A more human way to shop

Good coffee.
Fair price.

Jan wants a bag of specialty coffee for under €25. He has limited vision, so dense product grids and tiny filters turn a simple purchase into a chore.

Meet Jan's new shopping assistant

Meet Jan

Jan holding his phone near his mouth to shop by voice

“Good coffee. Under €25.”His words. His budget.

One clear need Too much friction

01 · The problem

A simple purchase
turns into work.

What Jan faces

A screen built for scanning

01Dense product grids
02Small controls and long filters
03Reviews scattered across sites

The result

The price is hidden inside the process

Jan can afford the coffee. What costs him is the time and effort required to find it.

Accessibility should not depend on how well someone can scan a screen.

02 · The approach

Move the work
from Jan to the agent.

Voice first

Jan states his need once, in his own words. No grids, no filters, no fine print.

Grounded in the market

Every answer comes from a current listing with a real price and seller evidence.

Checked before claimed

A separate validation agent reviews each listing. Missing evidence stays visible.

The solution

One spoken need. Live research. Evidence you can inspect.

03 · Jan starts with his voice

He already knows
what he wants.

"Find me a smooth specialty coffee under €25. Whole beans, not too acidic, and from a seller I can trust."

Under €25Whole beanLow acidity

One conversation, less searching

04 · Markit gets to work

One useful question.
Then it searches.

01
Listen

Markit picks up the budget, format, taste, and seller requirement.
02
Clarify, if needed

At most one important missing detail. If Jan says go ahead, it searches.
03
Search

It looks for current offers that stay within his €25 limit.
04
Check, then answer

Listings are validated before Jan hears which options meet his needs.

Jan can do the whole search without reaching for a keyboard.

05 · From voice to the market

One spoken request
becomes a live search.

Jan's voice Plain-language request Budget and preferences included

Understood

Markit.ai One structured search Jan's hard budget stays enforced

Live listings Price and availability Retrieved for this request

Product evidence Offer and seller details Kept with each result

Voice requestHard budgetLive resultsSpoken reply

06 · The budget is a rule

Under €25 means
under €25.

Markit only shows coffee with a current, matching price in Jan's currency. Unknown prices and near-budget substitutes stay out of the results.

"
If no verified offer fits, Markit says so instead of quietly moving the goalposts.

A hard constraint

01Current price taken from the live search result

02Budget check currency and price must match Jan's request

03Clear limits missing evidence is shown as missing

07 · Every listing gets checked

One product.
A dedicated second look.

Price check

Does the listing support the price and offer shown on the card?

Offer check

Do the product details support the claims Markit is presenting?

Seller evidence

How much useful seller information is present in this listing?

While Jan listens

Validate before recommending. Seller evidence is not a guarantee.

08 · Results built around Jan

A shortlist he can hear.
Evidence he can inspect.

Illustrated coffee ritual with a cup, whole beans and an unbranded lavender pouch — Jan's coffee brief Concept artwork · not a live offer

What Jan asked for

Maximum price of €25
Whole coffee beans
Smooth, low-acidity taste
Useful seller evidence

What Markit returns

A current matching price
A short reason for the match
Visible audit results
A direct link to the offer

The voice answer stays brief. The product cards hold the detail.

09 · Jan makes the choice

The research fits
into one conversation.

Before

Jan speaks naturally

He gives Markit the same details he would tell a knowledgeable shop assistant.

During

Markit does the research

It searches live offers, enforces the budget, and checks every listing.

After

Jan chooses

He hears the shortlist and opens a card only when he wants more detail.

The outcome

Jan gets a better route to the same market.

Markit.ai

Good coffee.
Less friction.

Markit turns a spoken need into a grounded product choice.

Say what you needChoose with confidence

The team

Built by
four people.

Oskar Kalbarczyk

@oskalbarczyk

LinkedIn ↗

Denys Storozhuk

@denys-storozhuk13

LinkedIn ↗

Andrei Shumko

@andrei-shumko

LinkedIn ↗

Frensi Qatipi

@frensiqatipi

LinkedIn ↗

Your turn

Less browsing.
More finding.

Tell Markit what you need. Let the research start with a conversation.

Open the live demo

Sign in, allow your microphone, and say hello.

Not sure where to start?“Find me smooth whole-bean coffee under €25.”

Take Markit with you QR code linking to the Markit.ai live demo

Scan to open on your phone

markit-ai.dalist.workers.dev

Good coffee.Fair price.

A simple purchaseturns into work.

A screen built for scanning

The price is hidden inside the process

Move the workfrom Jan to the agent.

Voice first

Grounded in the market

Checked before claimed

He already knowswhat he wants.

One useful question.Then it searches.

Listen

Clarify, if needed

Search

Check, then answer

One spoken requestbecomes a live search.

Under €25 meansunder €25.

One product.A dedicated second look.

Price check

Offer check

Seller evidence

A shortlist he can hear.Evidence he can inspect.

The research fitsinto one conversation.

Jan speaks naturally

Markit does the research

Jan chooses

Good coffee.Less friction.

Built byfour people.

Oskar Kalbarczyk

Denys Storozhuk

Andrei Shumko

Frensi Qatipi

Less browsing.More finding.

Good coffee.
Fair price.

A simple purchase
turns into work.

Move the work
from Jan to the agent.

He already knows
what he wants.

One useful question.
Then it searches.

One spoken request
becomes a live search.

Under €25 means
under €25.

One product.
A dedicated second look.

A shortlist he can hear.
Evidence he can inspect.

The research fits
into one conversation.

Good coffee.
Less friction.

Built by
four people.

Less browsing.
More finding.