YC Root AccessHightouch: Revolutionizing Personalized Marketing With The First Agentic Marketer
CHAPTERS
Hightouch today: outcome-driven, agentic personalized marketing
Gustaf introduces Hightouch and frames the company’s current mission: an “agentic marketer” that helps teams achieve specific marketing outcomes. The founders describe how an AI agent can personalize campaigns at the level of an individual customer, using both online and offline activity data.
- •Hightouch positions itself as “outcome-driven marketing,” starting from a marketer’s goal
- •AI agent generates propensities and campaigns tailored per customer
- •Vision: “one marketer per customer” personalization at database scale
- •Example use case: B2C upsell/cross-sell based on a customer’s full history
Formation story: quarantine house, pivot hell, and deciding to merge
The founders recount how, during early COVID quarantine, they were all pivoting separate ideas while living in the same house. Instead of anchoring on a single thesis, they bet on the team’s ability to persist and figure it out together.
- •March 2020: three founders in the same house, rapidly cycling ideas
- •Decision to work together emerged gradually through daily collaboration
- •Core belief influenced by YC: startups die when founders give up
- •Team-first thesis: trust, work ethic, and persistence mattered more than initial idea
Finding the wedge: why Reverse ETL emerged
After exploring other directions post-travel collapse, they noticed a consistent customer pain: business teams couldn’t access warehouse data easily. As warehouses like Snowflake/BigQuery grew, “Reverse ETL” became the way to move data from the warehouse into SaaS tools used by marketing and sales.
- •Customer discovery across roles revealed a shared issue: data access for business teams
- •Industry shift: companies centralizing data in modern warehouses
- •Reverse ETL concept: moving data from warehouse → SaaS tools
- •Early focus on enabling marketing teams to activate warehouse data
Getting first customers: network intros, communities, and selling while building
They landed early customers (including Retool) through a mix of network introductions and proactive outreach in online communities. They effectively sold the product as they were building it—sometimes in extremely scrappy ways—to ensure it worked end-to-end for paying users.
- •Early customers included Retool plus a couple of others in parallel
- •Acquisition channels: founder network and Slack/community discovery
- •Sold while launching; built and delivered simultaneously
- •Early scrappiness: at one point the product ran off a founder’s laptop
Signals of product-market fit: broad demand and true self-serve signup
Two moments stood out: realizing diverse global industries shared the same use case, and later seeing multiple customers onboard without talking to the team. That “pull” motion (SEO/content-driven inbound) validated the category and the go-to-market they believed in at the time.
- •Validation through variety: same need across countries and industries
- •PMF signal: a week where multiple customers onboarded self-serve
- •Inbound driven by SEO/content marketing rather than outbound push
- •Early principles: pull-based demand, self-serve setup, broad abstraction (SQL → SaaS)
Three phases of the company: category creation → business users → AI reinvention
They describe Hightouch’s evolution as three major phases. First, they created the Reverse ETL category and gained many inbound logos; next, they shifted toward business-user workflows and verticalized into marketing; later, they competed in the CDP space and chose to innovate again with an agentic marketer.
- •Phase 1: Reverse ETL category + inbound growth, lower ACV logos
- •Phase 2: build for business users beyond SQL; choose marketing as the key vertical
- •Go-to-market adjustments: more top-down selling and different hiring needs
- •Phase 3: competing with major CDPs (Salesforce/Adobe) → build agentic marketer
“Innovate or die”: why they keep rethinking the product
Even during strong growth, they repeatedly questioned whether the current product was enough for the long-term market. Their view is that marketing/data is inherently competitive and obvious; to survive, they must proactively find the next disruptive tranche rather than wait for it.
- •Market is huge and crowded; personalization has been a long-standing “unsolved dream”
- •Internal tension: investors/team asked why not just focus on what’s working
- •Founders’ mindset: continuous innovation is necessary for survival
- •Proactive strategy: search for the next opportunity instead of waiting for threats
Early hiring philosophy: high-aptitude builders from personal networks
In the earliest days, they leaned on personal connections and hired people who could grind, learn, and adapt—often before they had a formal interview process. They optimized for “startup-ready” traits: speed, resilience, and the ability to figure things out in ambiguity.
- •Early hires came primarily from personal networks (not just prior employers)
- •They didn’t initially “know how to interview,” but found strong talent anyway
- •Selection heuristic: people they’d seen work extremely hard or who clearly wanted to
- •Emphasis on high aptitude over traditional resume signals
The Hightouch interview: one 90-minute “core interview” focused on thinking
Their engineering interview process is intentionally lightweight and high-signal: a single 90-minute, no-prep, architecture/problem-solving session. It prioritizes communication, system thinking, and product/business judgment over live coding, and candidates tend to respond positively to the simplicity.
- •One main engineering interview (“core interview”), 90 minutes, no prep
- •No coding; focuses on architecture, system design, and open-ended solutioning
- •Evaluates communication, product sense, and broader problem framing
- •Candidate appeal: avoids long multi-round gauntlets; enables faster hiring decisions
Team scale and what they sell candidates on: ownership, impact, and agency
They share current headcount and how they attract talent despite competing opportunities. The pitch centers on high ownership and being rewarded for impact—especially for people who can identify what matters most and execute without being micromanaged.
- •Team size ~170; ~45 engineers plus design and PM functions
- •Culture: high ownership/agency; “make it happen” environment
- •Performance and promotions tied to impact (not strict adherence to instructions)
- •Learning environment: working alongside motivated, smart peers
Founder motivation, co-founder dynamics, and handling stress at scale
The founders discuss what keeps them going: technical challenge, responsibility to employees, and ambition to build a massive business. They also explain why multiple co-founders matter—someone is always in a low moment, and the others keep momentum—while repeated crises eventually desensitize founders as they learn quickly and move on.
- •Motivation: continuous technical challenge and changing problem space
- •Responsibility: delivering great outcomes for people who trusted them with careers
- •Co-founder benefit: emotional support and faster divide-and-conquer during thrash
- •Stress handling: frequent wins/losses normalize; focus on learning and speed of response
Staying low-profile, who they’re hiring, and where they work
Despite reaching large-company scale, they note Hightouch has remained relatively low-profile. They close by describing the type of people they need for the next AI/decisioning wave, the breadth of roles they’re hiring for, and their in-person hubs in SF and New York.
- •Low-profile brand relative to company size, despite strong performance
- •Hiring for “won’t stop until the job is done,” especially for AI tinkering/iteration
- •Wide hiring across functions (engineering, PM, sales, marketing); many open roles
- •Work setup: in-person offices in San Francisco and New York, with hybrid flexibility
