High-alpha GTM data tools and Forward-Deployed Researcher methodology for cracking long-tail deep-tech markets. Sales as research. Researchers as sales. The lab's brain in the field.
The market for AI in deep tech is a long tail of small verticals, each one more idiosyncratic than the last. Corrugated-paperboard process control. Satellite-bus thermal modeling. Cathode dry-room humidity in a battery gigafactory.
These markets are smaller and more intimate than traditional AI SaaS markets. Unique dynamics. Unique elites. Unique standards. Unique trends. Unique counter-positioning opportunities. Unique media. Unique people. Unique data. Unique envs.
Louder GTM helps here. Use your AI SDRs and your MEDDICC and all of it. But smarter, more curious GTM — like a better search algorithm — is what really conquers these markets.
The great restaurateur Danny Meyer, who has conquered the NYC dining market with restaurants ranging from the Gramercy Tavern (I've been — excellent) to Shake Shack (ditto — great shakes), said the secret to taking over a market is Always Be Collecting Dots so you can Always Be Connecting Dots. Only through curiosity and research and trial and error can you really earn the right to talk with folks and say, "People like us do things like this."
Before he got into restaurants, Meyer was the top seller of Checkpoint anti-shoplifting systems. Sold thousands and thousands of little electromagnetic scanners that detect RF tags as you leave the store, making millions. He hit the pavement and did customer demos. But he also looked around:
"In the New York retail world, many of them were related to each other — there was a huge Syrian Jew population. I started to develop a sense for the family trees, and I got to know who knew who, and I would take that as far as I could possibly take it."
You can run your demo at a thousand bodegas. Or you can cook one nice lamb dinner on a Friday. You have to find the elites, learn the standards, develop your own expertise, design experiments, build the datasets (maybe even sell them), make envs, put red string to bulletin board, just really think. Today's key-logged SaaS sellers are smart and observant, but burned out and spinning their wheels. Acknowledge their role as researchers bringing back vital bitter feedback, empower them to think more like founders, and you unlock types of intelligence the competition doesn't have and AI GTM solutions don't offer. Let them find the elites, develop the elites, and run a strategy validated by unique research data more than traditional Salesforce KPIs.
I applied this at Garden Intel, a patent analytics startup where I was the first and only true GTM hire and where I spent a year running demos from nine to nine, closing deals that ranged from under fifty thousand to north of a million, cold email to close. The company eventually sold for $150M. The job was entirely research — who are these people, what do they care about, which dogs hunt with them. It's the same Who/What/When/Where/Why research-loop I used as a deep-tech consultant, VC analyst, and Stanford Daily reporter.
I went to high school at Saint Ignatius in San Francisco — a Jesuit school that, for reasons the order doesn't advertise, produces an outsized share of GTM movers and shakers at Anthropic, the FAANGs, Oracle, half the great YC startups, the CEO of Hims. The pattern isn't just aptitude. It's that the Jesuits were the original forward-deployed engineers — wily, sent for better and for worse to gather souls in half the courts and tribes of the world. Matteo Ricci spent decades learning Mandarin and the Confucian classics until the Ming emperor made him imperial astronomer. Francis Xavier did the same in sixteenth-century Japan. The Jesuit going into a foreign court had to read everything the locals had read, learn the family trees of the local elites, and hold a substantive conversation with the brightest courtier.
Take two modern AI-industrial startups, same product space, same year. The first hires two SDRs, books conferences, runs LinkedIn ads, cold-calls every plant in the SIC code, tokenmaxxes, gets a McKinsey tokenmaxxer medallion. By month nine: thin pipeline. The second sends its FDEs into the field with no quota and one instruction — go learn the vertical. They map the actual buyers — six firms, not six hundred. Three months on the standards committee teach them which clauses are about to be rewritten. The six operators they befriended become a customer advisory board. The market doesn't have an AI-readable ontology, so they write one. Who do you think captures economics long term?
Ken Stanley calls this open-endedness — Why Greatness Cannot Be Planned. Great stepping stones get collected by people searching for what's interesting, not by people optimizing toward a known goal. Quotas pre-converge the search. Curiosity keeps it open. The few long-tail markets that bloom will be the ones a curious FDE was already inside, by accident, six months before anyone could have known to point them there.
Ashiba Research develops high-alpha GTM data tools and helps enterprises crack long-tail markets by dramatically expanding the surface area of intelligence and empowering the sales function to be a research function. Whether you are just selling Claude Code or Codex to local influencers in a new domain or whether you are trying to convince a Korean HBM manufacturer to buy a new $500,000 metrology machine, either way you have to use AI to really niche down and get curious in a way that few FDEs ever had to do in the past.
A deeper claim would be that the forward-deployed researcher is the lab's brain in the field, bringing back the gradient that decides the production possibility frontier of intelligence. The bitter algorithm picks most of the rest; environment selection is the residual decision and it's the only one that matters. Whoever wins the right envs wins.
— Cooper Veit, Ashiba Research
Map the six firms that matter, not the six hundred in the SIC code. Identify the standards-body clauses about to get rewritten. Surface the operators worth turning into a customer advisory board. The output is research, not a CRM.
The tools that empower a small FDE team to operate at the intelligence density of a fifty-person sales org. Not AI SDR replacements. Researcher amplifiers — the ones that let one curious FDE map a vertical in ninety days that a McKinsey-tokenmaxxed team will be lost in for two years.
The Jesuit playbook adapted for AI-era deep tech. How to recruit researchers who can sell, train them on a vertical, deploy them with no quota and the right instruction, and capture the bitter feedback gradient they bring back. The output is a research function that funds itself by closing deals.
The strategic layer above sales. The bitter algorithm picks most of the rest; environment selection is the residual decision and it's the only one that matters. Ashiba GTM is the partner for that decision — the team that already lived inside the vertical for six months tells you which envs are worth winning.
Ashiba GTM is for AI-industrial startups, deep-tech enterprises, and frontier labs trying to enter long-tail verticals — anyone selling AI software, agents, evaluation environments, or specialty hardware into markets that have unique elites, unique standards, unique data, and unique counter-positioning opportunities.
The most useful first conversation is concrete. Bring:
From there we can answer the only question that matters: is this a vertical where curiosity beats cadence, and if so, what does the field-deployment look like for your team?
Read the manifesto as PDF → · Ashiba Deep Tech → · cv@ashibaresearch.com