1Mind's AI Superhuman Now Sources 76% of the Pipeline

Zapier CEO Wade Foster opens Agents of Scale with Amanda Kahlow, founder of 1Mind, the company that just raised $40 million to sell what she calls AI superhumans: video-avatar teammates that engage, qualify, and convert. He introduces her as the founder who built 6sense into a billion-dollar-plus business. She immediately corrects the credit. She was there in the early years, then stepped down; she did not take 6sense to the $5 billion marker herself.

The longer backstory starts at 22. Working at Giga Information Group — later bought by Forrester — she asked for a raise, was told to wait six months, and spun out CI Insights instead. That company ran for 16 years as a "big data before big data was sexy" shop, stitching unstructured and structured data for Cisco, Intel, and other enterprise buyers. 6sense was born when she wanted to purpose-build software from a Cisco project and realized she needed $10 million she did not have.

Building after ChatGPT, she says, is a different sport. At 6sense she made every mistake in the book, including copying Salesforce's playbook without Salesforce's resources and hiring Salesforce people who could not transplant the machine. Now the old SaaS go-to-market framework is obsolete again. "I thought, oh, now I know the playbook of what SaaS framework and go-to-market is, and I think we're just walking into a whole new world that none of us know how to approach." Foster, still figuring out how to turn Zapier's long PLG history into enterprise sales, agrees: if a playbook existed last decade, he never quite cracked it, and there is not one for doing this with AI.

What 1Mind is actually shipping, Kahlow says, is a go-to-market superhuman: a face, a voice, and — the part she wants you to double-click — a brain. Think of it as a go-to-market Glean with agency, able to take action across the full cycle from first inbound touch through upsell. The face, she is blunt, is a gimmick to get people to talk to it. She does not believe that is where the core IP lives. The IP is the ability to carry on a conversation "exponentially better than a human."

On a website the superhuman is not a chatbot. She can share slides, give the pitch, run a live demo, read the screen, and see what the buyer is doing. A human seller walking through a deck cannot, in real time, pull the one relevant slide out of 10,000 when the conversation veers. Humans forget. They cannot go deep and wide at once. A company like Databricks sells to anyone who needs data, across every industry; its salespeople cannot speak every vertical's vernacular. The superhuman can.

The buying experience she is trying to kill is the one every B2B buyer already hates. You get qualified by a 22-year-old SDR who does not know your business, handed to an AE who cannot go deep technically, then to a solutions engineer, and after you buy you start over with customer success. "Oh my god, it's a nightmare what we do to our buyers." One context window, she argues, is itself a product opportunity.

Asked what is practically working, as opposed to the hype, Kahlow does not point at the core of enterprise sales. She points at the gaps. The value, she says, is in the places where there is no business model to put a human. She calls the motion AILG: AI-led growth. In a PLG funnel you cannot tell whether a visitor is a $500-a-month deal or a $5-million-a-month deal, so you cannot staff humans on every conversation. A superhuman can still uncover pain, explain the product at CEO altitude, and move free-trial users toward paid.

HubSpot is the named proof point. After putting a superhuman into its PLG motion, HubSpot increased revenue by 25%. A second customer, a very large social networking platform she is not allowed to name, cut average sales cycles from 22 days to two and doubled average selling price in the down-market commercial segment. Those are not "chatbot engagement" metrics. They are revenue and cycle-time numbers in the long tail that sales-led teams usually ignore.

Foster's co-founder Mike studies where AIs are strong and humans are weak, and the inverse. Kahlow's version of that split, mapped onto the go-to-market org, is pointed. Account executives are the safest role right now because of the softer skills: navigating the org, reading who actually has power. But she would say 80% of a deal is the conveying of information — the buyer's pain, the solution, the cheapest reliable path into production. On that transfer, "the AI is exponentially better than a human."

People keep asking Kahlow whether the AIs will hallucinate. It is one of her favorite questions, because she has a one-line answer: do your sellers hallucinate? They do, she says, and they do it nefariously. They know they are hallucinating to get the deal done. The superhumans, she claims, are likely to do it exponentially less. Humans still meet buyers when they are active and ready. The agents take the information job.

That split is why she is not predicting the death of the AE next quarter. SDRs whose job is qualifying, and solutions engineers whose job is mostly the demo, sit on the wrong side of the 80/20. Their core work is information transmission and recall — exactly the thing a system can do with a full history in context, in any language, in any vertical, while pulling the right slide from thousands of decks. The AE's remaining 20% is politics, trust, and getting a deal over the line. That, for now, is still a human job.

Foster asks what is happening to the go-to-market org chart. Kahlow's answer is that the current split — marketing, sales, customer success — exists because humans have capacity limits. They have time limits, recall limits, they get sick, and when they are good they leave for a competitor. Teams stay in silos and talk on an update call without a common goal. Once agents can support a buyer from first touch through upsell, she says, you will have to have one team managing the whole lifecycle.

She does not pretend the new playbook is written. She wants firms like Winning by Design to define where humans still sit once information transfer is automated. The human role, in her sketch, is to manage a customer from beginning to end, not to own a slice of a funnel. Train the superhuman well and she will not quit on you — "unless you don't pay your 1Mind bills."

1Mind is using its own AE capacity accordingly. Kahlow wants a small number of salespeople to manage relationships and close. She wants superhumans to be always on, always selling, then onboarding, then supporting customers after implementation. The next build is an onboarding superhuman; after that, one that keeps data fresh and lets customers "see into her brain." Eat your own dog food, or as she prefers it, drink your own champagne.

The champagne has numbers. Over the last 18 months, 76% of the opportunities in 1Mind's pipeline were sourced by the company's own superhuman, Mindy. She is not just first touch and not just creating the opportunity. She is shortening the cycle. When Kahlow sold 1Mind to HubSpot, she sold Kip, the CMO, and Kieran, the head of AI. That was not the deal. Twenty other people still had to get on board.

Instead of taking those meetings herself, she sent them to Mindy. Marketing-ops people asked data-integration questions; she told them to talk to the superhuman. She watched about 50 conversations land. A typical enterprise cycle that would have been around 90 days closed in 30, because the questions were already answered. She then got on a call already knowing they had asked about APIs, integrations, and hallucination.

Foster wonders whether that only works because they were buying Mindy — curiosity as a close. Kahlow says tech-to-tech buyers will talk to it, and the rest of the world is catching up because talking is not "using a tool." Owner, which sells to restaurants, cloned CEO Adam so owners can hit talk and ask about scoring. HubSpot saw 88% of visitors who landed on the page where she lived actually talk to her; 35% of those had a deep conversation — pain, solutioning, or a push to trial or purchase. A chatbot is lucky if 5% engage and maybe 2% convert, and even then it is booking a meeting with an SDR who cannot represent the business.

With traditional roles outside the AE getting automated, Foster asks where human capital goes next. Kahlow has a policy she admits is not fully papered and not yet through the board: if you fully replace your own role, she wants to forward-vest equity or otherwise reward you financially, then find you another job. People who can do that are smart and curious; there is no shortage of internal agent and workflow work. Some people will not make the transition. She is hiring for people who can. "We're doing the world a disservice when we tell people that their jobs aren't going to be replaced, because that's not the reality that we're moving towards."

That clarity collides with loss aversion. Foster notes that humans feel loss more strongly than opportunity, so even people who theoretically know the next task will be valuable still freeze. Kahlow's own story is the opposite of a gifted-kid narrative. She lived in 22 houses before she was 18; schools put her in remedial classes; she believed she was slower and dumber than everyone else until someone in her twenties took a bet on her. That person had just died, and she has to pause. The lesson she wants in the room is belief plus curiosity — or you get left behind. People who say this is not moving as fast as it is, she argues, are failing the people who need to hear the truth.

The companies that actually buy are usually buying top-down. CROs and CMOs have a board or CEO mandate to use AI, and this is a rare window when you will not get fired for trying. Five years ago the wrong tool could cost you the job. The blockers, she says honestly, are often lower in the org. People see a superhuman and think it might replace them, then slow the project, withhold content, or park the agent in a corner where nobody sees it — which then becomes evidence that it does not work. She cannot fix that from the outside. Executives have to make middle managers feel they still have a job, even if it looks different.

Foster asks what the 1% of companies do differently. Kahlow has seen the other extreme: one company had everyone go build agents, thousands of them, half doing the same thing. She can talk out of both sides of her mouth on it. Experimentation is beautiful. Production-ready it is not. Individual contributors should all know how to build a workflow; leaders still have to decide build versus buy. She cites Megan Eisenberg, CMO of Samsara, who built an in-house AEO stack — the equivalent of Profound — that is working, while also buying tools and mandating the team figure it out together.

Foster's observation from Zapier is the practical one. Orgs that make everyone build a first AI workflow or agent raise the floor. They may not raise the ceiling. They give every employee a tactile feel for what is possible, and the fear fades into the background. People can still see where they are needed. The work becomes problem-solving instead of fear-mongering. Kahlow would rather be part of the change than left on the sidelines.

The new job she is hiring for does not have a settled name. Someone has to manage AI across the whole company so the CTO, head of product, and head of customer success can stop holding weekly meetings about which agents to build. She likens it to a chief of staff for agents: the person who understands the full org because agents have to hand off across functions. Prompting, she says, anyone can learn. What she will not do is call forward-deployed "engineers" and prompt "engineers" engineers. That title belongs to people who actually trained for it. Foster prefers "builder." She adds that the work is art plus science, and the creative side is what will be most valued once agents do the tasks.

Day to day, Foster says, the job looks like reviewing agent output, holding a quality bar, tuning rubrics and evals until thumbs-up is consistent. Kahlow has already been burned by AI slop that looked great at first glance and fell apart on paper. He asks if she buys the dead-internet theory — that most of what you see on major social networks is now AI talking to AI. She had not heard the phrase. She does think the destination is agent-to-agent, and 1Mind will have to build the infrastructure for it, even if she cannot always get her head around whether it will work.

If chess is any guide, Foster says, humans will still want to watch humans even after the machines are better. Kahlow goes further: what if basic needs on Maslow's hierarchy are actually met, and people get to do the work that raises their energy? She will not get into her politics. She will say that most people are in jobs they do not love, even if they tell themselves otherwise. She loves problem-solving, people, and building something that makes change — not necessarily building superhumans as such.

Her oldest brother was a CEO, then dropped out of capitalism, spent time as a Buddhist monk, and is now a chaplain. He sits by the bedsides of people who are dying with nobody else there, and he walks homeless camps asking if anyone needs to talk. There is no money in it. "He's doing God's work." Her hope for the new world is that work like his actually gets valued — that once machines take the information jobs, the leftover human work is the part the economy used to treat as worthless.

For a 22-year-old who loves relationships and wants to be in sales, her advice is not methodology. Process, product talk, and information transfer are going to agents. What is left is creating connection. Her father told her he did not care if she learned anything in college as long as she learned how to have a conversation. A friend from Ethiopian royalty was taught at five how to walk into a party, say hello, and ask questions that invoke curiosity. That, Kahlow says, is the curriculum: learn how to engage people and lean into what makes you human.

Foster names resilience as the human skill that will be in higher demand. Kahlow, a mother of a three-year-old and a five-year-old, thinks her modest success is a trauma response. She lived through the 22 moves; she loves being told no; rejection is fuel. She has learned to feel something, get it out, and be over it — a survival mechanism. She does not want that path for her kids. Her oldest daughter is adopted; the life she would have had in a one-bedroom in Philadelphia is not a curriculum Kahlow is willing to rerun. She does not believe in gentle parenting, or in abuse. The one iron rule for childcare: if the kid is crying, do not give them what they want. Calm down, ask in a reasonable voice, or it is an instant no.

The same taste for "no" is how she builds a category. In the early 6sense fundraising days she heard no constantly. 1Mind still hears plenty of both. Too many yeses actually worry her, because that means the product is too close to the current reality — a slightly better chatbot, an incremental point solution inside today's playbooks. Category-defining companies get no because most people will not accept what is not already in the confines of what they know. She does not want everyone to get it.

She has not operationalized that instinct into a dashboard. She just loves hearing the no. Early on, at around a million dollars in revenue, she and her head of sales, Katie, would virtually kick each other under the table after a rejection. Katie would grimace; Kahlow would say it was good. If everybody loves this, they are going to fail. The nos also diagnose the market: why did they say no, and what does that tell you to build? You do not create anything big by getting yes from everyone. That just means you are slightly better.