Afleveringen
-
In our latest episode, we talk with Anshul Gupta, who recently raised $22.5M from Bain Capital and First Round Capital for Actively.
Our favorite takeaways:
* Personalization at scale is actually "laziness at scale": "The amount of emails I got saying 'Anshul, congrats on your recent funding round' followed by a totally unrelated value proposition... if you're looking at thin logic like 'someone raised funding, throw them in a sequence,' that's not working anymore."
* Top performers do less, not more: "When you sit down with the best reps, they're not the ones doing the most volume. If you look at the leaderboard, they may even be in the bottom half from an activity perspective. What we see is they're incredibly targeted and laser-focused."
* Relevance beats volume: The alternative to personalization is "relevance at scale" – going incredibly deep with research, building proper hypotheses based on customer needs, and focusing on quality over quantity.
* AI flips the paradigm: "The revenue orgs in two years will be flipped on their head. Systems of intelligence will, instead of humans asking them what to do, be processing all information and guiding humans where to spend their time."
* AISDRs have limited applications: "If you're a company just starting up with no sales reps and want to get meetings by hook or by crook, AISDRs could work. But fundamentally upmarket, in complex ecosystems with multi-product motions, that model doesn't work."
* The revenue frontier concept: "Rather than constraining headcount and increasing productivity, increase your ambition. Companies are nowhere near their revenue frontier, and imperfect go-to-market execution is a huge contributor to that gap."
* Custom reasoning models: "Ramp has their own Actively reasoning model, Ironclad has their own – trained on the hyper-specific nuances of their products, go-to-market motions, and as those continue to evolve."
What makes Actively different:
* Their focus is increasing rep productivity and the quality of pipeline each human can generate
* They train custom reasoning models specific to each client's business
* They deploy a "forward deployed engineering model" similar to Palantir to customize for enterprise needs
* The system continues to improve through active learning (hence the name!)
What's not working in the market:
"All these volume-based solutions purport to help you send more emails or do more calls. It's all about chasing volume when the top performers are actually doing less but with more focus. That thin logic of 'if someone raises funding, throw them into a sequence' is what's corrupted the concept of personalization at scale."
What do you think about Anshul's take on personalization vs. relevance? Are you seeing "laziness at scale" in your inbox? Have you identified the revenue frontier for your business? 👇
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit agenttalk.substack.com -
In a candid conversation at Paid’s launch party, EQT Ventures investor Doreen Huber shared her razor-sharp perspective on what’s working in AI agent investing, why traditional SaaS is losing ground, and what founders need to bring to the table to secure funding in this new era.
Our favorite takeaways:
* SaaS is no longer the focus to the B2B Software team at EQT Ventures:"My team is not investing in traditional SaaS anymore. Our strategy is to go for agentic, AI-native companies, and we tend to disqualify what doesn’t fit that bucket."
* True agents only:"We only want to support companies doing something end-to-end—not just enhancing customer care with AI-drafted emails. We’re looking for agents that do the actual work from start to finish."
* Commercial DNA matters:"I definitely have a thing for founders with commercial DNA. If someone comes from an engineering side, they absolutely need to learn this... the best CEO is also the best product person."
* Founder qualities:"I personally love the outliers, the underdogs, or someone with a crazy CV. I'm not into the typical business school, textbook founder. I love it when someone shows up with an edge."
* Legacy SaaS is under pressure:"Many legacy SaaS companies will lose market share to agentic players. A lot of them are struggling—they don’t have the AI talent, and they’re stuck in outdated stacks."
* On industry hype:"Some of the big players are slapping AI labels onto old products. That’s not agentic innovation. That’s legacy software trying to catch up."
What Doreen is looking for now:
* Enterprise-ready agentic sales and marketing solutions – not just slim use cases, but holistic systems
* Agentic cybersecurity – solving modern threats with AI-native architecture
* Vertical AI applications – especially where AI is applied to labor, not just software budgets
What’s working:
“Most companies moving faster than others have that AI-native mindset. They want lean teams and ask: ‘Can we do this with agents instead?’”
What’s not working:
“BDR email sequencing or scheduling tools... they look impressive at first, but in reality, these problems won’t exist in a year or two. That’s just a GPT anyone can build.”
What are your thoughts on Doreen's take that traditional SaaS is finished? Do you agree that legacy companies can't catch up with AI-native startups? 👇
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit agenttalk.substack.com -
Zijn er afleveringen die ontbreken?
-
In our recent conversation with Amos Bar-Joseph, CEO and co-founder of Swan AI, he shared his radical vision for the future of business: autonomous companies achieving $10M ARR per employee through strategic AI implementation. Beyond the typical AI hype, Amos details a practical roadmap for companies looking to scale with minimal headcount.
Here are the key insights from our discussion:
The autonomous business revolution
* The next 5 years belong to SMBs: "We believe that the next five years are the year of the SMB, the year of the small business, small lean teams that are showing the world that using AI agents, you can reach massive scale, like never seen before."
* Rethinking business fundamentals: "It's about reimagining how humans and AI collaborate together, rethinking fundamentally the operating system of a business."
* The Swan Metric: Targeting $1M+ ARR per employee (with Swan pushing toward $10M) through strategic AI implementation.
Breaking the "throw bodies at problems" mindset
Amos said something that we think many founders have thought about: "The first instinct that the old playbook got you to do was throw bodies at the problem, right?".
But Amos does things a bit differently, and we like that a lot:
* Self-imposed constraints drive innovation: At Swan, they created the constraint that they "can't throw bodies at a problem" - forcing them to find more creative and intelligent solutions.
* “Ops Wizards”, not just more headcount: "Every team should have that ops wizard" who can bridge technical understanding with business orientation.
* Human-in-the-loop design: Always start with humans supervising AI and providing feedback to create "a self-learning system that takes your knowledge in a collaborative way."
The future of sales is human + AI, not AI replacing humans
When I asked about which parts of the sales cycle will be replaced by AI, Amos offered nuanced insight:
* It depends on ACV: "The higher the ACV, the tougher it is to replace any part of the sales cycle... When you look at $20 million deals, then you want a human in the loop."
* Low ACV should be marketing-driven: For $19/seat products, "from a unit economics standpoint, it doesn't make sense to do outbound."
* Sellers love winning, not prospecting: "Sellers love winning. And for that 1% that they are winning, they love that notion... It's the best moment of their day when they get that yes on the screen."
* The 100X seller: "The future of sales is first of all reimagining how sellers work with human and AI collaboration at the core. And it's more about finding the path to the 100X seller."
The three types of businesses in the AI revolution
According to Amos, businesses will fall into three categories across a spectrum:
* Biggest losers: "Those trying to automate their workforce, replace their workforce with digital workers. And they will be left behind."
* Partial winners: "Implementing AI agent tooling across their entire stack... but disparate solutions for different functions."
* True unicorns: "Lean teams that are building AI agents as their product, able to rally their entire agentic workforce around that specific agent."
Practical advice for outbound sales:
Amos had some pretty practical advice too: Move from “digital workers" to storytelling engines.
How?
* Goodbye sequences, hello relationships: "We never pitch with a hard CTA. Never... It's only about, look, something happened in your business. There's an event that is relevant to your day-to-day as a VP of sales."
* Non-deterministic outreach: "Swan looks at a timeline always, looks at what happened before, what is happening now, and will create a recommendation how to engage that lead."
* Monitoring relationships at scale: "What we're envisioning is a future where SWAN can monitor relationships, top of the funnel relationships with thousands of accounts in parallel."
What's your take? Are you seeing companies in your industry successfully implementing AI to scale without proportional headcount growth? 👇
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit agenttalk.substack.com -
In a recent deep dive with Sequoia Capital partner Pat Grady, he shared surprising insights about what separates winning AI companies from the rest, and challenged conventional wisdom about AI moats, pricing models, and what investors truly value.
Our favorite takeaways:
* Building AI companies is just building a company. It’s 95% the same and people problems still dominate
* Trust is the critical design pattern most AI companies miss. Users need to see how you arrive at your results
* Most AI products achieve 80% functionality quickly, but the final 20% takes 5-10x longer and is what builds actual trust.
* The greatest moat in AI isn't data or tech - it's founders with relentless execution.
Pat also added some extra wisdom that we appreciate:
* The "data flywheel" appears in 100% of AI pitches but only 1% of companies actually demonstrate it works - Pat demands evidence, not theory
* AI pricing will standardize around outcome-based models with huge variation - the most successful companies think about both "input" (work done) and "output" (value created)
* For investors, negative gross margins are acceptable in early AI companies because token costs are dropping 99% and multi-tenancy is becoming more accepted
* Domain-specific AI products that build real trust can carve out defensible positions against foundation model providers in vertical markets
* The most successful AI companies avoid "vibe revenue" (temporary excitement) by focusing on engagement and retention using consumer internet metrics even for B2B products
What's your experience with AI products and pricing models? Have you found user trust to be the limiting factor? Share your thoughts below 👇
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit agenttalk.substack.com -
Wade Foster, co-founder and CEO of Zapier, shares how they completely revamped their pricing model - and actually REDUCED revenue in the short term - to drive explosive long-term growth.
Since launching in 2011, Zapier has grown to help millions connect their apps without code, reaching a valuation over $5B with minimal VC funding. Their recent pricing change offers fascinating insights for SaaS founders.
Key takeaways:
* Zapier eliminated their dual-metric pricing (Zaps + Tasks) to simplify the customer experience
* They made unlimited Zaps available on all plans - removing a major friction point
* Every plan now includes pay-as-you-go options beyond the base subscription
* They stopped counting utility features as billable tasks - providing more value
The result? Short-term revenue dropped significantly, but task consumption and customer happiness soared. This bet on long-term growth would have been impossible for most VC-backed companies.
More insights:
* Being profitable and bootstrap-focused gave them freedom to make radical customer-first decisions that sacrificed short-term revenue
* Pricing "debt" accumulates over time when you experiment with different models - eventually requiring a reset to first principles
* The best pricing aims for customer "love" - Wade literally used this word - not just reluctant payment
* Competition constantly forces pricing innovation - especially with direct competitors who counter-position against market leaders
* AI is rapidly democratizing entrepreneurship - Wade sees teams of 10-20 people reaching millions in revenue with minimal engineering, driven by domain expertise paired with no-code tools
What pricing changes have transformed your business? Share your experience below 👇
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit agenttalk.substack.com -
Manny hosts Mickey Haslavsky of Enso, who's flipping the script on AI agents in the most refreshing way.
Mickey says their mission at Enso is to reduce the prices of services, just in general, for small businesses.
While lots of AI companies chase enterprise dollars or build complex multi-agent systems, Mickey's team is printing money with a simple contrarian approach: deliver "good enough" AI agents to small businesses at a flat $49/month.
Some key insights from our conversation:
* "SaaS is pretty much dead... some of the agents that we're building, you would have to raise like a seed round like three years ago, like raise like $5 million and have a team of like 10 engineers to build them."
* "I don't buy this story of losing jobs. What's really important is reaching those who couldn't afford an agency before, who couldn't pay $3,000 for SEO or $5,000 for social media design."
* "SMBs don't churn because they go out of business - they churn because they pivot constantly. That's why our unlimited access model works."
Mickey's belief that LLMs will become completely commoditized, allowing companies like his to use cheaper models from anywhere (including Deepseek) to deliver services that previously required expensive agencies.
The future might not belong to the fanciest AI, but to those who make it accessible, affordable, and actually useful for the millions of SMBs who just need to get things done.
What do you think? Is this the true democratization of AI, or are there limits to the "good enough" approach?
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit agenttalk.substack.com -
Rob Litterst (HubSpot exec & SaaS pricing guru) just dropped some mind-blowing insights about how AI is completely disrupting traditional software business models. Whether you're building AI products or just trying to stay ahead of the curve, this episode is packed with spicy takes on where the industry is headed.
Three explosive themes that'll blow your mind:
* The Death of Per-Seat Pricing: The old way of charging per user is getting absolutely demolished by AI agents. Rob breaks down how companies are being forced to evolve or die as AI native startups come gunning for their lunch. Traditional SaaS companies are scrambling to figure out how to price their products when algorithms are replacing humans left and right.
* The Human Premium Paradox: Humans are becoming a premium feature in software. While AI handles the grunt work, companies are starting to position human expertise as the ultra-premium tier. But this creates an insane challenge: how do people gain expertise if AI is handling all the entry-level work? Rob dives deep into this existential crisis facing professional services.
* The Revenue Bloodbath is Coming: Companies that don't adapt their pricing models are about to get absolutely wrecked. Rob shares war stories about how AI native startups are growing to $100M+ ARR with tiny teams, while traditional players are stuck in old pricing models that don't match how value is created anymore. The race is on to figure out outcome-based pricing before it's too late.
This conversation gets into the nitty-gritty of how companies are navigating this transition - from baby steps toward outcome-based pricing to full-blown business model transformation. If you want to understand how AI is reshaping the fundamentals of software business models, this episode is a must-listen.
Whether you're a founder, product leader, or just fascinated by how AI is changing the game, Rob brings deep insights from both the trenches at HubSpot and his broad view across the industry. Don't miss this one - the future of software pricing is being written right now.
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit agenttalk.substack.com -
"35% of projects fail, and the numbers haven't budged in 20 years."
That's what kept Shawn up at night before founding Coworked. After two decades in project management, he'd seen every methodology, tool, and framework come and go - yet the fundamental problems remained stubbornly unchanged.
Shawn is building an AI project manager that could transform how work gets done. But not by replacing humans - by unleashing them.
Harmony, Coworked's AI agent is rewriting the rules of project management. While most AI companies are focused on chatbots and automation, Coworked is tackling something far more ambitious: creating an AI teammate that handles everything from resource allocation to risk analysis.
What's fascinating is how they've navigated enterprise adoption. When they first pitched to project managers, they hit a wall of resistance.
But then they had an insight - they were talking to the wrong people.
PMO leaders immediately got it. Not as a way to reduce headcount, but as a solution to a persistent problem: how to handle 2x the projects without 2x the team.
Thanks for reading Agent Talk! This post is public so feel free to share it.
Project managers spend up to 80% of their time on process tasks rather than strategic work. As Shawn puts it:
"There's a subtle but crucial difference between replacing a job and replacing a role."
Harmony handles the routine while enabling humans to focus on what matters.
Their early results are turning heads. Fresh out of Techstars, they're already landing enterprise clients who see the potential to transform their project delivery.
Perhaps the most interesting part is their insight about AI adoption: success comes not from elimination, but elevation. It's not about replacing project managers - it's about making them superhuman.
AI transformation isn't just about automation - it's about reimagining how work gets done. And sometimes, the biggest breakthroughs come not from replacing humans, but from freeing them to do what they do best.
Just as we saw with Zendesk revolutionizing customer support, Coworked is showing us what's possible when we think bigger about AI's role in enterprise transformation.
The question isn't whether AI will transform project management - it's how fast organizations will adapt to this new reality.
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit agenttalk.substack.com -
Zendesk's journey isn't just another AI adoption story - it's a masterclass in how even enterprise companies can reinvent themselves in the age of AI.
While others rushed to add AI features, Zendesk took the bold step of completely reimagining their business model.
What's fascinating is how they navigated the classic innovator's dilemma: "It's one thing to understand it, it's another thing to be in it and live it," as Kelly puts it.
They didn't just add AI - they fundamentally changed how they deliver and capture value.
Think about their evolution:
* Started investing in AI years before ChatGPT
* Built global AI teams across three continents
* Developed models trained on 18 billion support tickets
* Shifted from seat-based to outcome-based pricing
While most SaaS companies are stuck in the "more seats = more revenue" mindset, Zendesk recognized a fundamental truth:
"You've got these dueling factors... seat contraction and AI arriving on the scene."
Their response? Align their success directly with customer outcomes.
It’s interesting to note that even their private equity owners, typically viewed as cost-cutters, became their biggest champions for transformation. As Kelly says
"We've got investors that are actually pushing us to go faster and to spend a lot of time on this very topic versus playing defensively."
Zendesk’s approach shows that successful AI transformation isn't about following trends - it's about solving real business problems at scale - just like with last week’s Synthesia episode.
Instead of charging for tokens or interactions, they focused on what customers actually care about: resolved support errands and tickets.
Notice how Zendesk is being a pro at handling the transition:
* starting with clear success metrics
* offering predictable pricing through commits
* and finally building trust through incremental value delivery.
This story isn't just about AI - it's about how enterprise software evolves when technology enables fundamentally new business models.
The question isn't whether to transform, but how to do it while keeping your customers' trust.
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit agenttalk.substack.com -
Synthesia's success wasn't about chasing the latest AI trends, but about solving real business problems with a relentless focus on enterprise-grade features and workflows.
In this first episode, Manny talks to Victor Riparbelli, founder and CEO of Synthesia that raised $180M in Series D.
What's fascinating is how Synthesia navigated the classic enterprise software challenge: balancing the "wow factor" with sustainable business value.
While most AI companies get caught in the "sugar high" of viral demos, which is why we’re seeing massive churn problems right now. Synthesia on the other hand focused on the unsexy but crucial elements: ISO certification, enterprise security, and seamless workflows. As Victor puts it, "These boring things unlock bigger and bigger deal sizes."
Think about their evolution:
* Started with avatars that grabbed attention
* Built out comprehensive video editing capabilities
* Added enterprise-grade security and collaboration
* Developed deep integration with business workflows
The results speak volumes: Their customers aren't just trying AI - they're fundamentally transforming their video production.
As Victor says, "For the same budget of producing two traditional videos, we can now make 400 AI-powered ones."
Watch how they're shifting from charging per video creation to focusing on video consumption and engagement. It's a masterclass in aligning pricing with actual business value.
Synthesia's journey shows that building for lasting impact beats chasing the latest AI trends.
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit agenttalk.substack.com
