maX an AI assistant for leadership coaches
It takes over the analysis of questionnaires and interviews, so coaches can focus on their clients instead of the paperwork.
AI pilot
One process, one AI pilot, one fixed price
You don't get a report on what AI could do. We pick one process, and typically within 4–6 weeks you find out whether it works on your own live data.
For small and mid-sized companies that spend a lot of time on repetitive text work or on preparing decisions.
Related work
Other work
-
Ask it in writing, and it answers from this site's content. You can try it right now.
AI in live products, not just in demos
We don't give talks about AI. We build working AI features, and we use AI in our own work every day.
- maX: a live AI assistant for leadership coaches that takes the analysis of questionnaires and interviews off their hands.
- The chat assistant running on this site, which you can try right now.
- We develop with AI agents, but a specialist decides how the software is structured and whether the result is accepted.
- We clearly flag the use of AI wherever it matters to the user – for example, when they are talking to an AI system or receiving an AI-generated answer.
What does it cost?
We use the same pricing bands for three project sizes across every service area:
-
MVP or pilot
HUF 3–8M
Typically 4–12 weeks
-
Production system
HUF 8–20M
Typically 3–6 months
-
Platform
from HUF 20M
From 6 months, quoted individually
Before every project: Review and technical plan — HUF 0.9–1.5M, 2–3 weeks.
What each range includes, and what drives the price
Prices are net of VAT and indicative. The final price is set after the review.
Already have a plan or a prototype?
If you have already worked out your idea with ChatGPT, Claude or another AI tool, we don't need to start from scratch. We review the specification, the screen designs, the prototype or the source code, and tell you what can safely be built on. What works, we keep; what needs it, we fix or redesign.
When we are not the right choice
- If an off-the-shelf tool, such as a subscription chatbot, does what you need, buy that: it is cheaper and faster.
- If the data the AI would work on isn't at hand, or nobody knows where it lives, sort that out first.
- If you need compliance advice, for example on the EU AI Act, that calls for a lawyer, not a developer.
Frequently asked questions
What happens if the pilot doesn't work?
We tell you in writing that it doesn't, and why. That is a result too: you find out at a fixed price within a few weeks, not after a failed rollout.
Do I need my own data?
Yes, that is what makes the pilot worth anything. We test on your real documents, emails or spreadsheets, not on invented examples. What we use and how we handle it is agreed together at the start.
Who is responsible for a wrong answer?
That depends on the situation and the decision process. Before building anything, we set out where a specialist checks the AI's answer and where the AI makes the decision, what happens in case of uncertainty, and who is notified of an error. In maX, for example, the final call rests with the coach; Appific's and the client's contractual responsibilities are recorded separately.
Is development cheaper with AI?
AI can speed up writing code and some repetitive tasks, but that does not necessarily make the whole project cheaper. Planning, integration, testing and expert review remain substantial work.
Andor Kerti
Founder, agentic software engineering
At Appific, we have used an AI-agent-based development system since 2024: the agents work within strict technical rules, while a specialist decides on the system's structure and accepts the result. I have more than twenty years of experience as a software developer and architect; I previously worked on fintech systems. I now put that experience to use designing systems, reviewing them and managing their technical risks.
LinkedIn (opens in a new tab)Tell us what you are working on, and we will show you how we would start.