Pick a product, load a sample company and watch how Credit Atlas collects the evidence, applies a credit policy and returns a decision-ready view. A preview of the interface, with sample data.
Both products run on the same registry-backed evidence. Full Credit Scoring adds bank transaction intelligence, your credit policy and the AI management overlay.
Product 01
A structured first-pass view of a company built from official registries, compliance sources and public filings. No bank data required.
KYC, PEP and sanctions screening
Registry-backed entity and ownership checks
Tax arrears, court filings and insolvency history
Preliminary intelligence score with prioritised red flags
Typical run: under a minute
Product 02
The complete underwriting workflow: external intelligence plus categorised bank transactions, cash-flow analysis, your credit policy and a committee-ready recommendation.
Everything in External Intelligence
Proprietary Categorization of 12 months of transactions
CFADS, debt service cover and policy-aligned scoring
Approve, refer or reject pathway with the AI overlay
This is the intake your analysts or your origination system would submit. In this preview the fields are display-only: load one of the sample companies to fill them and run the flow.
Load a sample company
No data is collected or sent anywhere. The preview runs on prepared sample evidence.
Step 3 · Processing
Seven steps, every one of them traceable.
Each stage records what it collected, what it checked and what it concluded, so every number in the report can be followed back to its source.
1
Data Collection
Official registries, compliance sources and partner networks.
2
Data Processing
Validation, normalisation and structuring for analysis.
3
Enrichment & Categorization
Financial, operational and behavioural signals categorised.
Transaction categorisation needs bank data. Included in Full Credit Scoring.
4
Risk Scoring
Your thresholds and factor weights applied to the evidence.
5
AI Management Overlay
Key insights and exceptions surfaced for analyst review.
The AI overlay reviews the full credit file. Included in Full Credit Scoring.
6
Output & Delivery
Decision-ready report and API payload delivered.
7
AI Agents
A team of specialised agents searches, verifies, analyses and projects, then streams the verdict into your interface.
Assessment logWaiting for input
Run an assessment to see each step report what it found.
Step 7 · AI AgentsInput → risk process → a team of specialised agents. Each one streams its result straight into your interface.
InputCompany details and bank data from your client
POST /v1/assessments
Risk processCredit Atlas evidence and policy engine
Your interface
Agents working…
Risk score · 0/100
Committee summary ready, opening the report
Step 4 · Results
Decision-ready output, with the reasoning visible.
· ·
No assessment yet. Choose a product, enter a company and run the assessment to see the report here.
In production the agent runs on Claude by Anthropic inside your environment. It reads the same evidence your analysts see, explains the drivers behind the score and drafts the committee summary. It never takes the decision.
Grounded in the evidence trail
Every answer cites the registry record, transaction category or policy rule it relies on.
Analyst stays in control
The agent surfaces exceptions and what-if scenarios. Approval, referral and rejection remain your team's call.
Preview note: the answers here are prepared from the sample assessment. No data is sent to any model.
Credit Atlas AgentAgent · analyst assistant
Run an assessment and I will walk you through the drivers, the weakest evidence and what would change the decision.
Integrate the same workflow via API
Everything you just saw is one API away: create an assessment, receive a signed callback and read the policy outcome with every check, ratio and reason code behind it. One contract for every market.
A pilot connects Credit Atlas to your credit policy, your data sources and your workflow. Banks, leasing companies, fintech lenders and advisory teams start with one product and expand from there.
The companies, scores, findings and agent answers shown here are prepared sample data. No registry lookups, transaction analysis or AI reasoning run on this page. To see Credit Atlas working on real cases, with your data and your credit policy, book a live demo with our team.
External IntelligenceFull Credit ScoringRunning step {n} of 7Completed in {t}sWaiting for inputLoad one of the sample companies first, then run the assessment.Registry lookup {country} · entity {reg} found · status: activeSanctions, PEP and adverse media lists queried · {hits} hit(s)Tax authority and court records retrieved · {filings} filing(s)Bank statements ingested · {months} months · {tx} transactionsIdentifiers validated · VAT {vat} matches registry recordFinancial statements normalised · {years} reporting periodsTransactions de-duplicated and currency-normalised{tx} transactions categorised · {cats} lending categories · {conf}% confidenceRevenue, CFADS and debt-service series builtOwnership and group structure resolved · {ubo} beneficial owner(s)Transaction categorisation skipped · no bank data in External IntelligencePolicy "{policy}" applied · thresholds and factor weights loadedFactor scores computed · financial {fin} · business {bus} · behavioural {beh} · compliance {com}Overall score {score}/100 · band: {band}AI overlay reviewed the file · {flags} finding(s) prioritisedExceptions routed to analyst queueAI overlay skipped · available in Full Credit ScoringReport generated · decision object ready · pathway: {pathway}API payload delivered to the origination endpointLow RiskMedium RiskHigh RiskApproveReferRejectOverall Risk ScoreIntelligence ScoreDecision pathwayPreliminary viewKey MetricsRegistry ProfileLow level of credit risk based on the analysed data. Evidence suggests a stable financial footing and reliable operations.Moderate credit risk. The business is viable but several drivers sit close to policy thresholds and warrant analyst review.High credit risk. Material weaknesses in the evidence outweigh the positives under the selected policy.Meets the approval threshold of the {policy} ({min}+). No blocking exceptions.Falls in the referral band of the {policy} ({min}-{max}). Analyst review required before a decision.Below the {policy} floor of {min}. Blocking exceptions present.Preliminary screening only. A lending decision requires bank data and a credit policy (Full Credit Scoring).Standard SME policyConservative policyGrowth lending policyScreening rulesRevenue (12M)CFADS (12M)Debt service coverNet debt / EBITDAOverdraft utilisationDays past due (max)Registered sinceShare capitalDirectorsBeneficial ownersTax arrearsCourt filings (36M)NoneCompany registry ({country})Entity status, directors, share capitalVIES / tax authorityVAT number validityEU / UN / OFAC listsSanctions, PEP, adverse mediaCourt and insolvency registersProceedings, enforcement, insolvencyTax authorityArrears and payment statusBank statements{months} months categorisedCredit policyThresholds and factor weightsVerifiedClear{n} finding(s)AppliedNot includedYouThere is no assessment to discuss yet. Run one above and ask me again.In this demo I answer four scripted questions about the file. Pick one below, or run a pilot to ask the live agent anything about your own cases.{company} scores {score}/100 ({band}). The strongest factor is {best} at {bestv}/100; the weakest is {worst} at {worstv}/100. Under the {policy} this lands on the "{pathway}" pathway. The drivers: {drivers}.The score is {gap} points below the approval threshold ({min}). Resolving the top finding ("{flag}") would lift the {worst} factor by an estimated {delta} points, which moves the file to "{next}". A smaller facility or a longer term would also improve debt service cover from {dscr}x.The file already meets the approval threshold ({min}) with {gap} points of headroom. It would drop into the referral band if the facility rose above roughly {maxamount} EUR at this term, or if the "{flag}" finding escalated.External Intelligence gives a preliminary view only. To reach a lending decision, run Full Credit Scoring: bank data would confirm cash-flow capacity and the credit policy would set the approve, refer and reject thresholds.The weakest evidence is {worst} ({worstv}/100). {flag} Supporting sources: {sources}. I would ask the analyst to confirm this point before the committee.Committee summary: {company} ({country}, reg. {reg}) requests a {amount} EUR facility over {term} months. Overall score {score}/100, {band}, pathway "{pathway}" under the {policy}. Key findings: {flags}. Evidence: {sources}. Recommendation drafted for analyst review; decision remains with the committee.Screening summary: {company} ({country}, reg. {reg}). Intelligence score {score}/100, {band}. Findings: {flags}. Evidence: {sources}. Preliminary view only; no lending decision without bank data.Reviewing the credit file for {company}. Registry evidence verified, {policy} applied. Overall score {score}/100, {band}: {tail}Collects registry, court, tax and sanctions evidenceCross-checks identifiers, filings and ownershipCategorises transactions, applies your policy, scores the fileProjects 12-month cash flow and stress scenariosDrafts the committee summary and flags exceptionsPulls registry, court, tax and insolvency recordsScans the web for debts, adverse media and reviewsScreens sanctions, PEP lists and beneficial ownersReads payment behaviour and trade signalsScores the screening and drafts the summary{sources} registry sources, {filings} court filing(s){debts} debt notice(s) online, {reviews} reviews at {rating}/5Sanctions and PEP: {hits} hit(s), {ubo} beneficial owner(s)Payment behaviour: {behaviour}on time in trade referencesoccasional delays, up to {dpd} days{dpd} days past due, returned paymentsScreening complete for {company}. Registry, court and web sources reviewed, compliance lists checked. Intelligence score {score}/100, {band}: {tail}Intelligence ScoreDebts found onlineNone found in debt registers or enforcement notices{n} enforcement notice(s) and unpaid-invoice listingsOnline reviews{rating}/5 from {reviews} reviews, {sentiment}mostly positivemixedmostly negativeAdverse mediaNo negative coverage in the last 24 months{n} article(s): reports of unpaid subcontractorsCourt and insolvencyNo proceedings on record{n} filing(s) in the last 36 monthsWeb presenceDomain active since {year}, identity consistent across sourcesPayment behaviourQueuedWorkingDone{sources} sources collected, {hits} sanctions hit(s)Identifiers verified, {ubo} beneficial owner(s) resolved12-month projection: DSCR {dscr}x, {stress}x under stressProjection needs bank data, available in Full Credit ScoringOperating costsPayrollDebt serviceTax and leviesRetained cashJanFebMarAprMayJunJulAugSepOctNovDecno blocking exceptions left open.analyst review required before a decision.blocking exceptions present, decline recommended.Agent verification started · streaming events to the clientAgent verdict delivered · {score}/100 · report opened in the interfaceFinancialBusinessBehaviouralComplianceEntity active, filings up to date, directors verified against the registry.No sanctions, PEP or adverse media matches on the entity or its owners.No tax arrears recorded; VAT number valid and matched to the entity.Consistent monthly inflows; CFADS covers proposed debt service comfortably.Top customer accounts for 41% of inflows over the last 12 months.Revenue grew 38% year on year while margins compressed; working capital under pressure.Overdraft utilisation above 80% in 4 of the last 12 months.One civil claim filed in the last 36 months, settled; no enforcement.Beneficial owner resides outside the EU; enhanced due diligence recommended.Tax arrears of 142,000 EUR recorded; payment plan not confirmed.Negative equity in the latest filed accounts; two consecutive loss-making years.Two enforcement proceedings open; one insolvency petition withdrawn in the last 24 months.Returned payments and 47 days past due on an existing facility.Adverse media: reports of unpaid subcontractors in regional press.
We use analytics cookies only with your consent, to see how the site is used. Nothing is loaded or stored before you decide.