Teahose.
SIGN IN
NEW HERE — WHAT TEAHOSE DOES
We read the entire AI & tech firehose — so you don't have to.
PODPodcastsAll-In, No Priors, Acquired…
NEWNewslettersStratechery, Newcomer…
PAPPapersPhysical AI research
PHProduct Huntdaily launches
VCInvestor ScoutSequoia, a16z, Benchmark…
CLAUDE DISTILLS →
7 reads, 30 sec each — free, 6 AM ET.
+ a live graph of the companies, people & themes underneath.
HOME/DATA DRIVEN VC/🔥AI and Automation for Platform…
NEWS
// NEWSLETTER ISSUE
DATA DRIVEN VC

🔥AI and Automation for Platform Teams

DATE October 25, 2026SOURCE DATA DRIVEN VCPARTICIPANTS ANDRE RETTERATH
AI AgentsVenture CapitalEnterprise SaaSLarge Language ModelsAI InfrastructureGo-to-Market & Sales
In this episode
AnthropicOpenAILightspeedBlackRockHarmonicData Driven VCEarlybirdAtomicoGoodwinBaldertonIndexHV CapitalFindemGetroAndre RetterathAnna OttJay Anna Harris-TheisHari Kolam
// SUMMARY

1. Key Themes

Theme 1: AI is commoditizing standard platform work, so judgment, trust, and context become the product

AI now covers the routine advice, research, and benchmarking

The panel's shared view is that the work that once filled a platform lead's week is largely automatable.

"Their shared conclusion: AI now covers most of the standard advice, research, and benchmarking that used to fill a platform lead's week. What founders still call for is trust, context, and judgment, especially when things go wrong."

Founders are using LLMs themselves and then asking platform leads to review

The platform lead's role is shifting from producing analysis to validating it.

"Founders now pull benchmarks and market maps from Claude or ChatGPT themselves, and many send that output to their platform lead for a second opinion. Treat that review as part of what you offer, and put the hours AI saves you into the companies that need you most."

Reallocate saved time toward the most under-resourced founders

HV Capital directs freed-up capacity to first-time, earliest-stage founders, where the VC's brand adds the most.

"At HV Capital, that means first-time founders at the earliest stage, where Findem's founder points out a VC's name gets replies the startup's own name often can't."

Theme 2: Platform is a scale problem, which makes it the ideal AI test case

Lean teams, growing portfolios

The function is the least standardized in a VC firm, and often staffed by one person.

"Of all the functions inside a venture firm, platform is the least standardized. Many platform leads work as a team of one while the portfolio grows every year, which makes the function one of the most interesting test cases for AI in venture."

Concrete example of the scale gap

"Anna Ott, VP People at HV Capital, who brings 26 years of HR experience and supports roughly 170 to 200 portfolio companies on talent as a team of one."

Convert recurring advice into reusable AI skills

The tactical answer to the scale problem is to codify repeated advice.

"Most founder questions on onboarding, comp, or performance reviews already have tested answers. Log your calls and transcripts so a skill can draft responses you curate in minutes."

Theme 3: Proprietary, verified first-party data is the real moat for agents

Agents are only as good as their underlying data

Market benchmarks are a commodity; owned or verified data differentiates.

"Agents are only as good as the data underneath them, and market benchmarks are a commodity anyone can buy. What sets a fund apart is data it owns or can verify: cap tables, founder agreements, and network relationship data, which the panelists have yet to see another firm solve fully."

The unsolved opportunity

The panelists flag that no firm has fully solved first-party data, an open problem and a potential product or advantage.

Build an action layer on top, with flexible access

"Point your own agents at it to find founders, source talent for portfolio teams, or map the warmest intro into a deal. Then choose how your team accesses it: skill, MCP, chat, or dashboard."

Theme 4: There is a significant AI talent gap in VC platform teams

Skills are scarce relative to demand

"According to Findem's data, only about 3% of roughly 7,000 VC platform and operating professionals list AI as a skill, while about 70% of job listings now require it."

Platform teams sit mid-pack on AI maturity

"Our DDVC Landscape 2026 placed platform teams in the middle on AI maturity, and the talent gap is wider still."

Hire for learning velocity, not tool lists

"At Earlybird, we hire for people who experiment, build things themselves, and learn fast, which tells us more than any list of tools."

2. Contrarian Perspectives

Perspective 1: Measure platform success by relationship depth, not volume or hires Most firms default to quantifiable output metrics. HV Capital's VP People watches whether founders keep coming back. This is a deliberately qualitative signal that is also AI-resistant, since trust cannot be automated.

"HV Capital's VP People watches whether founders keep reaching out: 'It's less about the hires and the volume. It's about the depth of the relationship.'" Earlybird, by contrast, "track[s] OKRs," so the article shows the metric choice is itself a strategic decision: "Decide what founders call you for, then pick one signal to track."

Perspective 2: AI-generated output doesn't reduce the platform role. It adds a review service The obvious fear is that founders using ChatGPT or Claude makes platform teams redundant. The panel argues the opposite: that behavior creates a new demand for second opinions.

"Founders now pull benchmarks and market maps from Claude or ChatGPT themselves, and many send that output to their platform lead for a second opinion." Supporting reasoning: when stakes are high ("especially when things go wrong"), founders want trusted human judgment rather than generic answers.

Perspective 3: A VC's name is more valuable than the startup's own for earliest-stage outreach The reasoning supports prioritizing first-time founders, and suggests platform teams can lend their brand as an operational lever.

"A VC's name gets replies the startup's own name often can't."

3. Companies Identified

Findem

  • Description: People intelligence platform for VCs; co-hosted the session and sponsored the issue.
  • Why mentioned: Provides verified people and talent data used in the proposed workflows, accessible via MCP.
  • Quotes:

    "Findem makes it queryable. 800M+ verified profiles, 4M+ companies, and the warm paths your team already has, available over MCP." "Findem's 3D People Graph is one option: it connects people, companies, and career history across 100,000+ sources, with profiles verified and refreshed over time." "Findem mapping the talent market."

HV Capital

  • Description: Venture firm; case study in a lean, talent-focused platform model.
  • Why mentioned: A team of one supports roughly 170 to 200 portfolio companies on talent.
  • Quotes:

    "HV Capital treats it as talent advice" "At HV Capital, that means first-time founders at the earliest stage..."

Earlybird

  • Description: Venture firm; runs platform inside a broader "Flywheel" team.
  • Why mentioned: Case study in AI-fluent hiring and OKR-based tracking.
  • Quotes:

    "we run it inside a larger Flywheel team at Earlybird" "At Earlybird, we hire for people who experiment, build things themselves, and learn fast"

Getro

  • Description: Platform connected to Hari Kolam's teams, serving VC and PE firms.
  • Why mentioned: Contributes to the scale of Findem's VC/PE footprint.
  • Quotes:

    "a two-time founder whose teams across Findem and Getro serve more than 700 VC and PE firms."

Claude / ChatGPT

  • Description: General-purpose LLM tools used by founders.
  • Why mentioned: Founders use them to self-generate benchmarks and market maps, shifting the platform role.
  • Quotes:

    "Founders now pull benchmarks and market maps from Claude or ChatGPT themselves"

Balderton & Harmonic

  • Description: VC firm and AI startup-discovery company co-hosting an upcoming webinar.
  • Why mentioned: Upcoming event partners for "The Agentic Fund."
  • Quotes:

    "'The Agentic Fund' webinar on Oct 28 with Balderton & Harmonic"

Goodwin & Harmonic

  • Description: Law firm and Harmonic co-hosting the Helsinki event.
  • Why mentioned: Partners for the DDVC x Slush Breakfast.
  • Quotes:

    "DDVC x Slush Breakfast in Helsinki Nov 18 with Goodwin & Harmonic"

Accel, Atomico, Balderton, BlackRock, Index, Lightspeed

  • Description: Leading investors whose AI-and-automation deep dives are in The Lab library.
  • Why mentioned: Signal the caliber of practitioners sharing workflows.
  • Quotes:

    "100+ other AI and automation for VC deep dive recordings from leading investors at Accel, Atomico, Balderton, BlackRock, Index, Lightspeed, and many more."

4. People Identified

Anna Ott

  • Description: VP People at HV Capital.
  • Why mentioned: Panelist; exemplar of a one-person platform team covering a very large portfolio.
  • Quotes:

    "brings 26 years of HR experience and supports roughly 170 to 200 portfolio companies on talent as a team of one." "'It's less about the hires and the volume. It's about the depth of the relationship.'"

Jay Anna Harris-Theis

  • Description: VP Portfolio Excellence at Earlybird.
  • Why mentioned: Panelist; runs a multi-function platform ("Flywheel") covering talent, GTM, tech partnerships, and knowledge sharing.
  • Quotes:

    "advises our startups on talent, HR advisory, tech partnerships, go-to-market, and knowledge sharing as part of our Flywheel team."

Hari Kolam

  • Description: Co-Founder & CEO at Findem; two-time founder.
  • Why mentioned: Panelist providing the people-data and venture-network perspective.
  • Quotes:

    "a two-time founder whose teams across Findem and Getro serve more than 700 VC and PE firms." "Findem's founder points out a VC's name gets replies the startup's own name often can't."

Andre Retterath

  • Description: Author of Data Driven VC; moderated the roundtable.
  • Why mentioned: Newsletter author and session host.
  • Quotes:

    "I was joined by..." "which is all about becoming a better investor with data and AI."

5. Operating Insights

1. Build a "hiring plan stress test" skill before every founder call Automate the preparation step so human time is spent on judgment.

"The workflow I'd copy first: a second skill that stress tests every new hiring plan against competitor hiring, talent supply, and comp, with Findem mapping the talent market. You go into the first founder call already knowing which roles are realistic and which need a headhunter."

2. Log calls and transcripts to build a corpus of reusable advice skills Capture your own best answers first, then draft and curate.

"Log your calls and transcripts so a skill can draft responses you curate in minutes."

3. Define one success signal and give your team time to build with AI Pick a single metric aligned with what founders actually call you for, and protect experimentation time because lean teams rarely find it themselves.

"Decide what founders call you for, then pick one signal to track." "Give your platform team time to build with AI, since most of them run lean and their portfolios keep growing."

6. Overlooked Insights

1. Access architecture is a deliberate design choice (skill, MCP, chat, dashboard) The article closes by noting that how a team consumes data matters as much as the data itself, and MCP is positioned as an emerging standard for connecting proprietary data to agents.

"Then choose how your team accesses it: skill, MCP, chat, or dashboard." Related resource: "Build an MCP server for investors (with or without coding)"

2. Firm-level AI maturity is benchmarked, and platform sits in the middle A landscape study ranks functions by AI maturity, implying that sourcing and other functions may be ahead and that platform has a catch-up opportunity (and a talent constraint).

"Our DDVC Landscape 2026 placed platform teams in the middle on AI maturity, and the talent gap is wider still."