Implementation
AI applications
We design and build the AI application for the work you chose; we test it with your real data and roll it out with your team. Applications are offered on a monthly subscription; scope and price are set in a written proposal.
Three applications
-
Website chat assistant
Answers your visitors' questions from your company's own documents; says when it doesn't know and points them to you. You manage the documents and assistant settings from a Turkish-language panel and see the conversation history and usage.
Example: The chat assistant in the bottom-right corner of this site.
-
Knowledge assistant and corporate memory
Your employees ask questions about company documents in their own words; the answer comes with the document it is based on. Everyone sees only the documents they are authorised to see.
- Internal knowledge assistant
- Corporate memory: Decisions, meetings and past work
- Dealer and partner assistant
-
Workflow automation
Connects repetitive work into a single flow: Trigger, data, rule or AI, approval, action and follow-up. At important steps, a person makes the decision.
Quote and email drafts, weekly reports, customer and collections follow-up, approval chains, routing incoming requests to the right person, and more.
Workflow families
Which flow is built for your company is chosen in the discovery call or the AI Opportunity Map and written into the proposal.
Draft: Quote, email, contract and report drafts
Handover: Work and data passing from one team or system to another
Report: Daily and weekly reports, management summaries
Follow-up: Quotes, customers, collections, tasks and orders
Approval: Approval chain from manager to finance
Request: Classifying an incoming request and routing it to the right person
Notification: Telling the right person when something happens
Reminder: Reminders for work that is due
Escalation: Moving work that isn't resolved in time to a senior manager
Verification: Passing AI output through rules and a person
Data collection: Bringing data from forms, email or systems into one record
Classification: Sorting requests, emails or documents by type
Document generation: Quotes, reports or forms from templates and data
Synchronisation: Keeping records in step across sheets and systems
Planning: Task, appointment and shift plans
Exception: Retrying a failed step, then handing it to a person
Archive: Filing finished work in the right folder and record
Underneath all of them: security and connections
Security and KVKK: Who can see and change what; a record of actions; where personal data may and may not go. Important decisions need human approval.
Connections: The application talking to email, spreadsheets, forms and the systems your company uses. Which connections are built is written into the proposal.
How it is built
- Scope: In the discovery call or the AI Opportunity Map we choose the work to build; scope, subscription and price are in a written proposal.
- First version with your real data: Not with made-up examples, but with your documents and your work.
- Trial: Tested with a few users on real work, then updated and refined based on their feedback.
- Roll-out: Your team learns to use it; the scope of support and updates is written into the subscription.
Cards on the table
Rules
- Subscription in writing from the start: What is included, the monthly usage limit, support, and how your data is handed over and deleted if you cancel are set out in the proposal.
- Least privilege: The application only accesses the information it needs for its job.
- Large engineering work: For work such as custom ERP development, image processing or phone systems, we design the solution, write the technical specification and take part in choosing the implementer and in acceptance testing; the implementer you choose does the build.
Training teaches what is possible, consulting chooses what to build, implementation builds it: Training · Consulting.
Frequently asked questions
Subscription or one-off?
Applications are offered on a monthly subscription. The chat assistant also has a one-off licence option that runs in your own cloud account; with a licence, support and updates are agreed separately.
What happens if we cancel the subscription?
Your data is exported and deleted under the conditions set out in the proposal.
Do we have to do the AI Opportunity Map first?
No. If it's clear which work to build, I send an implementation proposal straight after the discovery call. If it isn't, the Opportunity Map is the better first step.
How much does it cost?
It depends on the scope; after the discovery call I send a written proposal with a fixed price.