Public case study · no identifying data

How an MCP client researched customer-support agents.

An external MCP client used Agentery to identify a customer-support market, establish its pricing, narrow providers by budget and integrations, and compare five finalists. Observed 24–25 July. The figures below are that session's historical evidence, not today's current prices — see the live index on the homepage.

Connect Agentery MCP Browse the same data
What happened

One task became a market, a price and a shortlist.

The client never had to name a niche or read a pricing page. Agentery resolved the market, returned tier ranges, and shortlisted providers that actually belong to it — then compared each one against a compatible cohort.

The session also exposed real weaknesses: an early search returned generic infrastructure listings, and a high-intent query filtered on five ticketing integrations returned nothing. Those failures directly changed the product.

01
Describe the buying problem

The client asked broadly for customer-support agents and their pricing — a small business wanting help under roughly $99/month, ideally with Zendesk, Intercom or Freshdesk.

02
Identify the market

Agentery mapped the free-text task to the relevant customer-support niche (Customer Support — Tier 1) rather than guessing from keywords, and returned the market's price ranges by buyer tier.

03
Add constraints

The search was narrowed by a monthly budget, ticketing needs and preferred integrations — the real constraints a buyer cares about.

04
Shortlist providers

Niche-first procurement search returned providers from inside the matched market instead of generic directory noise; where an integration wasn't indexed, that preference was relaxed rather than returning nothing.

05
Compare finalists

Five providers were compared using their OWN observed price against a genuinely comparable market cohort — same buyer tier, same billing unit, standing (not promotional) price.

The five finalists, compared like for like.

Each provider's own observed price beside a compatible market cohort. Where no comparable cohort existed, Agentery said so rather than inventing a position.

Chat Thing$11.67/mobelow market
CrafterQ$29/mobelow market
Recursive Support$49/mobelow market
Deskcrew Support$24.17/mobelow market
Wonderchatinsufficient comparables

Price positions use the same buyer tier, billing unit and standing price as the market median — never a cheapest-plan-versus-all-tier comparison.

What it changed

The failures shipped fixes.

  • Search became niche-first: it resolves the buyer's market and ranks inside it, instead of matching keywords across the whole directory.
  • Integration filters became soft preferences: an unindexed integration relaxes the constraint rather than returning zero results, and is labelled “unverified”, never “absent”.
  • Comparisons became like-for-like: each provider sits beside a genuinely comparable market cohort.
Run it yourself

Try the same research from your MCP client.

Connect Agentery and paste the starter prompt to reproduce this exact workflow on any market.

Starter prompt

Find customer-support agents under $100/month. Compare their pricing and tell me which have confirmed Zendesk, Intercom or Freshdesk integrations.

Claude Code
claude mcp add --transport http agentery https://agentery.com/api/mcp
Adds Agentery as a remote streamable-HTTP MCP server.
Published safely: no caller hash, gateway profile, precise timestamps, geography, account details or raw telemetry. The external client is an MCP caller; we do not claim it is a specific person or organisation.