Every vendor deploying AI sales agents faces the same problem: not enough Forward Deployed Engineers to make the tech actually work inside customer environments. Traditional CS teams cannot fill the gap, and the agents cannot deploy themselves. Here is what that means for sales organisations hiring in 2026.
Australia's $335B sovereign wealth fund is reviewing 10 investment and ops roles after deploying new data systems. Expec...
Funding & StartupsSydney quantum sensing startup Deteqt closed a $5 million seed round led by Main Sequence to scale diamond chip manufact...
SaaStr's Jason Lemkin hasn't opened Canva in months despite being a happy customer for 8 years. The pattern: specialist AI tools like Reve, Opus Pro, and Higgsfield are eating specific use cases while team usage masks power user churn. When your most engaged champions quietly disengage, that is a leading indicator every B2B vendor should track.
The SaaStr.ai Index tracking 25 leading B2B software companies hit a 50.5% decline from October 2025 to April 2026. This is not a correction. The market has structurally re-rated software as an asset class, now trading below S&P 500 multiples for the first time ever. Two forces: AI capex displacing traditional software budgets, and substitution fear that agents will replace seat-based revenue models.
Jason Lemkin's SaaStr replaced a 20-person GTM team with AI agents managed by 3 humans, driving revenue from -19% to +47% YoY. Most implementations fail because teams treat AI like set-and-forget software, not sales reps that need daily coaching. Here is what actually breaks.
## The Bottleneck No One Saw Coming Every company rolling out AI agents at scale is running into the same problem: forward deployed engineers are impossible to hire. The shortage is not just a hiring challenge. It is a structural issue about what it actually takes to get AI working inside a real enterprise. FDEs sit at the intersection of product, engineering, and customer success. They go on-site, understand actual workflows, and configure the product to work inside those workflows. They are not building from scratch. They are not doing basic support. They are doing the hard middle work of making software actually land in the real world. Palantir built their entire go-to-market around this model. You needed their people inside your organisation, configuring and training the system for your specific context. That model worked. It was expensive and it did not scale the way SaaS was supposed to scale. But it worked, because complex software in complex environments requires human judgement to deploy well. Now almost every serious AI product has the same requirement. And almost no one has enough people who can do it. ## Why CS Cannot Fill This Gap The instinct at most companies: solve this with customer success. CS is already post-sale, already focused on adoption. Just upskill them, right? Wrong. Traditional CS was built for a different era. The job was: help customers use software they have already decided to buy, make sure they hit their renewal metrics, escalate bugs. It was reactive, relationship-driven, and optimised for retention. FDE work is different in almost every way. It is proactive, technical, and optimised for deployment. You are going in before the problem exists and configuring the system so the problem never happens. The skill set required is closer to a solutions engineer or a junior product manager with strong customer empathy than it is to a traditional CS rep. Most CS teams do not have it. Retraining takes longer than most companies want to admit. ## Agents Cannot Deploy Themselves. Yet. The whole premise of AI agents is that they automate work. But deploying an AI agent is itself significant work, and it is work the agent cannot do for you. Not yet. Someone has to understand the customer's workflows deeply enough to know where the agent fits. Someone has to train the agent on the right data, the right context, the right edge cases. Someone has to test it, catch where it breaks, and iterate. Someone has to get internal buy-in from the people whose jobs will change when the agent goes live. That is FDE work. And it is manual, high-judgement, human work. Palantir announced recently that they have gotten deployment times down over 90% using forward deployed engineers. That is remarkable. It also means the best-in-class operator in this model is still deploying manually, just faster. 90% reduction in deployment time is not the same as automating deployment. The human is still in the loop. ## What This Means for Sales Teams Every serious AI vendor is now competing for the same small pool of people who can do this work. The companies that came up through Palantir, the solutions engineers from the major cloud platforms, the implementation consultants from the enterprise software world: everyone wants them, and there are not enough of them. Meanwhile the demand is exploding. Every enterprise that decides to deploy AI agents needs FDE-calibre people to make it work. For sales teams, this creates a few realities: **Longer sales cycles.** If you cannot deploy the product, you cannot prove value. If you cannot prove value, the deal stalls. **Higher implementation costs.** Companies are paying premium rates for FDEs. That cost gets passed somewhere, usually to the customer or to margin. **New comp structures.** Some vendors are tying commission to successful deployment, not just closed deals. If the product does not land, the rep does not get paid out fully. The companies winning at deployment are building serious enablement programmes: not just documentation, but hands-on training that gives customer-side operators the skills to configure and train agents themselves. If you are selling AI tools, ask your leadership what the deployment plan actually looks like. If the answer is "CS will handle it," the answer is wrong.
## Future Fund cutting 10 roles, banking $15m from tech automation The Future Fund is reviewing 10 roles across investment and operations teams after investing in data systems and automation. The cuts follow a tech overhaul that CEO Raphael Arndt says will save $10-15 million in FY2026/27. The $335 billion sovereign wealth fund, which manages public sector super liabilities, expects the technology investment to shave 5-7% off operating costs next year. Further savings are projected for subsequent years. "We're baking in the benefits and maximising the efficiencies of our technology overhaul," Arndt said. The fund has around 140-150 total headcount, concentrated in Melbourne and Sydney. The role reviews cover both investment professionals and support functions. The fund said it is consulting with affected staff before finalising decisions. ### Finance sector automation trend The Future Fund joins a growing list of financial institutions cutting roles after deploying AI and automation. Bendigo and Adelaide Bank announced hundreds of job cuts in April after signing two technology deals. Unlike traditional B2B companies, the Future Fund operates without sales teams or commercial revenue targets. It invests globally in equities, fixed income, property, and alternatives. The fund does not have a CRO or VP Sales: its structure centers on portfolio managers and investment analysts. Arndt said the tech investment has been "critical to investment performance" and positions the fund for what he calls a "new investment order" reshaping markets. The savings come from improved data systems and renegotiated external service contracts. Costs and staffing remain "appropriate for the scale and complexity" of the fund's mandate, Arndt said, but the organisation will continue assessing resource needs. Worth noting: this is a government entity with no traditional sales function, but the automation trend mirrors what commercial enterprises are doing. When large, well-funded organisations start cutting roles to automation, the rest of the market usually follows.
Deteqt, a University of Sydney spinout, closed a $5 million seed round today. Main Sequence led, with ATP Fund, BOKA Capital, Beaten Zone Venture Partners, Uniseed, and the university participating. The company builds chip-scale quantum magnetometers using diamond-on-silicon tech. Target markets: GPS-denied navigation for defense (drones, submarines), autonomous vehicles, and potentially portable MRI. They already have an Australian Defence Force contract. Founded in 2025 by CEO Dr. Jim Rabeau and Professor Omid Kavehei, with Rupal Ismin as COO. This follows a $750k pre-seed in March 2025. ## What This Means for Sales No sales team details disclosed. No CRO, no VP Sales, no AE count. This is pre-revenue, pre-GTM team. Defense tech sales roles at quantum startups typically look different from SaaS: - Longer sales cycles (12-24 months for defense contracts) - Heavy on government procurement experience - Equity compensation often outweighs base (early-stage defense tech) - Remote roles rare due to security clearance requirements For context: aerospace and defense contractor sales roles in ANZ are heating up. Quantum sensing sits at the intersection of deep tech and defense, a niche but growing market. Companies like Infleqtion (US-based quantum tech) offer remote roles, but most defense-focused positions require on-site presence. ## The Numbers Total raised: $5.75M ($5M seed + $750k pre-seed). Funds go to product development, diamond chip manufacturing scale-up, and team growth. No revenue disclosed. No current sales headcount disclosed. Deteqt is Sydney-based, targeting Australia-US-UK investor and customer networks. Named a 2025 InnovationAus Awards finalist in Defence and Dual Use. ## Bottom Line Early-stage deep tech with defense applications. If they build a sales team in the next 12 months, expect equity-heavy comp and a focus on government procurement experience. Not hiring yet, but worth tracking if you are in defense tech sales.
## The Numbers Look Great. The Usage Pattern Does Not. Canva went from $23M ARR in 2018 to $4B at end of 2025. That is 173x growth in seven years. They have 265 million monthly active users, 31 million paid subscribers, and their B2B segment alone is $500M ARR, doubling year over year. By every traditional metric, they are crushing it. Profitable for eight consecutive years. Sydney-based, built a global design platform that competes with Adobe. But here is the problem: power users are going quiet. ## What Stealth Churn Actually Looks Like No single competitor replaced Canva. Instead, specialty AI tools are eating individual use cases. Reve handles thumbnails. Opus Pro cuts video clips. Higgsfield does short-form video. Each tool does one thing better than Canva's all-in-one approach. The power user who drove the original purchase, who would have championed expansion, stops logging in. The team still uses Canva for social graphics and event collateral. Usage metrics look fine. NPS stays high. The account renews. But the person who would fight for budget at renewal just checked out. Quietly. Without even noticing. ## Why This Matters for B2B Sales When your topline is growing 100% year over year, you cannot see this pattern in your numbers. New revenue masks quiet disengagement at the edges. By the time it shows up in retention metrics, you have lost 12 to 18 months of leading indicators. Your most engaged customers are exactly the ones most likely to discover purpose-built AI alternatives. They care about output quality. They are early adopters. They are your expansion revenue. Your casual users stick around because switching costs still matter and $12 per month is not worth the effort to cancel. But inertia-based retention is the worst kind of retention. It means your product became a rounding error in someone's budget. Not essential. Just cheap enough to ignore. ## The Category Risk This is not a Canva problem. Canva will probably be fine. $4B ARR, strong execution, massive distribution. But every horizontal B2B tool faces this dynamic right now. When AI tools can do one specific job better than your all-in-one platform, power users will find them. Your metrics will not warn you until it is too late. Worth asking: who are your power users, and what are they actually using right now?
# Public Software Stocks Down 50% in Six Months The SaaStr.ai Index of the top 25 public B2B software companies hit a 50.5% decline over six months, from October 2025 to April 2026. Half the market cap, gone. This is not a 20% correction. This is a structural re-rating of software as an asset class. For the first time ever, public software companies trade at a P/E discount to the S&P 500. Not at parity. Below. Forward P/E multiples for application software collapsed from 84x in 2021 to 22.7x today. The market's implied long-term growth rate for public SaaS dropped from 4.7% three months ago to 1.1% now. The market is saying: software is no longer a premium business. ## What This Means for Sales Teams Two forces are hitting simultaneously: **Budget displacement.** When Anthropic hits $19B in annualised run rate, growing $6B in a single month, that spend comes from somewhere. Approximately 75% of new hyperscaler infrastructure spending in 2026, over $450 billion, targets AI infrastructure. That money used to buy Salesforce seats, ServiceNow modules, HubSpot licenses. Not anymore. **Substitution fear.** AI agents might replace seats instead of complementing them. Seat-based revenue models depend on headcount growth. If agents replace headcount, the model reverses. Investors are pricing this into terminal value, which explains why current earnings do not explain the decline magnitude. Category leaders are getting crushed: - **Atlassian (TEAM):** Down 57.91% in the recent quarter, 67.84% over the past year. Founded in Sydney in 2002, the company that built Jira for every engineering org is now 71.80% below its May 2025 high. - **HubSpot (HUBS):** Down 50%+ over the past year. $2B+ ARR, one of the best go-to-market motions in B2B history. - **Salesforce (CRM):** Down 30%+ in Q1. The defining CRM platform of the last 20 years. - **ServiceNow (NOW):** Down 30%+ in Q1, despite actually accelerating. - **Adobe (ADBE):** Down from $638 to under $350. These are not speculative bets. These are cash-generating, deeply embedded businesses. The market is treating them like they face existential risk. ## ANZ Context Atlassian maintains ~500-1,000 headcount across Sydney HQ and Auckland offices, focusing on enterprise sales to government and finance sectors. The company reports no major 2026 hires or cuts amid the downturn, relying on inbound and partner-led models over large direct sales teams. Sales organisation emphasises efficiency under President Anutthara Gose, promoted in 2024. Broader SaaS firms report stalled net retention at ~90% gross as AI shifts divert budgets. Sales teams are feeling this in quota design, comp structures, and territory planning. When your product category loses half its market cap in six months, quota relief conversations get harder. ## What Is Holding Up Not everything is down equally. Companies performing better share characteristics: **Palantir (PLTR):** +135% in 2025, cooling now but still outperforming. Rule of 40 score hit 127 in Q4 2025. Revenue growth at 70% YoY. They are not a seat-based SaaS vendor. **Cloudflare (NET):** Guided 2026 at $2.79B revenue, 28-29% growth. AI agents generate an order of magnitude more outbound requests than user-driven apps. All of that flows through Cloudflare infrastructure. **DigitalOcean (DOCN):** Up nearly 50% YTD in an index down 50%. Simpler stack, smaller companies. The bifurcation is clear: infrastructure that enables AI outperforms applications that AI might replace. For sales professionals, this matters. Comp packages are tied to equity. Territory planning assumes growth. Quota is built on market assumptions. When the market re-rates your entire sector by 50% in six months, every one of those assumptions changes. Worth noting: if you are carrying a bag at a company down 50%, your equity-based OTE just got a lot less attractive.
## The Numbers SaaStr deployed 20+ AI agents across SDR, support, and GTM functions. Headcount dropped from 20+ employees to 3 humans. Revenue growth shifted from negative 19% to positive 47% year over year. AI-attributed pipeline hit $4.8M, closed-won $2.4M. Outbound volume reached 60,000+ personalised emails with 5 to 7% response rates against an industry average of 2 to 4%. Jason Lemkin, SaaStr founder and former EchoSign co-founder (acquired by Adobe, scaled to $100M+ ARR), shared why most AI agent implementations fail. The pattern: teams deploy AI SDRs, check back in two weeks, see 0.02% conversion, and blame the tool. ## What Actually Works SaaStr's Chief AI Officer spends 30% of her time managing agents. The team invests 60 to 90 minutes daily: reviewing output, refining prompts, QA-ing emails, iterating on what converts. Lemkin is clear: give it 90 days of this discipline before drawing conclusions. The failures follow predictable patterns. Teams deploy AI on top of broken processes (bad messaging does not improve at 10x speed). CRM data quality collapses under AI load (agents need clean data, humans can work around it). Segmentation stays lazy (generic outreach scales into generic spam). SaaStr runs 15+ email variants across segments: company stage (seed through public), role (CEO, CRO, VP Sales), past event engagement, vertical, deal size. A Series A CEO who has never attended SaaStr gets different outreach than a returning enterprise sponsor's VP of Marketing. ## The AI Implementation Failure Rate Most AI projects fail because teams treat agents like software: deploy once, let it run. That works for static tools. It does not work for AI that needs training, feedback loops, and constant refinement. The companies succeeding with AI agents are managing them like junior reps, just with higher volume potential. Lemkin's advice: audit your CRM before deployment, fix your messaging with humans first, budget real time for daily management, and segment radically. If your best human SDR cannot book meetings with current messaging, your AI SDR will not either. AI scales what works. It does not fix what is broken. Worth noting: SaaStr's results came after months of painful iteration. The 47% revenue growth and $4.8M pipeline did not happen in week one. Most teams churn before they get there.
Microsoft's hiring freeze isn't a surprise—it's the inevitable end of tech's decade-long talent hoarding mistake.
When the poster child of Australian tech blames AI for cutting 10% of staff, it's not about the technology—it's about finally admitting they overhired during the free money years.
Woolworths cops $1B in underpayments, Optus pays $100M in penalties, and Sydney's top real estate agent gets banned—your SaaS comp plan is next.