{
  "video": {
    "id": "APqXGyCoGW4",
    "title": "Forward Deployed Engineering at Cursor — Pauline Brunet",
    "duration": 1247,
    "upload_date": null,
    "channel": "AI Engineer",
    "source": "AI Engineer",
    "event_format": "In person"
  },
  "analysis": {
    "video_id": "APqXGyCoGW4",
    "title": "Forward Deployed Engineering at Cursor — Pauline Brunet",
    "one_liner": "Cursor's global FDE lead maps where forward deployed engineering actually pays off — a matrix of customer digital maturity against product customizability — and gives concrete rules for hiring, scoping and proving ROI without sliding into staff augmentation.",
    "summary": "Pauline Brunet, who leads Cursor's forward deployed engineering team globally after 10 years of enterprise AI deployments, argues FDE is not professional services or staff augmentation: it only fits a middle band on a matrix of customer digital maturity × product customization, and the extremes should be served by self-service docs, traditional SaaS rollout, or advisory. She describes Cursor's own team — 5+ year software engineers only, hired from Spotify, Rippling and Palantir, organised by geography today and by industry later — working project-based with the economic buyer, building long-running cloud agents, automations and Cursor SDK apps directly on customer code bases. The operational half of the talk is about scoping directionally rather than selling FDE-months, involving the customer at every step, and measuring ROI as one of exactly three things: revenue up, cost down, risk mitigated.",
    "key_points": [
      "The FDE fit test is a 2x2 of customer digital maturity vs product customizability. High maturity + low customization → ship self-service with great docs. Low maturity + low customization → a traditional waterfall SaaS deployment. High maturity + high customization → FDEs as advisers who accelerate. Low maturity + high customization → embedded transformation. FDE's real value sits in the band across the middle, not at the corners.",
      "Explicit anti-patterns: FDEs should not act as solution architects, write bug tickets, update documentation, or run product 101/201 workshops — do that and the 10x engineers you paid to attract will get bored and leave because 'you sold them FDE that wasn't FDE.'",
      "Staff augmentation red flag is the phrase 'we're understaffed'; her counter-move is to always ask the customer who the named working team on their side will be before agreeing to a project.",
      "Cursor's hiring profile today is deeply technical engineers with 5+ years and customer-facing experience — no new grads, no early-career hires yet — sourced from Spotify, Rippling and Palantir; the role will later split into more customer-facing and more technical variants.",
      "Team structure: matrix org, geography-based now but expected to move to industry alignment, because 'if you talk to a bank and you're not talking about payment systems... asset management, risk, you kind of lost them.' Product SMEs exist per area — a named expert on long-running cloud agents, another on the Cursor SDK — who get pulled into others' projects.",
      "The FDE team is 'the tip of the spear' for use cases outside the software development lifecycle — HR, finance, supply chain, e-commerce, call-center ticketing, claims management, asset management — feeding those edge cases back to product and engineering.",
      "Scope directionally, never 'take two FDEs for 6 months and do whatever you want with them' — she calls that a recipe for failure. Define the problem, the phases, and a 6-week window, because you haven't seen the customer's processes, data or systems yet and both sides will want to pivot on what they learn.",
      "ROI is always exactly three things: increasing revenue, decreasing costs, or mitigating risk. A customer alarmed by an agent costing $2,000/day had never measured it against the cost of dispatching the right person to fix equipment — 'Absolutely, but I never measured it this way.'",
      "Listening changed the offering: asked six or seven times 'how do I change my organization now that we have these amazing tools — who do I hire, what are the job descriptions, how do I rearrange teams?', she is now building an FDE offering for it that she never planned."
    ],
    "takeaways": [
      "Before building an FDE function, place your target customers on the maturity × customizability matrix and only staff FDE where it lands in the middle band; serve the corners with documentation, a standard deployment, or light advisory instead.",
      "Scope every engagement to a problem and a KPI with a stated baseline ('this process goes from 3 hours to 20 minutes — is that success for you?'), agreed with the economic buyer or a senior champion up front, and keep the plan directional across phases rather than selling headcount-months.",
      "Insist on a named customer working team and top-down support before starting; co-develop in their code base rather than alone, so the thing isn't switched off when you leave.",
      "Say no when your product isn't the right tool — it builds the credibility that earns you the next, better use case, and it keeps you off the hook for engagements that will go badly.",
      "Push the low-value work to system integrators and consultancies who already have the relationships — change management, broad rollouts of mature products across telco and healthcare/life sciences — and keep FDEs on the edge cases."
    ],
    "topics": [
      "forward-deployed-engineering",
      "enterprise-deployment",
      "hiring",
      "agents",
      "roi",
      "scoping",
      "org-design",
      "ai-coding"
    ],
    "tools": [
      "Cursor",
      "Cursor SDK",
      "Cursor cloud agents / long-running agents",
      "Palantir",
      "Spotify",
      "Rippling",
      "Teams",
      "Forbes"
    ],
    "quotes": [
      {
        "text": "I always say, if you put in the latest and greatest tech in your organization, and you don't accompany people, no one's going to use it.",
        "at": "06:13",
        "url": "https://www.youtube.com/watch?v=APqXGyCoGW4&t=373s"
      },
      {
        "text": "So, my trick is I just ask for who are the people we're going to work with.",
        "at": "10:06",
        "url": "https://www.youtube.com/watch?v=APqXGyCoGW4&t=606s"
      },
      {
        "text": "We are still developing on top of their code base, but make sure you're not doing it alone. If you're doing it alone in their office in a little cubicle, we have a problem.",
        "at": "18:45",
        "url": "https://www.youtube.com/watch?v=APqXGyCoGW4&t=1125s"
      },
      {
        "text": "It's always three things, super simple. Am I increasing revenue? Am I decreasing costs? Or am I mitigating risks? That's it.",
        "at": "19:33",
        "url": "https://www.youtube.com/watch?v=APqXGyCoGW4&t=1173s"
      }
    ],
    "words": 4560
  },
  "summary_url": "/#APqXGyCoGW4",
  "transcript": {
    "html": "/transcripts/APqXGyCoGW4.html",
    "txt": "/transcripts/APqXGyCoGW4.txt",
    "vtt": "/transcripts/APqXGyCoGW4.vtt"
  }
}