From automatic Dashboard suggestions to a forecast of which spare parts to stock, from used device valuation to the diagnostic checklist suggested in the lab, from the voice command to open a repair to the Telegram support bot that answers by reading this guide: here is where Fixh AI actually works in the platform, today.
We hear about artificial intelligence everywhere, often as a magic word stuck onto any feature to make it look more modern. At Fixh IT we chose the opposite path: use AI, which we internally call Fixh AI, only where it genuinely solves a real problem for people working every day in a repair shop, and say so openly when something is not there yet. Fixh AI is not a single button or a generic chat: it is a set of six distinct features that step in at precise moments of your daily work, in the shop or at the counter. This guide brings them all together in one place, explains where to find them and, just as importantly, what is coming next and what it does not do yet.
Why we use artificial intelligence in Fixh IT
Anyone running a phone repair shop faces decisions every day that require experience: how much a used phone just brought to the counter is really worth, what to look at in this month's numbers before it becomes a problem, how to answer quickly a question about a feature of the back office without interrupting work. Fixh AI does not replace the people working in the shop: it gives them one more tool, quick to consult, that turns data you already have, repairs, sales, inventory, product documentation, into a practical suggestion, in natural language, right when you need it. Every feature is designed to fit inside a workflow that already exists, without adding unnecessary steps: the AI shows up when it helps and stays out of the way when it is not needed.
Smart suggestions in the company Dashboard
Opening the company Dashboard, you find a box with three short observations, generated by Fixh AI from your shop's real data over the last thirty days: number and revenue of repairs, average repair value, number and revenue of non-repair sales, how many products have dropped below the minimum stock threshold. The system builds a prompt with these numbers and asks the AI for three practical, actionable tips, each with a short title and a brief description, in the language set for your company. If the numbers for a given period are too low to say anything specific, Fixh AI does not leave the box empty: it still proposes general tips useful for growing the shop. The calculation no longer happens while you are looking at the page: an automatic process runs it once a week for every active company, so when you open the Dashboard the three tips are already ready and stay the same until the next weekly refresh.
Repair forecast: which spare parts to keep in stock
The company Dashboard also has a Repair Forecast box, which analyzes repairs across the whole network of repair shops using Fixh IT over the last thirty days and shows you the five most repaired models, with the most frequent declared fault for each. The result refreshes once a day and is designed to help you decide which spare parts to keep in stock before you actually need them, based on real trends across the whole network and not just your direct experience in the shop. If there is not yet enough data for a given day, the box says so instead of showing unreliable numbers.
Used device valuation with Fixh AI
When a customer brings a used smartphone, tablet or notebook to the counter and asks how much it is worth for trade-in or direct purchase, the Shop > Used device valuation section puts a standardized condition questionnaire in the staff member's hands, always the same for a given device category. Once every question has been answered, Fixh AI calculates a price range, a suggested price and a written justification, also taking into account the shop's own real purchases of the same model in recent weeks, which weigh more than general market knowledge alone. The result can be turned in an instant into the used device purchase document, already pre-filled. This feature is rich enough to deserve a guide of its own, which you can find in the Point of Sale (POS) section of the Knowledge Base, with the stepper explained step by step, real use cases and the most common mistakes to avoid.
Dictate the repair, transcription turns it into the form
In the wizard for opening a new repair, under Laboratory > New repair, you can also describe the device and the customer's problem out loud instead of filling in every field by hand. The browser transcribes what you say using built-in speech recognition, and only the transcribed text, never the audio, is sent to Fixh AI with the precise instruction to extract only the fields actually mentioned, device model, fault, cosmetic condition, accessories handed over, without ever inventing information that was not said and without ever touching the customer's name or phone number, which always stay typed by hand. The form is pre-filled with what was recognized, but it remains entirely editable: nothing is saved to inventory or to the customer record until you review and confirm it. It is designed for someone whose hands are busy examining the device while the customer talks, not to replace the staff member's attention during the final check.
A diagnostic checklist suggested by Fixh AI
When you save a new repair with the brand, model and the fault declared by the customer, Fixh AI searches the web for information about that specific fault on that model and generates, in the background, a short checklist of diagnostic checks to run in the lab, designed for a technician who already has the device in front of them. The processing does not block saving the repair: it runs asynchronously through an automatic process that runs every minute, so the checklist appears in a warning card at the bottom of the repair detail page a moment after opening it, with a loading message in the meantime. The text, always in the language set for your company, remains a general suggestion to verify against the real device, not a certain diagnosis: the card explicitly reminds you of this every time.
A Telegram bot that answers by reading this very guide
There is a fourth feature, less visible but curious: the platform's support Telegram bot, a single one shared by every repair shop that uses Fixh IT. When someone mentions the bot in a support channel or group by writing a question, Fixh AI reads it together with the list of every article published in this Knowledge Base, automatically recognizes the language of the question among the nine supported by the platform, and identifies the most relevant article by meaning, not by simple keyword matching. It then generates a short reply, a few sentences long, based only on the actual content of that article, with a direct link to read more, and sends it in the same language as the question. If it finds nothing genuinely relevant, it says so clearly instead of inventing a plausible but wrong answer. In a sense, every article in this Knowledge Base, including the one you are reading now, is also material Fixh AI relies on to answer support questions on Telegram.
What's coming soon
Two features are in the works and are not yet available in the back office: reports and statistics automatically commented on by artificial intelligence, to read a shop's performance at a glance without having to interpret charts and tables on your own, and spending forecasts based on the history of purchases and expenses, to anticipate spikes or drops before they happen. We will update this guide as soon as they are released, following the same logic as the other features: one more tool only where it truly helps, never a feature added just for the sake of announcing it.
Quick guide
- For automatic tips, open the company Dashboard and look at the dedicated box: it refreshes itself once a week.
- To see which spare parts to keep in stock, look at the Repair Forecast box in the same Dashboard: the data is aggregated across the whole network and refreshes once a day.
- To value a used device, go to Shop > Used device valuation, identify the device and answer the condition questionnaire.
- To open a repair by voice, go to Laboratory > New repair, start dictation and describe the device, fault, condition and accessories.
- After saving a repair with brand, model and declared fault, come back to the detail page after a moment: at the bottom you will find the diagnostic checklist suggested by Fixh AI.
- Always review the transcribed text or the suggested price before confirming: Fixh AI proposes, you decide.
- For a quick question about how a part of the back office works, write it in your company's support Telegram channel mentioning the bot: if a relevant article exists in this guide, it points you to it right away.
- If the bot finds nothing relevant, come back here to the Knowledge Base and use the search bar at the top to browse categories.
Real use cases
A shop owner opens the Dashboard on Monday morning and sees, among the tips generated by Fixh AI, that stock of a popular screen protector has dropped below the minimum threshold: instead of only noticing when a customer asks for it and it is not available, they reorder it right away.
Whoever handles spare parts purchasing looks at the Repair Forecast box every week and notices that a certain battery model is among the five most repaired across the whole network that period: they order a few extra before stock runs out, instead of finding out when a customer is already at the counter.
A technician opens a repair for a model they have never seen before with an unusual fault: a moment later they find, at the bottom of the detail page, a checklist of checks suggested by Fixh AI, a concrete starting point instead of having to search the web on their own for common causes of that fault on that model.
A customer brings in a used notebook during peak hours: while the staff member next to them is busy with another repair, whoever is at the counter answers Fixh AI's questionnaire, gets a justified suggested price within seconds and closes the deal without making the customer wait longer than necessary.
A technician has their hands busy taking apart the back of a smartphone while the customer describes the problem: instead of stopping to write, they start dictation, the transcribed text is interpreted by Fixh AI, and the new repair form already shows the fault and condition filled in, ready to check in a moment.
A new employee, still not very familiar with the back office, writes in the support Telegram channel asking how to activate automatic notifications for customers: the bot finds the right article in the Knowledge Base and replies within seconds with a direct link, without having to interrupt a more experienced colleague.
Tips and best practices
Treat every Fixh AI output as a reliable starting point, not as a decision already made for you: the Dashboard tips, the suggested price in used device valuation, the text transcribed by voice, and the diagnostic checklist in a repair should always be reread before acting, because they know the data you feed them but not the nuances only you have in mind. The diagnostic checklist mainly helps you avoid starting from scratch in front of an unusual fault, but checking the actual device remains the technician's job. When dictating, speak in simple, direct sentences, explicitly naming what you want to appear in the form: the clearer the sentence, the more correctly the corresponding field gets recognized. When writing to the support Telegram bot, phrase your question the way you would to a colleague, describing what you are trying to do rather than just the technical name of a feature: the search is based on meaning, not exact words, so it also understands naturally phrased questions.
Common mistakes to avoid
The first mistake is treating Dashboard tips or the suggested price in used device valuation as final figures to apply without checking: they remain suggestions based on statistics and should always be compared with what you know about your shop and the customer in front of you. Another mistake is expecting the voice command in New repair to automatically save the customer's name or phone number: by design, this data is always excluded from automatic transcription and must be typed by hand, precisely to protect the customer's personal data. Also avoid expecting the support Telegram bot to hold a free-ranging conversation about any topic: it only answers by drawing on the content published in this Knowledge Base, and if a feature is not documented here yet, the bot will not be able to explain it to you. Another mistake is thinking the Repair Forecast box only reflects your own shop's repairs: the data is aggregated across the whole network of repair shops using Fixh IT, so a model can appear among the most repaired even if you have not seen it in your shop yet. Likewise, do not expect the diagnostic checklist to appear right after saving a repair: it is processed in the background and may take a moment before appearing at the bottom of the detail page. Finally, do not confuse features that are already available with those announced as coming soon: AI-commented reports and spending forecasts are not yet active in the back office, whatever you may read elsewhere.
Frequently asked questions
Is Fixh AI a single feature or several different ones? There are six distinct features: the automatic Dashboard tips, the repair forecast to see which spare parts to keep in stock, used device valuation in Shop, the voice command for opening a repair in Laboratory, the diagnostic checklist suggested in a repair, and the platform's support Telegram bot.
Does Fixh AI make decisions on my behalf, for example saving data without me seeing it? No, in none of the six features: it always proposes tips, suggested prices, checklists or pre-filled text that you review and confirm, and nothing is written permanently to inventory, to the customer record or to a document without an explicit action from you.
Is the customer's personal data sent to the AI? In the voice command for repairs, the customer's name and phone number are explicitly excluded from automatic extraction and always stay typed by hand; in the other cases, only the aggregated company data needed for that specific feature is used, such as repair and sales figures for the Dashboard tips.
Does the Repair Forecast box in the Dashboard show my shop's data or the whole network's? It shows the five most repaired models across the whole network of repair shops using Fixh IT over the last thirty days, not just your own shop's repairs: it is designed to show you real industry trends, not just your direct experience at the counter.
How does the diagnostic checklist suggested in a repair work? When you save a new repair with brand, model and declared fault, Fixh AI searches the web for information about that specific fault and generates a short checklist of checks to run in the lab. The processing is asynchronous, so the card with the checklist appears at the bottom of the repair detail page a moment after saving.
Is the support Telegram bot dedicated to my company or shared? It is a single bot shared by every repair shop using Fixh IT: it answers by drawing on the public articles of this Knowledge Base, not on data specific to your company.
Are AI-commented reports and statistics and spending forecasts already available? No, they are in the works: this guide will be updated as soon as they are released, so if you do not see them yet in the back office it is because they have not been published yet.
What happens if Fixh AI cannot generate a response, for example due to a temporary service issue? Every feature is designed not to block your work: if the service does not respond, the tips box simply stays empty, the used device valuation can be repeated, dictation leaves the form editable by hand as always, and the Telegram bot reports that it could not find an answer instead of making one up.
