Process customer due diligence replies
AI reads returned due diligence forms and shows where they differ from the customer record. Approved changes are entered, and follow-ups are listed.
- Runs on your own PCs and servers
- Use Azure OpenAI, Amazon Bedrock, Google Vertex AI, or an in-house model
- Its engine is Descarty's open-source Dagu (GPL-3.0)
The problem today
Ongoing customer due diligence under anti-money-laundering rules covers, in principle, every customer, and periodically confirms details such as occupation and the purpose of the account. Institutions ask by post, compare each returned form with the customer record on screen, and type in what changed. Customers who do not reply also have to be chased.
The workflow in Kitewell
- AI reads occupation, purpose, beneficial owner, and other items from the scanned reply formAI judgment
- Get the current details from the customer record screenWeb
- AI shows, item by item, where the reply differs from the recordAI judgment
- Staff review the differences and decide what to update and whether the risk rating needs reviewApproval
- Enter the approved changes in the customer record and save the screenWeb
- List customers due a risk review and customers who have not repliedAutomatic
What AI does
Reads reply forms, including handwriting, and shows the differences from the record item by item. Items read with low confidence are shown with the matching part of the form.
Where people decide
Staff decide which items to update and whether to review the risk rating. AI never makes decisions that affect the customer, such as restricting transactions.
Prove Kitewell on your own systems
Try it free for 14 days on your own PC. To see it working first, ask for a 30-minute demo. There is also a one-month onboarding (¥500,000 before tax), where we build your first job together.
What to check when you try this work
- How reliably handwritten items on the reply forms can be read
- Whether the customer record screens can be operated from a computer on your internal network
- Whether the AI model that handles customer information can run in-house or in a closed network
How the onboarding runs
- Week 1Requirements and success criteria
- Weeks 2 to 3Build
- Week 4Verify on production screens
- Final dayReport
One-month onboarding
- Length
- One month
- Cost
- ¥500,000 (tax excluded). Your staff build, with the developer beside them
- If you go on to a license
- The onboarding fee is credited to the first year's contract
- Without the onboarding
- At Team scale, start on Team monthly (¥14,800 per computer a month, tax excluded), and cancel in any month
You can try it first on your own PC, in the free 14-day trial.
Frequently asked questions
- Can customer information be processed without sending it to an outside AI?
- Kitewell runs on VMs in-house or in a closed network and uses AI models you host. AI models that can read and operate a screen are limited, though, and small local models usually cannot operate one, so the PoC settles which model to use.
- Do our documents and screens leave the company?
- Kitewell runs on your own PCs and servers, and Descarty never receives your business data. You choose the AI model: a cloud API, or an in-house or local model with an OpenAI-compatible API. With an in-house model, no data leaves the company. With a cloud model, what the AI reads is sent to the model's provider.
- What if the AI misreads something?
- A person checks what the AI produced before the work goes on. In a sheet, every value the AI reads shows the quote it came from, and rows below your confidence threshold go to a person. Reviewers can send work back with notes on what to fix, and who approved what is kept in the run history.
- What happens when a screen changes or a run stops?
- When a screen changes, the AI reads it again, relearns the step, and carries on. The relearned step is saved, and later runs repeat it. If a run still stops, ask the assistant right there, and it finds the cause and proposes a fix. A proposal is not saved until a person applies it.
- What happens to the RPA and systems we use today?
- Keep them. Kitewell operates screens the way a person does, so it works with systems that have no API. Move the robots that keep stopping and the work you gave up automating because it needs judgment, and leave fast, fixed processing in your current RPA.
- What does it cost?
- ¥14,800 per computer per month, or ¥148,000 per computer when paid yearly, before tax. A computer costs the same whether it builds or runs. You can try it free for 14 days. The one-month onboarding, where we build your first job together, is ¥500,000 before tax, credited to the first year if you go on.
Related use cases
Answer inquiries that arrive on paper
For inquiries on paper, AI reads the form, the workflow looks up accounts across systems, and AI drafts the answer. Staff identify the person and approve.
Screen new business partners from public sources
For each new partner on an Excel list, AI searches news and public sources and flags anything to check. A person decides, and the result stays in the row.
Enter insurance claim documents
AI reads claim forms, certificates, and receipts, and the workflow looks up the policy and enters the claim. Staff see mismatches; an assessor decides payment.
See this work in a demo
A 30-minute demo shows this workflow running. The 14-day free trial lets you check it on your own systems.