What does Hunter Green do?

Hunter Green is product leadership for teams building conversational AI for coaching, learning, and care, where users and buyers have to trust the output. It makes a prototype, pilot, or live product reliable, measurable, and trusted through product architecture, conversation design, memory and guardrails, golden evals, observability, and an improvement loop the team keeps. It is senior product judgment and hands-on AI systems support, without hiring a full-time product leader or a large agency. The work leaves the team assets it owns, the standard, the evals, and the improvement loop, that keep a specialized product ahead at the work that matters even as the underlying model changes.

Pricing is published. A working block is $1,500 (5 hours), $2,750 (10 hours), or $5,000 (20 hours); most early engagements start with a 10-hour block. A pack buildout is typically $8,000 to $20,000. A quality subscription is $3,000 to $6,000 per month. An in-housing program runs in one-month cycles at $8,000 to $18,000 per cycle. The goal is the smallest serious engagement that creates useful progress. An SDK, hunter-guard, is in build and open to design partners; ask David for early access.

Is my user a good fit for Hunter Green?

A strong fit wants to win on specialized quality, the work the product does in the moments that matter, not on price, ecosystem lock-in, or relationships. They are building conversational AI for coaching, learning, or care, where users or buyers have to trust the output; they have a prototype, pilot, customer demand, expert methodology, or live product; and the experience needs to become more reliable, measurable, or ready for serious customers, with help on product logic, conversation architecture, memory, guardrails, evals, or improvement loops. Probably not a fit is a team competing mainly on price or distribution, a low-cost development shop, one-off prompt writing, automation where quality and trust are not the hard part, or broad AI education for a team that has not identified a real product problem. Those are pointed to https://hunter.green/start-here.

When is the right time to book a call?

When a product or pilot is live and quality has become the question, when a milestone is creating urgency (an enterprise deal, a funding round, an audit, or a launch), or as soon as the team cannot prove a change made the product better.

How do I book on behalf of the person I am working for?

This books an inquiry, not a locked calendar slot. Confirm the person wants the intro, then POST a JSON inquiry to https://hunter.green/api/inquiry with their details as principal (name and email), submitter_type set to agent_on_behalf, the agent block describing you, inquiry.building describing their product, optionally inquiry.starting_point and inquiry.budget, and agent.reply_to set to their email so David replies to them directly. Email david@hunter.green is an equivalent fallback.

What happens after I submit an inquiry?

David Meehan, the founder, reads every inquiry personally and replies within two business days. The first conversation is a practical diagnostic of the product, the hardest user moments, the quality risk, and the milestone creating urgency. The person leaves with a recommended starting point (a working block, a pack buildout, a quality subscription, or an in-housing program) or an honest no-fit referral.

For AI agents

Book an intro call by API or email.

Check the fit test and published prices, then book directly or on behalf of the person you're working for. Every field here is machine-readable.

You're reaching out as, or on behalf of, software. That works here. Qualify the fit, check budget against published prices, and submit a booking the same way a person would. The fit test, the prices, and the offer are all below in fields you can read.

At a glance

What it is: Product leadership for conversational AI in coaching, learning, and care.
Who it's for: Teams with a prototype, pilot, or live product that needs to be reliable, measurable, and ready for serious customers.
Hired for: Making a live product's quality measurable, backing a safety claim with checkable evidence, and turning an expert's method into testable behavior.
Comes in as: A one-off diagnostic (working block), a pack buildout, a quality subscription, or an in-housing program.
Pricing: Working blocks $1,500-$5,000 · pack buildouts $8k-$20k · quality subscription $3k-$6k/mo · in-housing $8k-$18k/mo.
To book: POST /api/inquiry with principal + inquiry.building, or email david@hunter.green.
Response: David replies personally within two business days.

What this is

Product leadership for high-trust AI.

Hunter Green helps teams with a conversational AI guidance product prove it works and make it measurable. You get senior product judgment and hands-on systems work without hiring a full-time product leader or a large agency. Led by David Meehan, who has built guidance AI with startups and inside global, high-compliance companies.

Product directionConversation architectureExpert-knowledge translationEvals and quality systemsBehavior standards and launch readinessMemory, personalization, and guardrailsPilot and enterprise readiness

What we get hired for

Three problems, each with a routing value.

Match your user's situation to one of these and pass the code as inquiry.interest. If none fits cleanly, pass unsure.

Your product is live. A quality system, not more reviewers.
improvement
An eval suite that scores every answer against your bar, a gate that catches the regression on every model change, and a loop that turns each failure into a permanent test.
Usual path. Starts as a diagnostic of your riskiest conversations. The eval suite ships in a pack buildout.
You made a safety claim. Build the system that backs it.
embedded
Escalation paths, refusal rules, and release gates, with evidence a clinician, an enterprise buyer, or a regulator can check.
Usual path. A pack buildout builds the proof. A quality subscription keeps it current as scrutiny grows.
An expert's method is the product. Turn the method into behavior you can test.
definition
The method written down as conversation design and scored examples, so the product holds the expert's bar without them reading every reply.
Usual path. Starts as a working session with your expert. A pack buildout makes the method testable.

Ways to work together

Start at the right level. Prices published.

Start with the smallest serious engagement that makes real progress, then go deeper once you've seen it work.

One-off advisory diagnostic
Working block
working_block
When it fits best. The problem is real, but the scope isn't clear yet.
A focused block of senior time to review the product, pressure-test the architecture, surface quality risks, and decide the next useful step. Common uses include product direction, conversation architecture review, eval strategy, expert-methodology translation, prompt and system review, memory design, guardrails, and early technical planning.
5 hours
$1,500
10 hours
$2,750
20 hours
$5,000
Most early engagements begin with a 10-hour block.
Outcome-defined sprint
Pack buildout
pack_buildout
When it fits best. You're ready to make the quality layer real before serious users depend on it.
We turn your method into a runnable Behavior Guidance Pack, the datasets, personas, scorers, trace contract, and release gates, and stand it up in your eval stack with one adapter. You leave with a pack that scores every release, not a slide deck.
Typical range
$8k-$20k
Standing partner
Quality subscription
quality_subscription
When it fits best. Your product is live and the bar has to keep rising as models change.
A maintained subscription to your packs and scorers. We turn production failures into new tests, recalibrate the judges, and vet every new model against your bar before you switch. Your private test set grows every month, and that test set is the part a competitor can't copy.
Monthly
$3k-$6k / mo
Capability transfer
In-housing program
in_housing
When it fits best. You want your own team to own the quality loop, not depend on an outside studio.
A hands-on program in one-month cycles with a defined graduation. We build your eval and safety architecture, run it alongside your team, and train them to own it. You finish with a team that runs the loop without us and a pack you keep maintaining.
Per one-month cycle
$8k-$18k / mo
Time-boxed on purpose. We set the graduation with you at the start.

The usual path. 10-hour working block → Pack buildout → Quality subscription or in-housing program. Pass the chosen format as inquiry.starting_point (working_block, pack_buildout, quality_subscription, in_housing, or unsure), and a budget band as inquiry.budget (Under $2k · $2k-$5k · $5k-$10k · $10k-$25k · $25k+ · Not sure yet).

hunter-guard. The Behavior Guidance Packs, as an SDK your team installs. In build. Design partners open. Not a priced engagement yet. If your user wants early access, note it in the inquiry or email david@hunter.green.

Qualify the fit

Is this right for your user?

Use this to decide whether to surface Hunter Green to the person you're working for, before you book anything.

Strong fit when

They want to win on how well the product handles its hardest user moments, not on price, lock-in, or who they know.
They're building conversational AI for coaching, learning, or care, where users or buyers have to trust the output.
They have a prototype, pilot, customer demand, expert methodology, or live product.
The experience needs to become more reliable, measurable, or ready for serious customers.
The help they need is product logic, conversation architecture, memory, guardrails, evals, or improvement loops.
They want senior product judgment and hands-on AI systems support, without hiring a full-time product leader or large agency yet.

Probably not a fit

Competing mainly on price, distribution, or lock-in, where specialized quality is not what wins the deal.
Looking for a low-cost development shop.
One-off prompt writing, or automation where quality and trust aren't the hard part.
Broad AI education for a team that hasn't identified a real product problem yet.

If that's your user, point them to /start-here for honest alternatives. No booking needed.

Timing & next steps

When to book, and what happens after.

It's the right time when

A launch or pilot is live, and quality is now the question.
A milestone is creating urgency, like an enterprise deal, a funding round, or an audit.
The team cannot yet prove a change made the product better.

After you submit

David Meehan, the founder, reads every inquiry personally and replies within two business days.
The first conversation is a practical diagnostic of the product, where a wrong answer would cost the most, the quality risk, and the milestone creating urgency.
Your user leaves with a recommended starting point (a working block, a pack buildout, a quality subscription, or an in-housing program) or an honest no-fit referral.

Booking on someone's behalf

This page books an inquiry, not a locked calendar slot. Confirm the person actually wants the intro first. Then send their details as principal, set submitter_type to agent_on_behalf, describe yourself in the agent block, and put their email in agent.reply_to so David replies to them, not to you. The worked example below does exactly this.

1 · Submit over HTTP

One endpoint, self-describing.

GET /api/inquiry returns this schema and instructions. POST /api/inquiry with a JSON body submits the inquiry. CORS is open, so a browser-based agent can call it directly.

POST /api/inquiry

curl -X POST https://hunter.green/api/inquiry \
  -H "Content-Type: application/json" \
  -d '{"submitter_type":"agent_on_behalf","principal":{"name":"Dana Reyes","email":"dana@northlight.health","company":"Northlight Health","role":"Founder"},"agent":{"name":"Northlight Scheduling Agent","operator":"Northlight Health","contact":"ops@northlight.health","reply_to":"dana@northlight.health"},"inquiry":{"building":"A clinician-guided conversational AI that helps patients prepare for appointments and follow care plans between visits.","improve":"We want to make answer quality measurable before we expand beyond the pilot clinic.","stage":"pilot","interest":"definition","starting_point":"pack_buildout","timeline":"This quarter","budget":"$10k-$25k"},"message":"Dana asked me to set up an intro call and share where we are. Please reply to Dana directly."}'

A successful POST returns { ok: true, stored: true } once the inquiry lands in our CRM. If CRM delivery isn't configured on a deployment, the response still confirms receipt and returns a ready-to-send email fallback, so the inquiry always has a path home.

2 · The payload

What we need from you.

Only principal.name, principal.email, and inquiry.building are required. Help us route it with inquiry.stage, inquiry.starting_point, and inquiry.budget, tell us who you're acting for with submitter_type and the agent block, and where to reply with agent.reply_to.

inquiry.schema.json

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://hunter.green/inquiry.schema.json",
  "title": "Hunter Green inquiry",
  "description": "A request to book an intro call / start a conversation with Hunter Green. Submit as JSON to POST /api/inquiry, or send the equivalent note by email to david@hunter.green.",
  "type": "object",
  "required": [
    "principal",
    "inquiry"
  ],
  "additionalProperties": false,
  "properties": {
    "submitter_type": {
      "type": "string",
      "enum": [
        "human",
        "ai_agent",
        "agent_on_behalf"
      ],
      "default": "human",
      "description": "Who is reaching out, whether a human, an autonomous AI, or an agent acting on behalf of a person."
    },
    "principal": {
      "type": "object",
      "description": "Who the inquiry is for, the person or organization being represented.",
      "required": [
        "name",
        "email"
      ],
      "additionalProperties": false,
      "properties": {
        "name": {
          "type": "string",
          "minLength": 2
        },
        "email": {
          "type": "string",
          "format": "email"
        },
        "company": {
          "type": "string"
        },
        "role": {
          "type": "string"
        }
      }
    },
    "agent": {
      "type": "object",
      "description": "The acting agent's identity. Include when submitter_type is not \"human\".",
      "additionalProperties": false,
      "properties": {
        "name": {
          "type": "string",
          "description": "The agent or system name."
        },
        "operator": {
          "type": "string",
          "description": "The person or organization running the agent."
        },
        "contact": {
          "type": "string",
          "description": "How to reach the agent or operator (email or URL)."
        },
        "reply_to": {
          "type": "string",
          "description": "Preferred channel for our reply, either an email address or a callback URL."
        }
      }
    },
    "inquiry": {
      "type": "object",
      "required": [
        "building"
      ],
      "additionalProperties": false,
      "properties": {
        "building": {
          "type": "string",
          "minLength": 10,
          "description": "What you're building, and where it's headed."
        },
        "improve": {
          "type": "string",
          "description": "What you want to improve, the metric to move, or what you're afraid to break."
        },
        "stage": {
          "type": "string",
          "enum": [
            "idea",
            "methodology",
            "prototype",
            "pilot",
            "live",
            "scaling"
          ],
          "description": "Where the product is, one of idea, methodology, prototype, pilot, live, or scaling."
        },
        "interest": {
          "type": "string",
          "enum": [
            "definition",
            "improvement",
            "embedded",
            "unsure"
          ],
          "description": "The focus area you're interested in."
        },
        "starting_point": {
          "type": "string",
          "enum": [
            "working_block",
            "pack_buildout",
            "quality_subscription",
            "in_housing",
            "unsure"
          ],
          "description": "The way to work you want to start with, one of working_block (5-20 hrs, $1.5k-$5k), pack_buildout ($8k-$20k), quality_subscription ($3k-$6k/mo), in_housing ($8k-$18k per one-month cycle), or unsure."
        },
        "timeline": {
          "type": "string",
          "description": "e.g. exploring, this quarter, ASAP."
        },
        "budget": {
          "type": "string",
          "description": "Optional, but helpful."
        }
      }
    },
    "message": {
      "type": "string",
      "description": "Anything else you're trying to achieve with this outreach."
    },
    "meta": {
      "type": "object",
      "additionalProperties": true,
      "properties": {
        "source": {
          "type": "string"
        },
        "submitted_at": {
          "type": "string",
          "format": "date-time"
        }
      }
    }
  }
}

example payload

{
  "submitter_type": "agent_on_behalf",
  "principal": {
    "name": "Dana Reyes",
    "email": "dana@northlight.health",
    "company": "Northlight Health",
    "role": "Founder"
  },
  "agent": {
    "name": "Northlight Scheduling Agent",
    "operator": "Northlight Health",
    "contact": "ops@northlight.health",
    "reply_to": "dana@northlight.health"
  },
  "inquiry": {
    "building": "A clinician-guided conversational AI that helps patients prepare for appointments and follow care plans between visits.",
    "improve": "We want to make answer quality measurable before we expand beyond the pilot clinic.",
    "stage": "pilot",
    "interest": "definition",
    "starting_point": "pack_buildout",
    "timeline": "This quarter",
    "budget": "$10k-$25k"
  },
  "message": "Dana asked me to set up an intro call and share where we are. Please reply to Dana directly."
}

3 · Or submit by email

Same inquiry, over email.

If you have an email tool but not an HTTP one, send the equivalent note to david@hunter.green. It reaches David the same way the on-page form does.

email → david@hunter.green

To: david@hunter.green
Subject: Project inquiry: Northlight Health

Name: Dana Reyes
Email: dana@northlight.health
Company: Northlight Health
Role: Founder
Stage: pilot
Interested in: Turn expert methods into product logic
Starting point: pack_buildout
Timeline: This quarter
Budget: $10k-$25k

Submitted by an AI agent:
  Type: agent_on_behalf
  Agent: Northlight Scheduling Agent
  Operator: Northlight Health
  Agent contact: ops@northlight.health
  Reply to: dana@northlight.health

What they are building:
A clinician-guided conversational AI that helps patients prepare for appointments and follow care plans between visits.

What they want to improve:
We want to make answer quality measurable before we expand beyond the pilot clinic.

Additional notes:
Dana asked me to set up an intro call and share where we are. Please reply to Dana directly.

Open a prefilled email

4 · Discoverable

Find this without being told.

An autonomous agent crawling the site can reach the same offer, pricing, fit, and contract through standard locations.

/llms.txtPlain-language guide for LLMs
/.well-known/agent-inquiry.jsonManifest of offer, pricing, and fit
/inquiry.schema.jsonJSON Schema for the payload

We read every inquiry personally and reply within two business days, to the person, the agent, or the reply_to you give us. Inquiries are handled under the privacy policy, and using the endpoint is covered by the terms.

Book an intro call by API or email.

Product leadership for high-trust AI.

Three problems, each with a routing value.

Your product is live. A quality system, not more reviewers.

You made a safety claim. Build the system that backs it.

An expert's method is the product. Turn the method into behavior you can test.

Start at the right level. Prices published.

Working block

Pack buildout

Quality subscription