Bain Capital Ventures

My first job after college was in customer support at Google. Back then, the Google ad products had started to take off, and as search engine marketing became a significant part of marketers’ budgets, a large group of young customer support folks like me were necessary to assist Google’s huge customer base as it adopted this new and at times confusing technology. Many of my best friends and former colleagues got their starts in tech in this cohort of customer support workers - Lane Shackleton, Andrew Stinger, and Alan Ringvald, among others. We started on email and chat support, and would respond painstakingly to hundreds of messages a week. When we moved on to phones, I lost my voice several times during our eight-hour shifts. It was hard work at times, but also great training in working with customers - how to handle people who are busy, or stressed, or frustrated, and leave them with a better perception of the product and brand than when they started. My lived experience with customer experience is one of the reasons I’m so excited for our partnership with Decagon, a leader in AI for customer support, which has earned rave reviews from discerning teams with complex products at Rippling, Duolingo, Faire, and many others. Decagon can “deflect” a huge proportion (for some customers, 90%!) of conversations that would otherwise go to a human. This frees up time for human customer support folks to deal with more complex, and *human* issues, that improve customer satisfaction and brand perception. Jesse Zhang and Ashwin Sreenivas are building a company that uses tremendous technology to greatly improve the work of support teams, and the customers they serve. They are hiring up a storm (in sales, engineering, marketing, product). On the GTM side, Evan Cassidy is building a high-performing, fun team of sales executives to bring this new technology to market amidst the sea change of AI. I think Decagon will be one of the generational AI companies for a problem that almost every business faces, and which has a real impact on the lives of people. (How many times have you been frustrated when dealing with customer support?) Let me know if you’d like an introduction to this team or would like to learn more about using Decagon to improve your company's support processes.

When I joined Bain Capital Ventures, the first question I received from a founder was, “How do I go hire a great People leader?” That question has many layers: What’s the right profile? How do I decide what attributes or experience to prioritize? How do I know if they’re good or not? And of course… how do I not mess this up?! Working with Laura Hilton, we finally decided to take our back-of-the-napkin thoughts and turn it into something more user-friendly. I’ve lived this role day-in and day-out, Laura has hired for this role multiple times over. So today, we launch BCV’s Field Guide to People Leaders in hopes that our experiences may help founders look at this challenge of hiring a great People leader in a slightly different light. 💡 https://lnkd.in/gH34HJFz

🔊 AI is making services sexy again. 🔊 The core promise of AI is the elimination of barriers to creation. Text, audio, video, code... all become easier to generate. When creation becomes that simple, it doesn’t just empower—it commoditizes. What once required deep expertise and high barriers to entry can now be built by many, intensifying competition in software. Meanwhile, service businesses—the ones we’ve all ignored because, let’s be honest, their margins weren’t great—have a shot at massive value creation. Here's the math: 👩🏭 Average professional services margin: 15-25% 👨💻 Average software margin: 70-80% 🤖 AI-powered services margin: Approaching software territory AI fundamentally alters the unit economics of human labor by scaling a single person’s productivity exponentially. It can convert a 5-person team into a 50-person productivity powerhouse. Services that traditionally lived in the low-margin corner are now knocking on the door of software-like profitability. In a commoditized software world, the moat shifts. It’s not about the code—it’s about who owns the customer relationship. It’s in trust. It’s in delivery. Services businesses that leverage AI effectively are on the verge of tremendous value creation. 📈 💰 The ultimate irony is that the technology that’s supposed to take over the world, might actually put the power back in human hands. As software commoditizes, the intangible stuff—relationships, service, trust—becomes the differentiator.

Talented people have many options when choosing their next job. It's important to pick the right set of problems to work on, with the right team, at a company with a culture that resonates. To help you quickly find top options, we're launching our "Great Job" series where partner Saanya Ojha speaks with a founder at a leading startup. First up, Ansa cofounder and CTO 👨🏻💻 JT Cho shares insights on Ansa's culture of being ambitious in one's craft but humble in execution. They're hiring generalist full-stack engineers. Apply by visiting www.ansa.dev or email team@ansa.dev.

"Services businesses that leverage #AI effectively are on the verge of tremendous value creation. 📈 💰" 💯 POV from Saanya Ojha 👇

Building, testing & shipping #AI is not all generative sunshine and compute rainbows. BCV Labs recently convened inference experts to discuss the latest optimizations and improvements. Among them: -- The fundamental constraint in #AIInference is memory. Cache size becomes massive quickly – for context: the Llama 7B model hits 4.3 GB & the 65B model reaches 21.5 GB. -- One solution to inference memory constraints = sparsity. Strategically skipping certain calculations & retrieval sequencing is critical. Fastgen and Sparse Sharding Attention propose different heuristics for choosing when to compute & store vs ignore or evict. -- Mixture of Experts (MoE) has sparsity built in. As demand for larger context windows grows, we might see more sparse MoE models become increasingly popular. Check out the explainer here from Rak Garg and Dan LaBruna: https://lnkd.in/e67G6yJA

🎉 Big news — Socure has signed an agreement to acquire real-time risk decisioning company Effectiv! This strategic acquisition pairs Socure's best-in-class solutions with a developer-friendly AI orchestration and decisions platform that will elevate the way the market combats fraud and verifies identities across the entire customer journey. Effectiv’s open platform seamlessly integrates any risk solution, including identity and payment fraud controls, credit underwriting, Know Your Business (KYB) and anti-money laundering (AML), to drive real-time decisions that solve the toughest risk challenges of our time – identity theft, account takeover, scams, real-time payment fraud. Stolen or compromised identities lie at the heart of nearly every fraud attack or scam. Socure has invested more than a decade in solving what is arguably the most difficult risk challenge in the digital landscape – saying ‘yes’ to more never-before-seen consumers at onboarding, while simultaneously deterring the most complex fraud attacks, at a tremendous scale. Innovation has centered on building a comprehensive portfolio of market-leading verification and fraud solutions that uniquely provide a 360-degree view of identity. With advanced AI and more than ten years of data and model training that includes over 3.5B known identities, delivering certainty in identity is the north star for Socure. With the Effectiv platform, Socure will be able to orchestrate and write complex, combinatorial rules on its own solutions, along with those from third parties, that leverage this 360-degree view of identity. This unified approach of identity certainty automates risk and trust decisions beyond onboarding at login, authentication, payments, account changes, account recovery and regulatory filing from a single platform, with an adaptive, stateful view of the same identity at every touchpoint across the enterprise. Read the news here: https://lnkd.in/e6NcFdBY

Are the implications of #GenerativeAI overhyped in commerce? Definitely not! Trishla Ostwal shared Shoptalk discussion takeaways in ADWEEK, including perspective from BCV partner Scott Friend: “An intelligent shopping agent understands our needs, budget, and inventory that we have in the house, and takes action on our behalf. It doesn’t mean you’ll stop following a creator you love or shopping at your favorite store. But mundane tasks—like planning dinner this week or my kid has a birthday party to go to and I need a gift for a second-grade girl—will ultimately be done on our behalf.”

🚀 Founders, knowing the pivotal moment when you need a dedicated marketing leader can be tough. Even more daunting is identifying the perfect fit for your startup's unique needs. BCV’s Field Guild to Marketing Leaders is your compass in navigating the diverse landscape of marketing talent. But remember, choosing the right archetype is just the beginning. In addition to deciding which superpower is best for you, keep these four things in mind: 🔎 Diagnose Your Needs: Identify specific marketing strengths that align with your startup's needs for effective scaling. 🤝 Stay Involved: Collaborate with your marketing leader to amplify your vision while welcoming fresh perspectives. ⏱ Don't Delay: Prioritize hiring a versatile marketer to build infrastructure, generate leads, and craft your company's narrative. 👑 Don't get distracted by titles: Consider rising stars in marketing who are eager to make their mark and can work effectively without a large team. https://lnkd.in/g4y6i_gu