PRAXIST
ENTR Contact us
  1. From the field
  2. 5,500+ of 6,000 invoices a month booked automatically
See the work

Forward deployed engineering: engineers embedded in your team

AI, inside
your business.

Praxist builds the AI systems that do your team's repetitive work: invoice processing, marketplace reconciliation, cybersecurity alert response.

Contact us
  1. Invoice processing
    5,500+of about 6,000 invoices a month booked automatically

    Invoice processingFinance

  2. Marketplace reconciliation
    MarketplaceERPCommissionFeesReturnsPayout
    Denokids · 55 rules in 11 categories

    Marketplace reconciliationFinance · Denokids · In progress

  3. Cybersecurity alert response
    12 sper alert, down from 15 minutes
    15 min12 s

    Cybersecurity alert responseSecurity

See the work

You're using AI. But are you applying it to your business?

Your best people should not spend their days copying data, checking records, resolving routine discrepancies, and building manual reports. At one client, a team of seven was processing about 6,000 invoices a month. When people left, the client came to us instead of hiring.

  • more than 5,500 of about 6,000 invoices a month booked automatically
  • from 7 to 3 people doing the same invoice work
  • from 15 minutes to 12 seconds response time to a cybersecurity alert

From the field

What our team has built

AI does the repetitive work at every step that needs it. People keep the exceptions and the decisions.

  • Invoice processing

    ResultMore than 5,500 of about 6,000 invoices a month are booked automatically. Three people now do the work that took seven.

    What was builtAI reads each incoming invoice, extracts its fields and books the clear ones in the accounting system. People review the rest.

    Finance

  • Marketplace reconciliation

    ResultIn progress

    What was builtFor Denokids, a children's apparel brand on about 27 marketplace channels: 55 rules in 11 categories that compare commissions, fees, returns and payouts between marketplace and ERP data and calculate the amount of every discrepancy.

    Finance

  • Cybersecurity alert response

    ResultFrom 15 minutes to 12 seconds per alert.

    What was builtA SOC (security operations center) automation product written by our AI engineer.

    Security

Incoming invoice flow

  1. Invoice is fetched automatically

  2. AI extracts its fields

  3. Clear cases are booked automatically

  4. A person reviews the rest

Applied AI: for your processes, in your systems, in production.

We don't advise from the outside and leave. This way of working is called forward deployed engineering.

  • Alongside your team

    Our engineers work alongside your team and learn how the work actually gets done.

  • Inside your environment

    We build the system inside your environment, on the systems you already run.

  • Until it runs in production

    We stay with the work until the system is running in production.

Team

The people who do the work

A founder who has run companies and an AI engineer with production experience.

Sencer Kutluğ

Founder, Praxist

Sencer founded Praxist to help businesses put AI to practical use. Having led a business himself, he understands the pressure to act, the difficulty of changing how an organization works, and the responsibility to do right by its people. His approach starts with understanding the work: what helps the business, what slows people down, and what no longer needs to be done.

Bilyoner
Served as CEO of a major Turkish internet company.
Google Turkey
Led the travel, finance, gaming, and startups verticals, as well as the data insights and product specialist teams.
Tez Tour
Built one of Turkey's largest online B2B ticketing platforms.
Citrix Systems
Began his career as a software engineer in the US.
Boğaziçi University
Computer Engineering
Ohio State University
MSc in Computer Science
INSEAD
MBA
LinkedIn

AI Engineer

Has worked on AI systems in production. As a forward deployed engineer, wrote the SOC automation product listed above.

Contact us

To discuss an operational problem:

[email protected]