Frequently Asked Questions
If your standard options don’t fit our requirements, is it possible to make development from your side?
Yes, it is possible to make additional development. Our goal is to provide a flexible service for our clients. You have to contact our sales managers for more details.
How long do you store data in your system, is it possible to delete it after some period or delete it by request?
We can store data forever or you can tell us how long we have to store it. Yes, we have an API method/Admin Panel option to delete applications(by default, it is turned off)
Are you doing a Proof of Address check?
Yes, We have POA Service.
- Address auto prefill and cleansing in the online form
- Comparison of user submitted data with the data in the provided document
- Automatic search of data in the provided document (First and Last names, address of residence)
- Reports in Dashboard (including downloadable PDF with highlighted data fields)
What is the average response time for verification requests?
Our SLA is 5 min. In general, we use a hybrid approach. There is a fallback on the manual check team if the system is not confident enough in the document check results. Even if we have a pre-processing of all images, there are still cases when photos passed to the system have some reflections/glare, or their general quality is not good (mostly, blur). These cases are processed by the manual team. We can propose a fully automatic solution that will process documents in seconds but this will affect the pass rate. Also, the visual authenticity check of the document will not be done. For automatic document check, we have a check of whether the provided document is in line with official rules for such document type (we know the data points, where and which data should be placed). MRZ validity check is a correct cross-validation of visually extracted data and data decoded from MRZ expiration date check. Facial-similarity and AML watchlists checks are being done in seconds as well.
What cloud providers do you support for SaaS (Software as a Service)?
For a cloud solution, how do you ensure that our data is safe and not shared with any other third party?
What are the rules about what the user needs to enter? For example, do they need to enter the native character format for names?
It is better to use Latin symbols (when they exist in a document as the main language or as a duplicate). We also extract the native ones as original name fields.
What about a case when there are 3 or 4 names? Where do we enter the middle name(s)?
If there is a separate field in a document for the given/father's name, we can automatically compare it with the document's data.
From our experience, the exact match comparison of the First Name and Last Name values that are provided by some clients during self-registration with data extracted from a document, decreases the pass rate of the application, since clients do not provide exact names as in the document, but only the main ones. For example, during initial registration/profile creation, a client provided the following information: Phillipe (First Name) Filiatre (Last Name). Whereas, Phillipe Rene (First Name) Filiatre (Last Name) were extracted from the document provided. As a result, there wasn't an exact match, and the "Needs Review" status was given. Our recommendation, in this case, is to approve a First Name and Last Name's match automatically. However, we can collect all names from the document and separate/combine them exactly how your workflow requires.
How do you check the ID document?
- We have a database of ID document templates that have information about data points of each document type and revision of it. Relative coordinates, types of data, borders, fonts, etc.
- AI engine processes the picture of ID document and makes the classification - to understand the ID doctype and a template with rules to be applied for checking.
- AI engine applies rules for the template.
- Date extraction - we extract data from the visual zone of the ID document and cross-check with template rules.
- We check the MRZ (if applicable) for its validity (overall checksum + internal checksum for dates, OCR type B font usage).
- We decode MRZ and cross-check with the visual zone (name, last name, dob, type of doc, date of issue, date of expiry). Any inconsistency is visible in the report.
- We define and check for the existence of applicable doc features (signature, portrait - identify that is a picture of a face, ghost portrait, barcodes, etc). These breakdowns and results are potential marks of a suspected document. You will be able to review it when you make a final decision.