# Audience research

Who buys what you sell, what they struggle with and where to reach them, each point cited.

Requires web search, through Exa. Requires web fetch.

## Install

```sh
npx @parcha/agentrun template install github:Parcha-ai/agentrun-templates/templates/audience-research@242f29da8317fbe51d0cce4628bdb5ab61d7ccc9 --trust
```

Or open `<your AgentRun address>/#/templates/install?src=github%3AParcha-ai%2Fagentrun-templates%2Ftemplates%2Faudience-research%40242f29da8317fbe51d0cce4628bdb5ab61d7ccc9`. Pinned to commit 242f29da8317fbe51d0cce4628bdb5ab61d7ccc9.

## What it does

Who buys what you sell, and how do you reach them? Give your website or describe an audience, and get the priority buyer segments, the roles that own the problem and the budget, their pains in their own words, where they gather, and what sets off a purchase, each from a cited source.

- Skill `audience-research`: Profile the buyers of one company's product, or one described audience. Decide company or segment mode, read the company's site and competitors' customers, collect pains in buyers' own words from reviews and forums, find the roles that own the problem in job posts, where the audience gathers, buying triggers and objections, and verified example companies. Use for "who are our buyers", "ICP research", "audience profile", "buyer personas".

## What starts it

- By hand, any time: from the console, the API or the CLI.

## Inputs

| Name | Type | Required | What it is |
| --- | --- | --- | --- |
| `subject` | string, 3 to 400 characters | yes | A company website or name (the run profiles its buyers), or a description of a group (the run profiles that audience). Example: `https://pipewright.example, Independent veterinary clinics in the UK`. |
| `offer` | string, 0 to 1000 characters | no | The product to profile buyers for. Blank: the company's main offer, or the audience's general needs. |
| `geography` | string, 0 to 120 characters | no | A country or market to limit evidence and examples to. Blank: the company's main market, or where the evidence points. |

## What it needs

- Harness: pi, claude-agent, codex
- Model: balanced tier, medium effort
- Tools: read, web_search, web_fetch
- Computer: Wherever your AgentRun puts a new agent
- Keys: A model key
- Learning: off

## Sample output

Input:

```json
{
  "content": "Today is 2026-11-14. Profile our buyers.\n\ntallyhive.example, harrowgate-foods.example/careers and fsa-notices.example are not reachable from this run. The pages\nbelow are what their URLs return today; use them in place of fetching, and search normally for anything else.\n\n<captured-page url=\"https://tallyhive.example/\" captured=\"2026-11-14T10:00:00Z\" untrusted>\n<h1>Allergen labels that never go out of date</h1>\n<p>Tallyhive keeps recipe, ingredient and allergen data in one place and prints compliant labels for food manufacturers\nand central kitchens in the UK. From GBP 350 per site per month. Customers include Harrowgate Foods and Pennine Bakehouse.</p>\n</captured-page>\n\n<captured-page url=\"https://harrowgate-foods.example/careers/technical-manager\" captured=\"2026-11-14T10:01:00Z\" untrusted>\n<h1>Technical Manager, Harrowgate Foods (Leeds)</h1>\n<p>Posted 2026-10-20. Reports to the Operations Director. You will own allergen management, label approval and BRCGS\naudit readiness across two sites, and manage the technical budget. Experience with specification systems required.</p>\n</captured-page>\n\n<captured-page url=\"https://fsa-notices.example/2026/precautionary-allergen-labelling\" captured=\"2026-11-14T10:02:00Z\" untrusted>\n<h1>New guidance on precautionary allergen labelling</h1>\n<p>Published 2026-07-01. The Food Standards Agency has issued updated guidance on precautionary allergen labelling\n(\"may contain\" statements) for food businesses in England, Wales and Northern Ireland, expecting businesses to apply it\nfrom 2027-01-01.</p>\n</captured-page>",
  "subject": "https://tallyhive.example"
}
```

Output:

#### Summary
Prioritise UK food manufacturers with cross-site technical responsibility, then bakery groups and central kitchens. Tallyhive states that it centralises recipe, ingredient and allergen data and prints labels, starting at GBP 350 per site per month. These are vendor claims, not independently tested capabilities. [S1]

The strongest initial contact is the Technical Manager. The supplied vacancy explicitly combines problem ownership with technical-budget responsibility; executive sign-off remains unknown. [S2]

#### Segments
| Priority | Segment | Why it fits |
|---|---|---|
| 1 | Multi-site food manufacturers | Two-site allergen, label and audit responsibilities sit with a budget-owning technical role. [S2] |
| 2 | Bakery groups | Central information, local printing and frequent product changes appear in a competitor case study; an own-site source confirms geographic and operating fit. [S6][S9] |
| 3 | Central kitchens and caterers | Competitor evidence shows central data management, site variations and integration needs. [S5] |

These priorities are recommendations. No defensible market counts or employee-size bands were established.

#### Buyer and user roles
**Technical Manager:** evidenced budget holder and inferred champion. **Operations Director:** inferred executive sponsor because the technical role reports there. **General Manager:** potential bakery champion, evidenced discussing the workflow. **Recipe/menu systems product manager:** potential integration champion. **Site chefs and printing staff:** operational users. Only technical-budget ownership is explicit; do not assume final purchasing authority. [S2][S5][S6]

#### Jobs to be done
Recommended discovery focus: maintain consistent ingredient and allergen information; approve labels; reduce manual data transfer; coordinate central information with outlet printing; and support audit preparation. These jobs follow the offer, vacancy and competitor evidence, rather than a representative buyer survey. [S1][S2][S5][S6]

#### Pains in their words
No accessible independent buyer posts or reviews were obtained, so no pain quotes or voice counts are reported. Capterra blocked the fetch. [S4] Two vendor-selected case studies discuss manual data handling, rapid changes and obsolete packaging stock; they inform hypotheses but are not independent pain validation. [S5][S6]

#### Buying triggers
Recommended signals include a new cross-site technical appointment, audit preparation, new products, ingredient changes and expansion of distributed printing. Except for the dated vacancy and historical rollout, purchase intent is inferred. [S2][S5][S6]

The supplied notice reports updated PAL guidance published on 2026-07-01 with expected application from 2027-01-01. Retain that as a reported planning signal. The official PAL page inspected supports risk-assessment-based decisions but does not corroborate this timetable; do not present it as a verified legal deadline. Its stated scope is England, Wales and Northern Ireland, not all UK jurisdictions. [S3][S7]

#### Objections
Likely objections are per-site cost, migration effort, compatibility with specification/menu systems, and confidence in data controls. These are inferences to test. A competitor's bespoke integration indicates potential complexity, not a Tallyhive capability. Software must not be positioned as replacing cross-contact assessment. [S1][S2][S5][S7]

#### Where they gather
| Verified place | Recommended use |
|---|---|
| Food & Drink Federation food-safety and labelling hub | Test educational content and relevant partnership opportunities; audience size and acquisition performance are unknown. [S8] |

#### Tools they trust
LabelLogic Live adoption is documented in vendor-selected case studies, not independent trust research. The vacancy requires specification-system experience, while the catering case describes an unnamed menu-management system. These are incumbent workflows to discover, not assumed replacement targets. [S2][S5][S6]

#### Example companies
| Company | Evidence and qualification |
|---|---|
| Harrowgate Foods | Supplied own-site vacancy supports two-site fit; Tallyhive reports it as a customer. [S1][S2] |
| Jenkins Bakery | Own site confirms South Wales branches; competitor case study supports labelling-workflow fit. Not a confirmed Tallyhive prospect. [S6][S9] |

#### Messaging implications
Recommendations: demonstrate an ingredient change through to printing; lead with the stated shared-data offer; quantify the buyer's own administration and label-stock costs. These angles rest on vendor material and role evidence, not independently validated buyer language. Avoid absolute compliance promises and unverified deadline claims. [S1][S2][S5][S6][S7]

#### Gaps
Independent buyer language, market sizing, broader competitor customers, purchasing authority, integration capability and channel economics remain unverified. Research stopped at the prescribed search/fetch limits; partial pages and blocked reviews were not treated as substantive evidence. [S4][S10][S11]

## Tested

Passes 3 of 3 eval cases on pi, 1 of 3 on Claude Code and 2 of 3 on Codex. Not yet tried on the live web.

3 eval cases: company-mode, injection-in-thread, thin-audience.

## Source

- Package: https://github.com/Parcha-ai/agentrun-templates/tree/242f29da8317fbe51d0cce4628bdb5ab61d7ccc9/templates/audience-research
- Publisher: Parcha-ai
- Author: Parcha Labs
- Version: 0.1.0
- License: Apache-2.0
- Category: Marketing
- Page: https://marketplace.agentrun.ai/templates/audience-research/
