Working together, starting with an AI pilot project
If you are thinking about an AI product for your team or organisation, we suggest starting not with a large project but with a small pilot. Working with Summits Lab follows four steps: conversation, scope, pilot and measurement.
Why we start with a pilot
How a process looks on paper and how it actually runs are often different. Who looks at which list, which step is done by hand and where information gets lost only become visible while the work is happening. So we first try a narrow scope in real work, and then decide together whether to grow it.
We build our own products the same way: we experiment, measure and ship. We follow the same approach when working with others. A pilot is a step that both sides can walk back from.
The kinds of problems we look at
We look at problems close to the ones we work on in our own products. The list below is not a fixed catalogue of services; it is a few examples to start the conversation.
- Scattered requests and follow-up. Requests arrive through different channels such as web forms, email, messaging and the phone, and follow-up lives in separate lists. You want to see each person's history in one place and not miss the next step.
- Explainable recommendations. You need to recommend the right option from many, and show the reason for the recommendation too.
- Records from photos or free text. You want to turn information that is typed freely or photographed into an orderly record that a person can check and correct.
- Planning and calculation. You need to see linked numbers such as measurements, budget or schedule on one screen and make a decision.
- Repetitive messages and notifications. The same messages are sent one by one. You want to bring order to them, with consent checks and approval before anything goes out.
In every case we follow the same rule: AI brings a suggestion, the decision stays with a person. Work that is expected to replace an expert's judgement, such as a medical or legal opinion, is not what this collaboration is for.
How a pilot works
1. Conversation
We listen to your process: which channels work comes in through, which tools the team uses today, where it gets hardest. The aim is to understand whether the problem can really be solved with a product. Sometimes the right answer is "you don't need a new product", and then we say so plainly.
2. Scope
We pick one workflow or one decision. Together we write down at the start what the pilot covers and what it doesn't, who will use it and how success will be measured. The measure is chosen to fit your work; for example, how long a request waits for its first reply, or how many steps are done by hand. Which data will be used, where it will be kept and who will have access are also discussed at this stage.
3. Pilot
The chosen scope is tried in real work with a limited team. If a suitable ready-made module exists, we start from it, and the missing part is set up around your process. During the pilot we collect feedback from the people using it and ship new versions when needed.
4. Measurement
At the end of the pilot we look at the measures we wrote down at the start. Depending on the result there are three ways forward: widen the scope, change the approach or stop. We make that decision together, based on what was measured.
Experience from our own products
This process is based on what we learned building our own products. Setting up Talepyolu follows a similar order: analysis of the current process, ready-made modules plus a setup tailored to you, migration of historical data, team training and go-live support. With BanaOto we learned to show the reason behind a recommendation, with Afiyetim to work without an account, and with Yerleşim Atölyesi to say plainly that a measurement is approximate. We describe our products and approach on the about page.
First step
Describe your process briefly: which work you want to make easier, how it runs today and who uses it. You can use the contact form on the home page or write to hello@summitslab.com.