From flow to function

• Online requests arrive with the data. A client fills in a request on the website. At the office an administrator finds it by tax number and adds it to the queue, so nobody retypes passport details.
• A queue for each request type. New certificate, renewal and reissue, blocking and revocation each have a list with filters in every column and bulk actions: send, download, print.

• Documents from Diia. The administrator scans a QR code in the client’s Diia app or types the code, and the passport and tax number load into the request.
• Photo and accession card in the same place. A webcam photo and the accession card are added right in the request’s documents.
• A signature you can check. Every attached file opens with its signature details: who signed it and when, with which certificate, the signature type and format (CAdES-X Long).

• Data as text, not disabled fields. In the first version every value sat in an input box, even where nothing could be edited. In the refreshed version values read as text, and inputs appear only where the administrator types.
• Status tabs with counters. The queue opens with how many requests are waiting and how many need rework, instead of a flat table with a filter in every column.
• Checks next to the decision. Documents, the register and the signature sit in one panel beside the Approve button, so the decision is made with everything in view.
• One calm palette. Neutral surfaces, the brand navy as the only accent, and colour for statuses only.

• Brand navy
• Deep navy
• Selection
• Background
• Divider

• Status at a glance. Shift states are coloured chips with text, open, closed or cancelled, instead of a line of orange text.
• Sheets instead of dialogs. Signing and the extra actions slide up from the bottom edge, closer to the thumb.
• The total on the pay bar. The sum stays in view above the Pay button while the cart scrolls.
• One accent colour. Amber marks only the main action and the selection; icons and secondary buttons stay neutral.

• What brings the money. The share of each product in sales for the period.
• How prices move. The sale price, the purchase price and the markup between them, by day.
• How much each product earns. Marginal income by day. Every widget opens in detail with a unit switch, sorting and the list of picked goods.

• Totals first. Revenue, cost, margin and gross margin for the period sit in tiles above the charts instead of one line under the table.
• One colour per product in every chart. The grey-blue shades were hard to tell apart; now each product keeps its colour across charts, and the palette is checked for colour-blind readers.
• The unit rule in the table itself. Once a product in kilograms is picked, goods in pieces step aside instead of failing later in the chart.
• Markup as a band. In the price chart the markup is the area between the purchase and the sale price, not a third line to decode.

• Blue
• Brand coral
• Violet
• Green
• Amber

• Calmer tables. Lighter dividers, numbers aligned to the right and status chips with a dot, so the eye goes to the data.
• The route as a chain. Parties are nodes on one line and fill in once they are set, so the order and the gaps read at a glance.
• Clear step states. Done, current and next steps differ in colour and mark, and the progress bar sits above the list.
• Same colours and structure. The product’s green and yellow, its layout and components stay. Spacing, hierarchy and type sizes are tidied up.

After the first screen the question is “what do I get next?”. I answer it with a scenario, not a feature list. The user:

• gets real data from the tests
• gets an explanation and a plan from the doctor
• can get answers and support quickly
• sees every change in one place

So the page shows one system that keeps working after the check-up, not separate features.

1. Perception
2. Trust
3. Value
4. Action

First I change how the product is perceived, from a one-off service to a system, then build trust with specifics and visuals, then explain the value through a usage scenario, and only then lead to the action. That lowers resistance at the start and makes conversion more likely.

• A policy calculator in the first screen. One promise, product tabs and an instant estimated price, so “came in → saw the price” takes one step.
• Products as cards with a “from” price. An icon, a price and an action instead of text lists, with the other types as compact chips. Cards with prices make comparison quick and reduce cognitive load.
• How to get a policy in 4 steps. The whole process up to the policy in your email. Insurance feels long and complicated, especially to people buying it for the first time.

• A strip of trust numbers under the hero. Years on the market, payouts, policies and partner service stations. After the emotional first screen people start to doubt, and that is where rational proof is needed.
• Reliability you can check. The licence, NBU and MTSBU supervision and the share of payouts made on time. The main objection in insurance is “will they actually pay?”, and it has to be answered with proof, not promises.
• The veterans programme as an equal block. A real photo and conditions in line with the products. A strong brand value is easy to lose among other content or to overweight with a separate heavy block.
• FAQ before the final step. Payouts, payment safety, refunds. Without answers people go looking on forums and get lost. It also helps SEO.

• Claims 24/7 as a separate block. Every channel (phone, Viber, online, the app), the steps and a status check, so clients reach help faster and the call centre carries less load.
• A concrete final call to action. A consultation request with a short form instead of a vague “Contact us”: the lowest barrier for people who read to the end but have not decided yet.

I replaced the old mix of styles with one system: brand navy as the base, green as the only action colour, the Manrope typeface, and colour and type tokens. Blue reads as reliability, the key emotion in insurance, and a single accent keeps attention on the action.

• Navy
• Blue
• Green
• Background
• Divider