The FindFST AI Engine: How AI Matches Field Service Technicians

What Is the FindFST AI Engine?

One engine behind every AI feature on FindFST. It reads what people upload and returns suggestions a person confirms. It has two rules that hold everywhere:

• It suggests. Nothing reaches a mission, a machine or a profile until somebody clicks and saves
• It advises. The match score informs who a company invites and what a technician applies for, and decides nothing on its own

What it reads, and what it returns

Reading a Machine From Its Nameplate

A mission is only as well matched as the machine behind it. Adding a machine takes a photo or two rather than a form, and the details it reads are what later drive the matching.

1. Add the photos

• Two slots: the nameplate, and the machine itself
• Either photo alone starts a read; both together give the richest one
• The nameplate is usually the better first photo, because it carries the data
• Further photos, manuals and datasheets can go alongside, up to 5 files

2. The engine reads them

• A plate or cabinet label is read as printed
• The machine photo tells it what kind of machine the plate belongs to
• A crop is offered for the machine photo, so the frame shows the machine and not the hall

3. Identity, marked Check

• Name, brand, industry, machine type and model are filled in
• Each carries an amber Check marker until you confirm it
• The brand is kept exactly as printed, never swapped for a better known maker
• Serial numbers are not copied into the profile

4. Technical details from the photo

• Control system, component brands, subsystems and suggested skills
• Private to your company and to technicians booked on that machine
• Selecting the machine on a mission later copies the relevant skills and brands into the mission requirements

5. Hazards and specifications

• Hazards such as moving parts, chemicals, working at height and high voltage
• Condition notes, and specifications such as supply voltage, frequency and year built
• Every field is editable, and nothing is saved until you press Save

AI Mission Descriptions: Drafting the Mission From Your Photos

The mission form is built so the photos lead. The engine turns them, the machine and the schedule into a first draft that you correct in plain words.

1. Select the machine and set your priorities

• The machine's brand, industry and type become the matching criteria, removable one by one
• Up to four questions: what matters most, where you could compromise, how much being nearby matters, and how much certifications matter
• The answers change how technicians are ranked, not who may apply

2. Add photos with a short note under each

• Show the machine, the fault and the nameplate
• A short note under each photo says what it shows
• These are the engine's main source, so a clear fault photo usually beats a wide shot of the line

3. Generate with AI

• One press drafts the mission title and the description
• Drawn from the photos and notes, the machine, the skills and the schedule
• 5 AI generations per mission, refines included

4. Review the suggested requirements

• Skills, certifications, parts, machine types, brands, industries and days needed
• Offered as chips, unchecked. Nothing is added until you click it
• Add them one by one, a group at a time, or all at once

5. Accepted into the mission

• Accepted chips join the skills already derived from the machine
• Parts are listed for the technician to supply, and shown with the mission
• Certifications are scored, not a gate: technicians without one can still apply

6. Refine in plain words

• Type the correction under the description and press Refine
• The description is rewritten and the suggested requirements are worked out again
• Skills the correction took out of the scope are offered for removal, marked in the skills list
• If the correction changes the work itself, the title follows, with Undo

7. The refined description

• A short lead line, then bullets of the work to be done
• Dates written in full, for example 15 November 26
• Everything stays editable before you post

How the AI Match Score Works

What is compared

• The mission: machine brand and type, industry, skills, certifications, mission type and your priority answers
• The technician: skills, brands, machine types and industries from the platform's own lists, certifications and years of experience
• Technicians covering the most requirements are shortlisted, and scored alongside everyone who applied

What comes back

• A score from 0 to 100 against a fixed scale, not a ranking against the others in the batch
• Up to four reasons and up to four gaps, in plain words
• The same number for the company and the technician

How your answers shape it

• What matters most shifts the weight between the exact equipment, the technical skills and the industry
• Where you could compromise lowers the weight of brand or industry experience
• Being nearby matters a lot, on an on-site mission, limits the list to technicians within 100 km. Applicants are still always shown
• Certifications set as essential weigh more than nice to have. Either way, they are scored rather than required

Reading a score sensibly

• The same pair can come back up to about 10 points apart, so treat a few points as level
• Scores are stored, and recalculated by themselves when the mission or the profile changes
• Read the reasons and gaps, not just the number. They say what fits and what is missing
• The score is advice. You choose who to invite and book

AI for Technicians: A Profile Built From Your CV and Reports

A profile is how the engine finds a technician, so a thin one is rarely matched well. The engine fills it from documents most technicians already have.

AI CV parsing for technicians

• Upload your CV as a PDF, up to 10 MB, on the Uploads tab of your profile
• Skills, industries, machine types, brands, languages, bio, certifications, role title and phone number are suggested
• Only numbers that are yours are offered: a referee's or an employer's number is not
• Companies can open your CV from your public profile

Accept what is right, dismiss the rest

• Every suggestion is a chip. Nothing is on your profile until you accept it
• Accept all at once, or one by one
• A dismissed suggestion is not offered again
• In this example one CV produced 20 suggestions across skills, industries, machine types and brands

A bio in short bullets

• A lead line and a few bullets on what you specialise in
• Written from the CV, in the third person, without contact details
• Use it, edit it, or keep your own

Past reports as proof of work

• Upload service or inspection reports: PDF or image, 10 MB each, up to 20
• Each one is read automatically after upload, a few at a time
• A summary is drafted and stays private until you publish it
• Companies only ever see the summary you publish

What a report adds

• Skills, machine types, brands and industries the report shows you working on
• Offered as chips, exactly like the CV
• A report is evidence a CV line is not, which is why it is worth uploading

What the AI Engine Never Does

How the Matching Improves

• Every completed mission is recorded against the match that led to it
• The record holds the score, the company's rating, whether there was a dispute, and how many days ran on schedule
• The engine is measured against those outcomes and its instructions are tuned from them
• The model does not retrain itself on your data. People review the results and change it deliberately

That is the difference from a recruiter's judgement: every recommendation is written down, and whether it worked out is checked.

Frequently Asked Questions

Related Guides