Amabis Data Quality
Analytics & DataClean, correct, and validate email addresses, phone numbers, postal addresses, and person or company identities with Amabis. Agents run contact records through real-time quality checks before they reach a CRM or a send list.
What This Integration Enables
Amabis treats the French market as a domain rather than a locale setting, and that is what separates it from the general-purpose validators. The postal address surface is not one endpoint but two approaches. Normalize Postal Address restructures a full address against postal norms and returns it split into house number, street type, street body, postal box, sub locality, postal code, city, state and country. It carries a quality code expressing the probability the mail will not arrive, and optional geocoding adds IRIS, FANTOIR, canton and priority district references. Alongside it sits a three-step guided funnel, Search Cities And Postal Codes to Search Streets to Search Buildings, where each step hands the next the geographic keys it needs. A shopper ends up selecting a registered building rather than typing into a free text box.
The same depth runs through the other modules. Check Email Address opens a dialogue with the domain mail server to confirm the mailbox actually exists and returns a verdict as specific as MAILBOX_FULL or TRASH_MAIL, while Normalize Email Address does syntax-only cleanup at much higher speed for volume work. Normalize Phone Number returns every common layout at once with line type and a suspicious flag for numbers that are syntactically valid but look fabricated. Normalize Identity untangles a jumbled name field into its parts and detects when two people, such as a couple, were entered into one box. Search Companies and Order Company Sheet reach the French company register for verified SIRET and SIREN data. FlowRunner's connectors are built and verified against the vendor's official API, so each of those verdict codes reaches the agent exactly as Amabis defines it.
Without FlowRunner
With FlowRunner
Use Case Scenarios
Enriching a French B2B list from the directory
A sales team has company names and addresses but no phone numbers. The agent calls Search Phone Numbers for each record, which either resolves a name and address to a listed number or, given a number, runs the lookup in reverse to recover the subscriber. Every result carries a prospecting objection flag, and the agent reads it before anything else. Records with the flag set are separated immediately. The rest are normalized into a consistent format, checked against the suspicious flag, and written into Pipedrive with the source of the number recorded on the field, so a rep can always see that it came from a directory rather than from the prospect.
Guided address capture at a French checkout
A shopper starts typing a postal code. The agent calls Search Cities And Postal Codes and returns ranked suggestions with population and INSEE code, carrying the CLEGEO and CLEAGG keys forward. When the shopper picks a city, Search Streets uses those keys plus the partial street text to return real streets with their CLEVOIE identifiers. When they pick a street and number, Search Buildings lists the buildings registered there, so the apartment complement is a selection rather than a typo waiting to happen. The order lands in the warehouse with an address the postal service already agrees exists.
Repairing an imported contact list before a send
A list arrives from an event with names in one column and emails of unknown quality. The agent runs Normalize Identity across the name column, which splits titles from first and last names, flags where the two were swapped back, reports probable gender, and returns a names array when a single field held a couple. It then runs Check Email Address on each address, which opens the mailbox dialogue rather than checking syntax alone, and routes anything returning BAD_USERNAME, BLACKLISTED or TRASH_MAIL out of the send list in Brevo. Get Pack Consumption runs on a schedule alongside it, so the pack is topped up before the credits run out mid-list.
Human-in-Loop Highlight
Search Phone Numbers returns something most enrichment APIs do not: a no_prospecting flag showing that the subscriber has objected to commercial prospecting. That flag is the whole reason this action needs a gate. The number was not given to you by the person it belongs to. It was found in a directory, along with a recorded statement that they do not want to be contacted commercially, and the record still looks exactly like a lead by the time it reaches a campaign tool. An agent that enriches a list and hands it straight to a dialer has quietly turned a legal signal into a marketing asset. So the FlowRunner agent splits the result and refuses to move: "Directory lookup on 1,240 companies returned 890 numbers. 214 of them carry a prospecting objection and have been suppressed. 47 more were found by reverse lookup from a number we already held rather than supplied by the contact. Confirm the suppression list and tell me whether the 47 have a lawful basis before I load the remaining 629." The campaign owner answers, the agent proceeds only with what was approved, and the suppression is on the record with a name against it. This is human-in-the-loop at the point where a data verdict becomes a decision about contacting a person who already said no.
Agent Capabilities
16 actions- Check Email Address Cleans an address and then confirms the mailbox exists by opening a dialogue with the domain mail server. Returns a verdict such as EMAIL_OK, MAILBOX_FULL, BAD_DOMAIN, BAD_USERNAME, BAD_SYNTAX, BLACKLISTED, TRASH_MAIL, EMPTY_EMAIL, QUERY_FAILED or UNKNOWN, a quality score from 0 to 3, the corrected address split into user name and domain, and the raw SMTP code and reply. Used at sign-up and before a send.
- Normalize Email Address Cleans and reformats an address using syntax analysis only, without contacting the mail server, which makes it far faster than a full mailbox check. Returns EMAIL_OK, EMPTY_EMAIL or BAD_SYNTAX with a quality score and a flag showing whether a correction was applied. Used for high-volume tidying where deliverability does not need proving.
- Suggest Email Addresses Proposes complete addresses from a person's first and last name plus whatever has been typed into the email field so far. Used to drive assisted entry on sign-up and checkout forms, where offering a valid completion prevents the typo that becomes a bounce.
Phone
2- Normalize Phone Number Cleans a number and returns it in international, national, national significant, national compact, E.123 and RFC 3966 layouts at once, with the detected line type, two-letter country code and country calling code. Reports IS_POSSIBLE or a specific failure such as INVALID_LENGTH or TOO_SHORT, and sets a suspicious flag on numbers that are syntactically fine but look fabricated.
- Search Phone Numbers Looks up numbers in the French telephone directory. Supply a person or company name with a postal address to find their number, or supply a number alone for a reverse lookup that recovers subscriber name and address. Each result carries a no_prospecting flag showing whether the subscriber objected to commercial prospecting, which has to be honoured before any outbound campaign.
Postal Address
2- Normalize Postal Address Restructures, standardizes and validates an address against postal norms, returning it split into house number, street type, street body, postal box, sub locality, postal code, city, state and country plus a ready-to-print formatted array. Carries a status code and a quality code from 0 to 3 expressing the probability of non-delivery, with alternative suggestions when the input is ambiguous. Optional geocoding adds coordinates and French administrative references.
- Search Postal Address Researches an address from a single free-form line and returns matching standardized addresses. Used when an address arrives as one unstructured string, from an imported file or a chat message, and has to be resolved before storage.
Guided Address Capture
3- Search Cities And Postal Codes Suggests cities, places and postal codes from partial input and opens the guided capture funnel. Each proposal carries a relevance score, place and city labels, postal code, INSEE code, population and the geographic keys the next step needs, plus a flag showing whether a street repository exists for that city.
- Search Streets Suggests street numbers and street names inside a city already chosen through Search Cities And Postal Codes. Takes the geographic keys from that step plus the street text typed so far and returns matching streets with the identifiers the building step needs.
- Search Buildings Lists the buildings and houses registered at a street number and closes the funnel, so the user selects a real building instead of typing free text into the address complement field.
Identity and Names
3- Normalize Identity Analyses a person or company name that may arrive jumbled across fields and separates it into title, first name, last name, maiden name, company name and department in the requested letter case. Reports a status code, a quality code, validity percentages for the personal and company readings, probable gender, and flags showing whether the title was corrected or first and last name swapped back. Multiple people in one field are detected and returned as an array.
- Suggest French First Names Completes French first names from partial input, optionally restricted to masculine, feminine or unisex. Used on registration and profile forms so names are captured in a known spelling, which improves later matching and deduplication.
- Suggest French Last Names Completes French surnames from partial input against the Amabis repository. Used to capture surnames in a recognised spelling and cut the typos that make record matching unreliable.
Company
2- Search Companies Finds French companies by autocomplete on a free-form string or by structured search on name, trade name and postal address. Each suggestion returns legal name, trade name and sign, full postal address including CEDEX details, flags for active, diffusible and head office status, and the identifier the detail sheet needs.
- Order Company Sheet Orders the detailed information sheet for a company located with Search Companies. Two sheet levels are available, so a lighter or fuller record can be requested. Used to enrich a B2B account with verified legal and establishment data before a quote, contract or credit decision.
Account
1- Get Pack Consumption Reports credit consumption for the pack behind the configured API key: total credits, used, remaining, and the pack start and end dates. Run on a schedule to alert on a pack close to exhaustion or expiry before validation calls start failing.
Frequently Asked Questions
What can FlowRunner do with Amabis Data Quality?
FlowRunner agents can run Check Email Address, Normalize Email Address, and Suggest Email Addresses in Amabis Data Quality, plus 13 more actions.
Does connecting Amabis Data Quality to FlowRunner require OAuth?
No. Amabis Data Quality connects to FlowRunner with an API key, no OAuth flow required.
Can Amabis Data Quality trigger a FlowRunner workflow automatically?
Amabis Data Quality doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with Amabis Data Quality
$100 in credits. No card required. Connect in minutes.