Telegram bot · 3PL provider

3PL provider → exporting manufacturers in Finland: what one real Prosp run returned

By the Prosp team · run on 2026-10-03 in production

Run date

2026-10-03

Job duration

9 min 52 s

Results came from

Live web search (2 of 8 from the knowledge base)

Delivered

8 of 8 requested

Rejected candidates

53

Discarded after target

1

The scenario

A third-party logistics provider wants Finnish manufacturers that export their products and would outsource warehousing and transport. The buyer is a profile we set up to test this market, not a customer. The run is real, and every company on this page came back from it.

Buyer

A third-party logistics company offering warehousing and export transport. An illustrative profile, not a named customer.

Target

Small and medium Finnish manufacturers, including furniture makers, that export their products.

The exact request

Typed into the Telegram bot:

1. You

   /find

2. Bot

   Let's set up a discovery.
   
   Step 1/7 — Description (required):
   Type a short description of the companies you want to find …

3. You

   Find Finnish manufacturers that export their products and need third-party logistics

4. Bot

   Step 2/7 — Industry categories (optional): …

5. You

   (taps) Manufacturing and Furniture Manufacturing, then Done ✓

6. Bot

   Step 3/7 — Business models (optional): …

7. You

   /skip

8. Bot

   Step 4/7 — Locations (optional): …

9. You

   (taps) Finland, then Done ✓

10. Bot

    Step 6/7 — Company sizes (optional): …

11. You

    (taps) Small, Medium, then Done ✓

12. Bot

    Step 7/7 — Keywords (optional): …

13. You

    exporter, production, valmistaja

14. Bot

    Who are we finding prospects for? …

15. You

    (taps) Skip for now

16. Bot

    Review your discovery: …

17. You

    (taps) Submit

18. Bot

    ℹ️ This search looks saturated — it's unlikely to find new companies right now.
    No credits have been charged.
    Tap “▶️ Run anyway” below to run it anyway (100 credits).

19. You

    (taps) ▶️ Run anyway (100 credits)

How this run was actually started: through Prosp's internal job endpoint, with exactly these criteria and a target of 8 companies, under an internal exploration account whose results never reach a CRM. That path does not charge credits. The transcript shows the same criteria entered in the Telegram bot's /find wizard; bot messages are shortened to the first line of each step, and step 5 (regions) does not appear because Finland has no regions to pick. The bot's reply to the hold is reproduced from the bot's English message templates, not captured from this run; the raw response this run actually received is shown below. The bot always asks for 10 companies and has no setting for the count, so the same request from Telegram targets 10, not 8.

The first attempt was held before it started

The first submission, with the same criteria but without acknowledge_low_yield, was held by the pre-flight check: ICPs like this one have historically delivered few results. No job was started and nothing was charged. It was then re-submitted with "acknowledge_low_yield": true (in the bot, the Run anyway button does this), ran, and reached its target, so the prediction was too pessimistic this time.

HTTP/1.0 409 Conflict

{
  "status": "low_yield_prediction_warning",
  "message": "Heads up: ICPs like this one have historically delivered few results (moderate). Suggestion: refine the industry / keywords — similar ICPs surfaced off-target industries. No credits were charged and no job was started. To run it anyway and be charged, re-submit with \"acknowledge_low_yield\": true in parameters.",
  "suggestion": "refine the industry / keywords — similar ICPs surfaced off-target industries",
  "acknowledge_param": "acknowledge_low_yield",
  "request_id": "0eb298fef92144eeb3236d203096ab7c"
}

What came back

Checking the registry ids

Why candidates were rejected

Not counted above: 1 more candidates were still being processed when the target was reached. They were discarded without being evaluated (terminal_job_discarded), so they say nothing about quality either way.

What went wrong

• Oras Group is stored under the name "Orasgroup website", a web-page title. By its own description it owns the Oras, HANSA, Gustavsberg and Vatette brands, which makes it a large group, but its company_size is empty, so the small/medium preference could not exclude it. Size is a preference, not a hard filter.
• Lappset Group and Stera Technologies also have no company size.
• Fit for a 3PL looking for exporters: Lapelland (cabin and sauna trailers for consumers in Finland) and Puuartisti (custom woodwork, mainly within Finland) describe a domestic market.
• 4 of 8 rows have no registry id: Lapelland's register match was not confirmed, and the lookup found nothing for Oras Group, Furninova and YK Collection.
• In the Telegram bot, this low-yield hold is worded as "This search looks saturated", which is the text for a different warning. The API response above is the accurate one.

How to read this run

Before spending anything, Prosp compares a new request with similar past requests. For this one it predicted a low yield and held the job: HTTP 409, no job, no charge, and a suggestion to refine the industries or keywords. You can take the suggestion or run anyway; we ran anyway. The prediction turned out too pessimistic: the run reached its target of 8. A hold like this is worth reading rather than clicking through, because it is the only point at which changing the request costs nothing.

The delivered companies make playground and park equipment (Lappset Group), contract-manufactured industrial assemblies (Stera Technologies), bathroom and kitchen fittings (Oras Group), office acoustic pods (Taiga Concept), sofas and beds (Furninova), designer chairs (YK Collection), cabin and sauna trailers (Lapelland) and custom woodwork (Puuartisti). For a logistics provider, the ones whose delivered descriptions mention customers outside Finland are Lappset, Stera (Finland and Estonia), Oras, Taiga Concept, Furninova and YK Collection.

The rejections spread across many reasons, led by country_not_in_target (7), industry_not_in_target_precrawl (5), industry_contradiction (4) and enterprise_too_large (4).

What this market can and can't verify

• Registers: yes, free and fast. The Finnish Trade Register publishes open data through PRH. On 2026-10-03 we looked up each of the four Business IDs Prosp delivered. All four matched on current name and city and were registered. Each lookup took under a second; the links in the table above go straight to the record.
• Gaps: four rows came back without a Business ID. For Lapelland the register match was not confirmed and was withheld; for Oras Group, Furninova and YK Collection the lookup found nothing.

Cost and time

• The job took 9 min 52 s. At submission the API estimated a median of 565 s.
• The held first attempt cost nothing. A discovery that runs costs 100 credits through the REST API, the MCP server or the Telegram bot. This run was started on an internal account and was not charged.

How to reproduce this run

• Telegram: open the bot, buy credits with /buy, then run /find and enter the criteria from the transcript. If the bot holds the request, tap Run anyway. See the Telegram guide. The bot will ask for 10 companies.
• REST API: send the criteria to POST /api/v1/discoveries; on a 409, re-send them with "acknowledge_low_yield": true. See the API reference.
• ChatGPT or Claude: connect the MCP server and ask for the same companies.

A repeat will not return exactly these eight. The web changes, and each run adds to Prosp's knowledge base, so a later run may be served partly from companies found by earlier ones.