Does action-space constraints data qualify? A nine-criterion scorecard for introducers

Short answer

An introducer scoring action-space constraints records should check nine things: company size, documented history, rights, a concrete list of restricted actions, whether software enforces the limits, whether escalations and approvals are logged, whether those records link, whether limit changes are dated, and who can authorize a licensing talk. Each scores strong, borderline or weak. A company strong on the records criteria and not weak on any gate is worth introducing.

Visual briefFull-size diagram
Nine qualification criteria are ticked one by one beside a semicircular gauge with weak, borderline and strong bands; the needle swings and settles at 16 of 18, in the strong band, for a fictional payments software company.
Illustrative score for a fictional 210-person payments software company. The scoring scheme is a partner aid, not SourceX's review.

This page is part of the guide to action-space constraints at a support desk, where a front-line agent can propose a refund but cannot issue one. Its job is a qualification scorecard. Use it after a conversation with the head of support or operations, filled in from answers, never from files.

How the scorecard works

Score each criterion 2 for strong, 1 for borderline and 0 for weak, for a maximum of 18. Three criteria are gates: company size, rights and the authorized contact. A weak score on any gate means the company is not ready for an introduction, whatever the total.

TotalReadingWhat to do
14 to 18, no weak gateStrong fitIntroduce the decision-maker
9 to 13, no weak gateBorderlineIntroduce, stating the weak criteria plainly
0 to 8, or any weak gateNot yetLeave the owner a list of what to check

The scheme is a partner aid built for this page. SourceX does not score companies this way, and a total is not a forecast of anything.

The nine criteria, with strong, borderline and weak answers

The first three criteria apply to any company. The middle five are specific to action-space constraints: they test whether the limits on what staff could do were recorded, or only described. The last confirms who can say yes.

CriterionStrong (2)Borderline (1)Weak (0)
Company size (gate)Clearly 50+ full-time employees at peakNear 50, with many contractors to excludeUnder 50 full-time employees
Operating historyFive or more years of records in the current systemsTwo to four years, or a recent helpdesk migrationUnder a year of retained records
Rights (gate)Company holds its own tickets and payment logsOutsourced provider holds tickets, or customer terms limit reuseRecords controlled by a vendor with no clear rights
Restricted actionsNamed list: refunds, credits, cancellations, price overrides"Bigger refunds need a manager"No restricted actions anyone can name
EnforcementSoftware blocks the action for the roleMix of software blocks and trained rulesRules exist only in training material
Escalation recordsProposal, escalation and approval each loggedApprovals logged, proposals notEscalations happen in chat or by phone
LinkageTicket ID carried through to the approval and the refund IDLinks exist for recent years onlyRefunds cannot be traced to tickets
Configuration historyEach threshold change dated, old values knownCurrent values known, changes roughly rememberedNobody knows what the limits used to be
Authorized contact (gate)CEO named, with the COO or head of customer operations involvedOwner named but not yet awareNo one can authorize a discussion

The enforcement row matters because research benchmarks draw the same line. τ-bench gives agents a domain policy and notes that some rules are enforced by the software while others depend on the agent applying them correctly. Records of both kinds can be useful, but they show different things: a blocked button proves the limit held, a trained rule shows whether people kept it.

Payment software makes strong answers easier to recognize. Stripe documents a tiered setup in which tier 1 agents get a role with refund permission removed (custom roles are in private preview), and its two-party approvals, in public preview, can hold a refund above an amount until a different team member approves it. A company using controls like these, in any system, is likely to score well on enforcement and escalation records.

Scoring three fictional support desks (Illustrative)

All three companies are fictional, and the scores are invented to show how the scorecard reads.

Criterion210-person payments software firm130-person online retailer45-person digital agency
Company size220
Operating history212
Rights212
Restricted actions211
Enforcement210
Escalation records210
Linkage110
Configuration history110
Authorized contact212
Total16, strong fit10, borderline7, not yet (size gate)

The payments firm blocks refunds above $500 for tier 1 in its billing tool, logs every approval, and lost some linkage in a 2022 helpdesk move. The retailer's support runs through an outsourced provider, limits are mostly trained, and approvals are logged but proposals are not. The agency has good owners and rights but is too small to qualify, so its score ends there. The graphic on this page plays the payments firm's score: nine ticks and a needle that settles at 16 of 18.

Five questions that move a score

Ask these of the head of support or operations and write down answers, not files.

  1. Which actions are restricted by role? A concrete list moves "restricted actions" to strong. "Only managers can do the big stuff" stays borderline.
  2. Is the restriction enforced by the software or only by training? This decides the enforcement score directly.
  3. When someone cannot act, what do they do, and is it logged? If the honest answer is "they message a manager on chat", escalation records score weak and the opportunity is mostly theoretical for now.
  4. Have the limits changed, and is the old configuration known? A threshold that moved from $250 to $500 last year changes which escalations were required. Dated changes score strong.
  5. Who owns these records and can authorize a licensing discussion? The COO or head of customer operations usually owns them, with finance involved because refunds move money; the CEO authorizes the discussion.

Before you ask, it helps to confirm the client is comfortable being introduced at all; see documenting client consent for a referral. For company-level basics that apply to every topic, the company qualification checklist is shorter.

Where the scorecard stops

The scorecard reads answers; it cannot verify them. It does not show how many escalations carry complete records, whether customer data can be handled, or whether any buyer wants this task type. Data companies do build business environments around permission models (Toloka's agent environments page, checked October 18, 2026, describes mockups that replicate real tool schemas, workflows and permission models), but that is evidence of category interest, not of demand for a given company's tickets.

The company test itself comes from the program terms: 50+ full-time employees at peak (contractors excluded), documented operating history, rights to license and an authorized contact. Companies in the U.S., Canada and the EU qualify, as do companies in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia when at least 70% of the data is in English.

If the score is borderline because records are patchy, the evidence gap analysis shows how to describe each gap. The wider record family is set out under support escalation processes, and a business introducer's guide to operational data licensing covers the conversation with the owner.

Next step

Refer a company. No cap on rewards. Default reward on a company's first qualifying transaction: the greater of $50,000 plus 5% and 15% of Actual Supplier Payout, with no cap; further qualifying transactions first agreed within 12 months of the first Closing pay 15%. Eligibility and payment conditions apply, including the USD 100,000 Minimum Transaction, and a signed Schedule A or addendum can differ. A lead, meeting or signature alone does not trigger payment.

If a client scores 9 or more with no weak gate, register as a partner, confirm the company against who qualifies, and read the program terms before you introduce the decision-maker.

  1. Step 1Share your linkSend your personal link to a company you know.
  2. Step 2Company appliesThe company applies itself at /apply.
  3. Step 3Company signs and is paidThe company signs, delivers its data and is paid by SourceX under its agreement.
  4. Step 4You get your rewardthe greater of $50,000 plus 5% and 15% of the actual supplier payout, with no cap. A reward becomes payable after SourceX has paid the referred company under its signed agreement and the payout conditions in the program terms are met, including finance approval and your tax and payment details.

Common questions

Is a high score a prediction that SourceX will accept the company?

No. The scorecard helps a partner decide whether an introduction is worth making. Acceptance depends on SourceX's own review of rights, record quality and buyer interest, none of which a partner can see from outside.

What if the company scores well but uses an outsourced support provider?

Score rights as borderline and note who holds the tickets. Many outsourcing contracts leave the records with the client company, but that has to be confirmed by the owner, not assumed.

Should I ask to see the permission settings to score enforcement?

No. Ask whether the system blocks the action or whether staff are trusted not to take it, and write down the answer. Screenshots and exports stay with the company.

Sources

  1. docs.stripe.com/get-started/account/teams/custom-roles
  2. docs.stripe.com/account/approvals
  3. arxiv.org/abs/2406.12045
  4. toloka.ai/virtual-environments-rl-gyms
  5. refer.sourcex.si/terms

Free resources

Know a company with valuable proprietary data?

Become a referral partner, get your link and introduce the company’s CEO (in an owner-run company, usually the owner). Referral partners can join from most countries. We can't work with partners in countries or regions under U.S. sanctions or on U.S. restricted-party lists. The companies you introduce should be based in the U.S., Canada and the EU, or in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia with at least 70% of their data in English.

Refer a company →

I own a business

Explore licensing your company's data directly to SourceX for AI training and evaluation. Start a short assessment; no uploads needed.

Start an assessment