks. The company raised $30M seed from N47, Menlo Ventures, and Unusual Ventures. Co-founded by Eubanks and CTO Srinivas Bandi, both with prior runs at VMware, Twilio, Nutanix, Meraki, and Harness. The platform has logged 41 million agent runs. No ANZ office or local traction visible yet. Offices in San Francisco Bay Area, Bangalore, and London. ## What It Means for Sales Ops Stack consolidation is not new talk. What is different: Eubanks is not pitching a better dashboard. He is pitching one data layer that makes agents useful instead of adding more noise to a fragmented stack. The math is simple. If your reps are selling 25% of the time and working inside 10 tools daily, you have an ops problem masquerading as a headcount problem. The question is whether collapsing the stack into one AI-native platform actually works at scale, or if you are just trading 22 vendors for one very expensive vendor with a better pitch deck. Worth watching: does this play as consolidation theatre, or do the agent runs translate into quota attainment? The $30M seed says investors think the latter. The proof will be in customer retention and expansion, not demo walkthroughs.

SaaStr
2 months ago
News

Anthropic, OpenAI hiring sales faster than engineering: 20% of roles are GTM

## The numbers tell the story Anthroptic and OpenAI are hiring sales roles faster than any other function. At both companies, roughly 20% of open positions are in go-to-market: account executives, solutions engineers, partnerships, revenue operations. This is not a vanity hire. Anthropic's ARR sits at $47 billion, up from $9 billion six months ago. Enterprise customers account for 80% of that revenue. More than 1,000 businesses now spend over $1 million annually with Anthropic, double the count from two months earlier. OpenAI is planning to nearly double headcount this year, from 4,500 to 8,000 employees. Enterprise account executives are among the most aggressively recruited roles. ## Self-serve hits a ceiling The driver is deal size. A $1 million annual contract does not close through a checkout page. It closes through an AE who can navigate a multi-stakeholder buying committee, a solutions architect who maps the model to the customer's technical stack, and a CSM protecting the renewal. Anthroptic recently posted a Head of GTM Systems role. The job description says the company is scaling toward "multi-billion dollar revenue organisation" and needs systems for CRM, CPQ, finance, billing, order management, and revenue recognition. That is enterprise sales infrastructure, not startup motion. Other openings include GTM Strategy & Operations roles for Startups and Industry/EMEA segments. The company is hiring sales ops and planning, not just quota-carrying reps. ## What this means for sales professionals Three role types are prominent: **Forward-deployed engineers and solutions architects:** Technical sellers who embed in customer environments and ship integrations. Not pitch-deck AEs who present and leave. **Enterprise account executives:** Multi-stakeholder deal runners. The companies building the most sophisticated AI products in the world still need humans to close enterprise logos. **Revenue operations:** CRM admins, sales ops, systems architects. The back-office that makes a high-velocity sales org function. Anthroptic filed confidentially for a U.S. IPO on 1 June, days after a $65 billion round pushed valuation to $965 billion. A company that still leads with AI safety is walking into public markets on the back of an enterprise sales engine it has been building for months. Comp data for these roles is not public yet. Worth watching: if Anthropic and OpenAI are competing for the same enterprise AE talent as Salesforce, Google Cloud, and AWS, they will need to match or beat those OTE packages. That will set a new baseline for AI startup sales comp.

Sales Hacker
2 months ago
News

Vercel AI agents handle 93% of support, replaced entire SDR team

## The numbers Vercel is running hundreds of AI agents internally and posting specific metrics, per CPO Tom Occhino at SaaStr AI Deploy: - **96% of marketing content** now starts as AI-generated drafts from Slack threads - **93% of customer support inquiries** handled with no human intervention - **SDR team reabsorbed** after a lead qualification agent took over repetitive work Those are operating claims from a conference talk, not verified company disclosures. But they match what Vercel sells: AI infrastructure for developers building agents. ## What this means for sales teams Vercel's model is product-led, not sales-led. The company has historically run on self-serve developer adoption with enterprise layered on top. So when Occhino says they replaced SDRs with an agent, context matters: this is a developer tools company with a different motion than traditional enterprise software. The agent that replaced SDRs handles qualification, the mechanical part. The people who did that work moved to higher-impact roles, per Occhino. No word on whether "higher-impact" means sales, customer success, or something else entirely. ## The GTM agent: DealOne Vercel's go-to-market agent, DealOne, ingests sales calls, generates notes with action items, and posts coaching suggestions. It is built on Vercel's own AI SDK and runs in production. Occhino's thesis: every company will build custom agents, not buy off-the-shelf "agents as a service." You cannot buy a one-size-fits-all agent any more than you can buy a one-size-fits-all website. What you buy is the tooling to build and run agents cheaply. ## The reframe Occhino and CTO Malte Ubl (ex-Google, created Core Web Vitals) frame agents as a shift from UI-first software to autonomous software that acts on your behalf. The UI becomes a leaf node for human decisions. The trunk is headless automation. Their conclusion after two years: humans stop doing work an agent can do and move to work that compounds. The support team only touches hard or valuable tickets. Everything else is signal: product gaps or misconfigurations. ## What we do not know Vercel is a late-stage private company. Revenue is not disclosed. ANZ headcount or local operating scale is unclear from public materials. The SDR reabsorption and support/marketing percentages come from Occhino's talk, not a company earnings report or verified HR data. Treat this as a case study in how one well-funded developer platform is restructuring around agents, not a benchmark for sales org design across B2B. ## The broader pattern Multiple AI SDR and agent vendors (Artisan, 11x, AiSDR, others) are pitching automation for outbound and inbound qualification. Salary surveys show SDR comp holding steady in 2024, but role elimination concerns are real. Vercel is not eliminating the function, they are redeploying it. That distinction matters when you are deciding whether to take an SDR role in 2025.

SaaStr
2 months ago
News

Anthropic hits $183B valuation in 4 years, $9.4M revenue per head

## The numbers Anthropic, founded in 2021 by former OpenAI researchers, raised a $13 billion Series F in 2025 at a $183 billion post-money valuation. That is four years from founding to one of the highest private company valuations ever recorded. The company is doing roughly $5 billion in annualised revenue with about 5,000 employees. That works out to $9.4 million in revenue per person. For context: when Salesforce crossed $30 billion in revenue, it had roughly 79,000 people. When Google hit that mark, it had around 32,000. Anthropic got there with 5,000. Apple generates about $2.5 million per employee. Alphabet does $2.1 million. Anthropic is at nearly four times that rate. ## What this means for sales The old model assumed revenue and headcount scaled together. Every dollar of new ARR required more AEs, more SDRs, more managers to manage the managers. Revenue per employee crept up slowly as businesses matured. That link is breaking. When your product is intelligence delivered through an API, you are not adding humans to serve each new customer. The marginal cost of the next million in revenue is compute, not a bigger org chart. Anthropic's growth has been driven by enterprise API usage and strategic partnerships with Amazon and Google, not a traditional long-cycle B2B sales motion. The company does have a sales organisation, but its structure and comp are not publicly disclosed. For ANZ, Anthropic's footprint appears limited and mostly indirect. No large ANZ office or disclosed headcount, which suggests the region is served through global enterprise sales and cloud channels rather than a local team. ## The compression continues Cursor, the coding tool built by Anysphere, went from founding in 2022 to a $60 billion deal with SpaceX in 2026. Four years. The fastest path to an exit of this scale on record, and the largest tech acquisition ever. Wiz, the cloud security company, went from founding in 2020 to a $32 billion all-cash acquisition by Google in 2026. Six years, 30x ARR multiple. These are not outliers anymore. This is the new ceiling for AI-native companies that sell intelligence, not software. The old benchmarks do not apply.

SaaStr
2 months ago
News

SaaStr built AI VP of Marketing, runs it alongside human CMO

# SaaStr built AI VP of Marketing, runs it alongside human CMO Jason Lemkin's SaaStr spent over $500,000 building what it calls an AI VP of Marketing. The agent, named 10K, does not write content. It orchestrates: plans campaigns, manages pipeline in Salesforce, connects to Zapier and vendor APIs, and assigns daily tasks to the two humans running marketing. The build happened because existing AI marketing tools only generate content. SaaStr already has 5,000+ pieces of content across 13 years. The bottleneck was not more blog posts. It was knowing what to do, when, and executing a coherent plan across channels. 10K pulls from 5+ years of SaaStr data, feeds it into Claude Opus, and outputs a Replit app that plans every marketing action through mid-2026. It updates daily. When a campaign underperforms, it flags it on day two, not day sixty. No ego, no sunk cost fallacy, no defending pet projects. The operating model: run the AI VP and the human VP in parallel. Compare outputs. Debate recommendations. SaaStr says the AI's honesty and lack of agenda beats most human analysis, even if it is not better than a great human marketer. What it cannot do yet: manage third-party AI agents directly. 10K can analyse outputs from tools like Artisan or Qualified, but it cannot train them or change their behavior. Each agent has different guardrails, input processes, and training requirements. Lemkin says that orchestration layer, a single pane of glass that manages all AI agents, is a massive opportunity for someone to build. The broader shift: SaaStr had 20+ full-time employees in 2020. Today it runs 3 full-time humans and 20+ AI agents at the same revenue scale. The AI layer connects to 10+ years of data and automates workflows for a media/events business with multiple revenue streams: content, conferences, sponsorships. For sales teams watching this: the AI VP is not hypothetical. The comp is real, the workflow is live, and the early signal is that orchestration matters more than content generation. If your go-to-market stack includes multiple AI tools, the next question is who (or what) coordinates them.

SaaStr
2 months ago
News

SaaStr runs on 3 humans and 21 AI agents, books 614 meetings

# SaaStr runs on 3 humans and 21 AI agents, books 614 meetings SaaStr, the B2B media company and $200M VC fund, operates with 3 humans, 1 dog, and 21 AI agents in production. Their Amelia agent handled 2.2M website sessions, 442,000 chats, and booked 614 qualified meetings. Average sponsor ASP: $85k. The stack matters because it shows what GTM efficiency actually looks like when you cannot afford traditional headcount. No BDR team. No army of CSMs. Just agents that started as dashboards and evolved through 600 to 1,000 commits each. **The agents doing sales work:** - **Amelia AI:** Qualified inbound. Replaces 3 BDRs SaaStr could not afford. 614 booked meetings, $85k ASP. - **Agent Force:** Dead lead revival inside Salesforce. Highest open rate because it has the most context. - **Ava/Artisan:** Warm outbound to past attendees and lapsed sponsors. Recovered ~$500K this year. - **Monaco:** Cold ICP look-alikes. Fills its own funnel, books meetings autonomously. - **QBee:** Sponsor success. Manages 150+ accounts with personalized outreach and real-time risk flagging. No full Salesforce integration yet, already outperforms 85% of human CSMs at force-ranking health. **Why this matters for sales teams:** Owner.com provides the other half of the story. The restaurant SaaS company hit $100M ARR with 83% of new customers starting via Gradr, their free AI website generator. The product costs $1 in compute per restaurant and converts into $1/month subscriptions. Owner's CRO Kyle Norton took ARR from $3M to $21M in 22 months using AI-first GTM. Klaviyo, the public marketing automation company, is rebuilding product and engineering processes around agents. Not customer-facing features. Core internal workflows. **The connective tissue:** Headless Salesforce. None of these agents work if they have to use the UI. They use the API directly, in real time. **What this means for ANZ sales teams:** If a sub-10-person media company can book 614 qualified meetings and manage 150+ enterprise sponsors with agents, the traditional SDR-to-AE ratio is under pressure. Not next year. Now. The question is not whether to deploy AI agents. The question is whether your comp plan assumes headcount that will not exist in 18 months. Owner showed the blueprint: build the free AI product, convert to paid, bundle from there. 83% of pipeline starts with the agent. The AE closes the expansion. SaaStr showed the ops model: agents own qualification, revival, warm outbound, and sponsor health. Humans own high-touch enterprise closes and strategy. The gap between these models and traditional sales org charts is widening fast.

SaaStr
2 months ago
News

Lovable hits $400M ARR with 197 people: PLG at $2M per head

## The Numbers Lovable, the AI app-building platform, hit $400M ARR with 197 people. That is $2M+ in ARR per employee. For context: most B2B SaaS sits at $150k to $250k per head. They doubled from $100M to $200M in four months, then doubled again. Elena Verna, Head of Growth, spoke at SaaStr AI about what is actually working when AI writes 80% of your code and competitors can clone your features in a weekend. ## Feature Differentiation Is Dead Verna's thesis: when AI collapses the cost of building, feature leads last weeks, not years. You ship something new, competitors replicate it by next sprint. The 15-year B2B playbook of "better product, better engineers" does not hold. What still works as a moat: - Hardware (genuinely hard) - Network effects (hard to create, compound over time) - Proprietary data - Security and compliance (slow and expensive, which is the point) - Brand ("Brand is back, baby") Notably missing: SEO, SEM. When everyone can build the product, the relationship with the customer is what is left. ## The GTM Implication Lovable runs product-led growth, not enterprise field sales. Self-serve, rapid adoption, AI positioning. The scale came from distribution and workflow integration, not a 50-person sales org. Verna's take on org structure: everyone ships code, no traditional PM-to-engineer ratios. A `#shipped` channel with multiple production releases daily. A `#feedback` channel where ideas go live in 24 hours if one other person agrees. She also went from leading a 200-person growth team at Dropbox to firing herself out of marketing at Lovable and going back to IC. Her read: the next decade's career flex is the high-powered IC with AI agents, not the VP title. Worth noting for sales leaders watching quota-carrying AEs debate management tracks. ## What This Means for ANZ B2B If feature parity is a weekend sprint, sales teams cannot lean on product superiority. The pitch shifts to: distribution, integration, trust, compliance. Enterprise AEs selling against AI-native competitors need to know what moats their company actually has, because "better roadmap" is not one of them anymore. Verna says Lovable is still "on the product-market fit treadmill" at $400M ARR. The category moves so fast they recapture PMF every month. Scale did not let them slow down. It raised the stakes on velocity. No public data on Lovable's ANZ presence or sales team size. The growth engine appears to be product-led, not field-sales-led. That $2M per head number? That is the new benchmark when AI does the building.

SaaStr
3 months ago
News

Seven ANZ startups raised $17 million this week, quantum tech leads

Seven ANZ startups raised $17 million combined this week. That is split across early-stage deals, not the $50 million Series B rounds that come with 20-AE hiring plans. QuantX Labs led the pack with $7 million USD ($7 million AUD) in seed funding from Serendipity Capital. The South Australian quantum sensing startup builds optical atomic clocks for defence, satellite networks, and critical infrastructure. The tech is 10 to 100 times more stable than existing microwave-based systems, smaller, and portable. Applications include GPS resilience, radar networks, and telecommunications timing. Founded in 2016, QuantX Labs is commercial now, selling into defence and infrastructure. Quantum sensing is a crowded global market, but the defence angle and Australian sovereign capability positioning give them an edge for government contracts. The other six startups in the funding tally include Nardo (sports tech, backed by former Socceroo Tim Cahill), Alloovium and Gutgutgoose (both Y Combinator-bound), and four others not detailed in the source material. Combined, they raised $10 million. What this means for sales teams: These are seed and pre-Series A deals. Hiring plans will be measured, not explosive. QuantX Labs might add 1-2 commercial or partnerships roles given the defence sales cycle. The Y Combinator startups will focus on product-market fit, not go-to-market scaling yet. For context, ANZ startup funding is down 60% from 2021 peaks. Seed rounds are steady, but growth-stage capital is tight. The companies closing deals now are capital-efficient and commercial-ready. If you are tracking ANZ startup hiring, focus on startups 12-18 months post-seed: that is when the first AE hire happens. QuantX Labs is Adelaide-based. Quantum tech is a federal priority, which helps fundraising but does not guarantee fast deal cycles. Defence procurement moves slowly. Expect long sales cycles, high ACVs, and a small, specialised sales team structure.

SmartCompany
3 months ago
News

Harvey hiring enterprise CSMs at $190M ARR, Assembly AI ditches CS title

## The Old CS Playbook Does Not Work for AI Products Three fast-scaling AI companies shared what they are doing differently in customer success, and it cuts against a decade of SaaS conventional wisdom. The details matter because these are the companies setting the pace: Harvey at $190M ARR serving 100,000+ attorneys across 60+ countries, Assembly AI selling API infrastructure to technical buyers, and Lovable scaling product at speed. ## Harvey: Enterprise CS Is Still High-Touch Tom Ronen, VP Customer Success at Harvey, runs an on-site, executive business review motion that looks like 2001. That is deliberate. Harvey is selling AI into 200-person law firms where partners have practiced the same way for 30 years. No automated health score gets you there. The company is hiring Enterprise Customer Success Managers with explicit focus on value realization, adoption, ROI, and C-suite legal stakeholders across the US, Canada, and Latin America. Harvey is backed by Sequoia and OpenAI's startup fund, positioning against legal workflow and document automation platforms. **Takeaway for sales:** If your product requires deep behavior change inside a skeptical enterprise org, high-touch CS is a competitive moat, not a relic. ## Assembly AI: The CSM Title Is Broken Ryan Seams, Head of Customer Success and Solutions at Assembly AI, watched technical buyers change body language the second he said "customer success." They got defensive. The title now signals QBRs and renewal pressure, not technical partnership. Assembly AI tried renaming roles to Technical Account Manager. Recruiting died: two candidates in two and a half months. They switched to Forward Deployed Engineer. Pipeline filled immediately. The actual job barely changed. The brand did. **Takeaway for sales:** Audit what your titles signal to technical buyers. The words on your Slack profile change whether a customer leans in or shuts down. ## Lovable: Stop Saying AI Monica Perez, Global Head of Customer Success at Lovable, stopped leading with AI in customer conversations. Her argument: saying "AI-powered" signals you are behind, not ahead. AI is becoming baseline infrastructure. Lovable's onboarding opens with what the customer will unlock, not what the technology can do. The moment you stop saying AI in every conversation is the moment you become AI-native. ## What Actually Drives Retention Bobby Cooper, founder of Retention Intelligence, shared platform data showing over 50% of CSM activities have no correlation with retention. CSMs accrete work over time, become jacks of all trades, and teams bloat. The fix: map every activity to whether it changes the product outcome for the customer inside the intended timeline. If it does not, it has no place in the playbook. **Takeaway for go-to-market:** Run a time-and-motion study on your post-sales team before you add headcount. You are likely funding work that does nothing for retention. ## Adoption Is Not Enough Anymore For traditional SaaS, decent seat utilization plus the same buyers at renewal meant CS did its job. That breaks with AI products. A law firm can log into Harvey constantly without changing how it operates. Logins are not transformation. Harvey tracks five ROI pillars and asks CSMs to tag value stories against them, because the goal is reducing non-billable work and speeding up matters, not racking up active users. **Worth noting:** These are AI-native companies redesigning CS around product speed and technical buyer expectations, not legacy process. The competitive set matters. Harvey competes with legal-tech enterprise vendors. Assembly AI competes with speech-to-text and voice AI infrastructure providers. Lovable is growing extremely fast in a category where users build software in plain English. The CS model follows the product motion, not the other way around.

SaaStr
3 months ago
News

Token budgets hit SaaS: AI spend forcing sales and CS headcount cuts

## Token Budgets Are the New Headcount Fight Engineering leaders are getting a new budget question: keep your 400-person team, or cut to 300 and spend the other 100 salaries on AI tokens. The numbers coming out of early adopters: Uber caps engineers at $1,500/month in token spend, roughly $18k/year or 10% on top of a $200k engineer. Brandon at McCor says they now spend more on tokens than engineering salaries with 80 engineers. The EDA software market, historically the most tool-heavy segment, runs about 13% of engineering spend on tooling. Jason Lemkin's read: the best shops push past 10%, maybe a third or more. Rory O'Driscoll thinks that math is aggressive. Nobody has run a full org at that ratio yet. ## What This Means for Sales and CS Roles The first cuts in 2024 were about survival. The second wave is about choosing tokens over marginal headcount. Bottom of the QA list, CS roles that can be automated, inbound SDRs closing $3k deals. The roles that survived the first layoff wave are now being measured against AI cost. SaaStr already made the call: cut B players, spend on tokens instead. The bet only works if your VP of Engineering can ship product with fewer people. Fast-growing companies say yes. Slow-growing companies say it cannot be done. ## Why Comp Matters Here If corporate America settles at 10% token spend relative to engineering cost, you get to the Anthropic and OpenAI growth curves without mass layoffs. At 33%, you are looking at one in three or one in four roles across product and engineering getting replaced. Sales and CS roles that touch product or require technical context are in the blast radius. The other data point: Anthropic raised $65b and filed to go public in the same week. ARR up 28% since the last quarter. When the best company of the decade goes zero to trillion in five years, it warps the bar for everyone else. VCs are now screening for billion-dollar positions, not billion-dollar outcomes. That means faster growth, bigger TAMs, and less tolerance for blockers like small teams or capped markets. ## The Takeaway for Sales Professionals If you are in a technical sales role, ask your leadership what the token-to-salary ratio is for your engineering team. That number is your leading indicator for where headcount lands in 2027. If your company is slow-growing and the ratio is climbing, start looking. If you are fast-growing and leadership is betting on tokens, make sure your role is tied to revenue, not process. Comp transparency matters more now. Real OTE, realistic attainment, ramp periods. The market is resetting fast, and the roles that survive are the ones that prove they drive pipeline, not the ones that just touch it.

SaaStr
3 months ago
News

Adelaide AI email startup Nitrosend raises $700k seed

## Adelaide AI email startup Nitrosend raises $700k seed Nitrosend, an Adelaide-based AI email marketing platform, closed a $700,000 seed round led by Eastend Ventures, with participation from Archangel Ventures and Aussie Angels. The company was founded by brothers Edward and George Hartley, who previously built SmartrMail, an email automation tool that scaled to 6 billion emails sent before being acquired by Relay Commerce in 2022. They brought back their former CTO, Kam Low, as a founding team member for the new venture. ### What they are building Nitrosend positions itself as an AI-native email platform. Instead of drag-and-drop template editors, users describe campaign goals in natural language (via Claude, ChatGPT, or the Nitrosend platform), and the system handles writing, design, sending, and performance reporting. The platform also automates workflows like welcome sequences, cart abandonment, and customer segmentation. Target market: SMBs that lack dedicated marketing resources or agency budgets. ### Early traction The company signed 190 users since April 2026, including early-stage startups Elita Genetics and Fast Lane. No public customer count, revenue, or sales headcount disclosed. ### Market context Nitrosend sits in the crowded AI email marketing segment alongside Mailchimp, Klaviyo, ActiveCampaign, HubSpot, and newer AI-native tools. Their differentiation: 10 years of email infrastructure experience combined with AI workflow automation. For sales teams evaluating martech stacks, the platform could reduce reliance on marketing ops resources for campaign execution, though enterprise buyers will want to see more proof of scale. Eastend Ventures founding partner Josh Garratt: "Experienced founders going after a market they already understand better than anyone, at exactly the moment that market is being reshaped by AI." Worth noting: the company appears to be operating with a founder-led commercial motion rather than a dedicated sales org at this stage. No sales hiring announced.

Startup Daily
3 months ago
News

SaaStr updates VP of Sales interview guide: AI fluency now mandatory

# SaaStr updates VP of Sales interview guide: AI fluency now mandatory Jason Lemkin has updated his widely-shared VP of Sales interview framework for 2026, and the changes reflect what is happening on sales floors right now: AI agents are handling meaningful pipeline volume, title inflation is worse than ever, and the VPs who do not know their Clay from their Artisan are getting left behind. The original advice still holds. Do not hire a VP of Sales until you have proven the motion yourself, either by making 1 to 2 reps successful first, or by hitting $1M to $2M ARR with founder-led sales plus AI SDRs. If you skip this step, you will fire whoever you hire in 9 months. No exceptions. But the screening questions have evolved. The classic filters are still there: team size estimation, deal size fit, direct management experience, recruiting ability, tool stack opinions. If a candidate cannot answer those fluently, pass. What is new for 2026: two critical additions that separate modern sales leaders from people who just have the title on LinkedIn. **Question 5: How are you using AI agents in your current sales motion?** If they hedge, say "we are exploring it," or defer to ops, pass. The right answer sounds like specifics: Artisan for outbound, Qualified for inbound chat, Gong for call analytics, Clay for enrichment, Agentforce for win-back. If your candidate is not fluent here, they are going to be an expensive observer of what your competitors are doing. **Question 6: How has your ideal rep profile changed in the last three years?** Listen for evolution. The best VPs are hiring fewer SDRs because agents handle prospecting. They are hiring AEs who can run their own AI workflows. They are cutting the bottom 20% faster because AI raises the floor on what an average rep can produce. If they describe the same rep they hired in 2022, they have not been paying attention. The underlying reality: team sizes are often smaller than they would have been in 2023. If a candidate is pitching you 20 SDRs when 4 SDRs plus an Artisan deployment would do the same job, that is a tell. They are playing the old playbook. The full 15-question framework is worth reading if you are hiring or evaluating a VP of Sales role. The questions are designed to create dialogue and surface whether you have a real VP candidate or someone who just has the title. Give them enough data ahead of time so they can answer with substance. If they did not do the homework before the interview, that is your answer right there. One constant that has not changed: 50% of the job of VP of Sales is recruiting. AI does not recruit your reps for you. Your VP does. If a candidate cannot describe the reps they hired and where they found them, they are not a real leader. They are a manager who got promoted into something they do not actually do. Worth noting: this guidance aligns with what sales recruiters and sales ops leaders are seeing in the market. The bar for VP of Sales has moved. AI fluency is not a nice-to-have. It is table stakes.

SaaStr
3 months ago
News

SiteMinder share price jumps after 20-year pivot to AI booking channels

## The Setup SiteMinder, the Sydney hotel commerce platform founded in 2006, just announced integration with AI booking channels via Model Context Protocol. Share price is up. The company supports 41,000+ hotels across 150 countries, primarily independent and SME properties using their channel manager to distribute inventory across OTAs. The timing matters: SiteMinder got caught in the early 2026 software sell-off alongside every other SaaS company that wasn't immediately profitable. Their valuation dropped despite consistent revenue growth. Now they are pivoting product strategy while the market reassesses. ## What Changed SiteMinder is positioning beyond traditional channel management into what they call an "open hotel commerce platform." The Model Context Protocol integration means hotels can surface inventory through AI-powered booking agents, not just Booking.com and Expedia. The company has been delivering growth year-over-year. The question now: can they convert that growth into profitability while expanding into AI distribution? ## The Sales Angle For ANZ sales professionals, SiteMinder represents a rare local success story in hotel tech. Founded in Sydney by Mike Ford and Mike Prewitt, the company built a genuine global footprint in a category they essentially pioneered. What we don't have: current headcount data, recent ANZ hiring numbers, or specific comp ranges for their enterprise sales roles. Public sources confirm the founders but not the current CRO or VP Sales. That leadership matters when evaluating a "next growth phase" story. Worth noting: SiteMinder competes with integrated PMS vendors (Cloudbeds, Hotelogix) and standalone distribution tools. Their positioning has always been strongest with independent hotels that need channel management without full PMS replacement. ## The Numbers We Need The original article promises a financial teardown but the available sources don't provide ARR, revenue multiples, or EBITDA figures. For a SaaS comeback story, those metrics matter more than share price movements. If SiteMinder is genuinely entering a growth-plus-profit phase, the sales team structure and quota models would be shifting. Enterprise AE roles at a 20-year-old company moving toward profitability look different than at a growth-at-all-costs startup. ## What To Watch AI booking channels could expand SiteMinder's addressable market or commoditise their core channel management play. Either way, it changes the sales motion. If you are evaluating enterprise sales roles in hotel tech, the competitive landscape just got more interesting. For now: strong distribution footprint, proven founder story, unclear path to profitability. The AI pivot is worth monitoring but show us the attainment data.

SmartCompany
3 months ago
News

Meta AI chatbot let hackers reset Instagram passwords by asking nicely

## The hack that required no hacking Over the weekend, attackers took over multiple high-profile Instagram accounts by chatting with Meta's AI support bot. Targets included Barack Obama's former White House account, Sephora, and the US Space Force Chief Master Sergeant. The method: open a chat with Meta AI, ask it to add a new email address to the target account, paste the verification code the bot sends, then click "reset password." In some cases, attackers used a VPN to match the victim's region. At no point did they need access to the legitimate email already on the account. Meta says the issue is fixed and affected accounts are being secured. The structural problem has not changed: an AI agent with the authority to modify emails and reset passwords is a security control point, and this one failed. ## Why this matters beyond Meta Many ANZ SMBs use Instagram for customer acquisition, paid social, and brand presence. Account compromise interrupts lead generation and damages trust. The reporting shows that users hit by takeovers struggled to escalate to a human, which underscores the operational risk of relying on automated support for mission-critical recovery workflows. Meta rolled out AI-assisted support across Facebook and Instagram to handle account recovery tasks at scale. That decision made the chatbot a high-impact control point. When guardrails are weak, AI support does not just scale service: it scales risk. ## The support angle This is being framed as a hacking story. It is also a support story. Meta's AI agent had the permissions to change account ownership without sufficient verification. The same automation designed to reduce support load became the vector for account takeover. For sales and marketing teams that depend on social platforms for pipeline and customer engagement, the lesson is clear: verify what access your support automation actually has, and whether the trade-off between efficiency and control makes sense for your business. Meta's AI-assisted support is now being judged not only against social platforms like TikTok and Snapchat, but also against the trust standard set by identity providers and customer-support automation tools. This incident will be referenced in procurement conversations for months.

SmartCompany
3 months ago
News

Florida sues OpenAI, names Altman personally in first state lawsuit

Florida became the first US state to sue OpenAI and CEO Sam Altman personally, alleging the company knowingly released ChatGPT while concealing serious risks. Attorney General James Uthmeier claims OpenAI suppressed internal safety warnings and deceived users. The lawsuit references two shootings where alleged gunmen asked ChatGPT questions while planning attacks. Florida opened a criminal investigation in April after a shooting at Florida State University. OpenAI maintains its models "repeatedly encouraged the individuals to seek real-world support, including from mental health professionals." The company says it co-operated with law enforcement in both cases and that ChatGPT is "used by hundreds of millions of people every day for legitimate purposes." ## What This Means for Enterprise Sales If you are selling AI tools or using ChatGPT in your sales process, the regulatory environment just got more complicated. This is the first state-led action against OpenAI, but it follows existing EU GDPR concerns and enterprise compliance questions. Key considerations: **Procurement risk**: Enterprise buyers are now dealing with a vendor under active litigation. Security and legal teams will ask harder questions during eval cycles. Expect longer deal cycles and additional compliance requirements in contracts. **Competitive positioning**: If you are selling against ChatGPT Enterprise or using it internally, this creates an opening. Competitors with cleaner regulatory records can lean into compliance and safety positioning. If you are an AE selling Anthropic, Google, or other alternatives, this lawsuit is relevant context for your discovery calls. **Internal usage policy**: Sales teams using ChatGPT for email drafts, research, or prospecting need clear company policy. Legal and IT will be reviewing what data goes into these tools and what comes out. Florida alleges OpenAI prioritised speed and commercial gain over safety. Whether that claim holds up in court, the message to enterprise buyers is clear: regulatory pressure on AI vendors is real and escalating. Factor that into your tech stack decisions and procurement conversations. OpenAI remains one of the dominant AI providers globally, competing with Anthropic, Google DeepMind, Meta, and xAI. The company has not disclosed revenue or headcount, but ChatGPT has mass-market reach. For ANZ enterprise buyers, this is another data point in the ongoing conversation about AI vendor risk and compliance requirements.

SmartCompany
3 months ago
News

SaaStr runs 4 AI SDR agents, not one: Artisan, Qualified, Agentforce, Monaco

## The Stack SaaStr is running four AI SDR agents in production. Not one platform doing everything. Four vendors, four contracts, four different jobs. **Artisan for outbound.** Three instances running parallel. Different personas, different ICPs. 40,000+ messages sent. Does cold outbound at scale, does it well. **Qualified for inbound.** 100,000+ sessions processed. Over $1M closed through the inbound motion. When someone hits the site, Qualified qualifies, books, routes. Purpose-built for that moment. **Agentforce for Salesforce-native reactivation.** Close to 200,000 messages sent. 72% open rates on win-back campaigns. Lives inside Salesforce, has every data point already. No integration mess. **Monaco for net new logos.** Different ICP data, different outreach pattern than farming the install base. Four agents. Four admin consoles. Jason Lemkin (SaaStr founder, ex-EchoSign) says they would do it again. ## The Case for Specialised All-in-one platforms promise simplicity. Reality: mediocrity across every function. Each motion (outbound prospecting, inbound qualification, reactivation, new logo land) has a different shape. Different data, different success metrics, different optimal behaviours. Outbound is volume with tight personalisation. Inbound is speed with intent signal interpretation. Reactivation is context with deep CRM history. New logo is research with ICP precision. No platform today does all four at A+ quality. A platform that is 60% good at all four costs you more than four platforms that are 95% good at their one thing. The cost shows up in pipeline, conversion rates, closed-won dollars. ## The Real Cost Running four platforms costs more than one. Four contracts, four minimum commits, four implementation cycles. Real dollars, not just work. You pay in operational surface area: four sets of credentials, four data flows, four sets of prompts to optimise, four vendor relationships, four security reviews. You need headcount. At least one person, ideally two, dedicated to managing deployment. AI agents are not zero-headcount tools. One person cannot hold tribal knowledge of how four platforms are configured. Why pay it? Because the quality gap between specialised and generalist AI agents is big enough that the extra cost pays for itself in pipeline. Difference between $1M closed and $300K closed. Difference between 72% open rates and 22%. ## If You Are Starting Out Most companies do not need four AI SDR agents to start. For 90%+ of B2B companies deploying their first AI SDR, one vendor can make material gains. Two at most: one for outbound, one for inbound. SaaStr runs four because they have pushed each motion into highly segmented territory after 10+ months of production. Roughly 100 effective segments across 1,000 contacts at a time. At that level of segmentation, no single platform handles all of it well. If you are standing up your first AI SDR, do not start here. Pick the one platform that covers your biggest motion and go deep. Prove that motion works. Then layer in a second tool when the first is producing real pipeline. Starting with four will get you four mediocre deployments instead of one great one. The tool matters far less than the strategy you bring to it. ## What This Means The AI SDR market is splitting. All-in-one platforms for teams getting started. Specialised tools for teams at scale with segmented motions. If you are evaluating AI SDR tools, the question is not which single platform wins. The question is: what is your biggest motion, and which tool does that one thing best? Once that motion is working, you can layer. Until then, the complexity of managing multiple agents is not worth it. The specialisation advantage only kicks in once you have outgrown the generalist answer.

SaaStr
Load More