01
Visits without a diagnosis
The technician drives out to find out what's broken. The second visit brings the spare part.
Service Decision Intelligence
Know what's wrong with the machine before you send a technician.
SDI by logicline connects machine file, CRM, ERP, telemetry and manuals and gives your service team evidence-based findings, equipment context and prepared next steps, right in Salesforce, in your portal or in your AI assistant.
Read-only, one machine type, results in 3–4 weeks.
Machine file · FlexWrap 5000
Overheating at the sealing jaw
Causes by confidence
Solid-state relay (SSR) stuck closed, heater no longer switches offSealing jaw 1 · SSR and heating cartridge
82%
Heating cartridge faulty or agedSealing jaw 1 · Heating cartridge
65%
Thermal load in the control cabinetControl cabinet · Cooling
28%
Evidence
Telemetry · 128 readings
Work order 28 Sep: same fault
Maintenance instructions
Data gap: operating hours
Now: Switch off and secure the heating circuit of sealing jaw 1, then replace the SSR and check the heating cartridge. The work order is prepared in Salesforce.
- Every diagnosis with source and confidence
- Single-tenant in your cloud environment
- Language model of your choice, open via MCP
- People decide, every step is traceable
The situation
The answer is spread across five systems and two colleagues' heads.
02
Scattered data
Case history in the CRM, parts in the ERP, signals in the IoT platform, the fix in the manual. Someone pieces it together by hand.
03
Knowledge in people's heads
The two experts who know every product line are the bottleneck. And one day they retire.
04
AI without machine context
A general AI assistant finds documents, but knows neither the serial number nor the build status nor last week's telemetry for this machine.
How it works
Connect. Diagnose. Act.
Your systems stay in charge. SDI reads the data where it lives and returns the result to where your team works.
Your sources
Machine file & CRMEquipment, build status, cases, contracts
ERPParts, orders, availability
IoT telemetrySignals and trends, where available
Manuals & solutionsDocumentation per machine type
SDIby logicline
- Connect: reads from your systems, no data copies to third parties.
- Diagnose: causes for this case on this machine, ranked by confidence, with evidence and similar resolved cases.
- Act: update the case or ticket, prepare the visit with parts, report patterns across the installed base.
Where your team works
Salesforce Service CloudDiagnosis right in the service case
Field ServiceVisit preparation for technicians
Agentforceas a skill of your service agent
Your AI assistante.g. Copilot or Claude, via MCP
No structured installed base? We bring it along: IOTAM by logicline, the Connected Machine File.
In action
From service case to evidence-based diagnosis, right in Salesforce.
Use cases
Where SDI fits in service.
SERVICE DESK
Case triage
Which case needs a visit and which can be solved remotely. With reasoning a colleague can check.
FIELD SERVICE
Visit preparation
Diagnosis, parts and approach are settled before the technician leaves. Fewer repeat visits.
QUALITY
Fault patterns in the field
Spot recurring causes per product line while only a few machines are affected.
WARRANTY
Warranty & RMA
Review claims with reasons: what happened, what the contract covers, what speaks for and against.
How it differs
Why not just an AI assistant?
The test for any service AI: does it know your machine or only your documents?
Document search & chatbot
Helpful for knowledge, blind to the machine.
- Finds matching passages
- Gives an answer in a chat window
- Knows neither serial number nor build status nor signals
- Rarely clear where a statement comes from
SDI
Diagnosis for this machine, with evidence.
- Knows the machine: build status, history, current signals
- Delivers a decision proposal with confidence
- Lands in the Salesforce case or work order
- Source for every statement, audit trail for every decision
Getting started
Measure first, then decide: a backtest on your own cases.
SDI re-triages your closed service cases for one machine type, seeing only what was known when each case was opened. Then we compare with the actual resolution.
- Week 0Kick-off
Agree on machine type, time period and export list. 60 minutes.
- Week 1Data delivery
Read-only export of 150–300 closed cases, master data, fault codes, service documentation.
- Week 2Backtest run
Each case is diagnosed with the knowledge available at the time and matched automatically.
- Week 3Review by your expert
An experienced technician blind-rates a sample. Their judgment counts.
- Weeks 3–4Results & recommendation
Readout for service management, with a clear recommendation: pilot, data foundation first, or no fit.
What you get
- Hit rate of causes on your cases
- How often the right spare part would have been suggested
- Visits that could have been avoided, confirmed by your expert
- How well your case data suits AI and which gap costs the most
- Projection in euros, based on your own figures
Fixed pricecreditable toward the pilot
Read-onlyno changes to your systems
EU hostingpseudonymized data
Not sure your data is enough? Take the free self-check
Backtest→
Pilot→
Operation
In the pilot we measure against the backtest figures from week 1.
Self-check, free
Are your cases ready for a backtest?
Seven questions about one machine type, about two minutes. You get a recommendation for the next step right away, no sign-up.
Recommendation
Thank you. We'll send you the checklist within the next few business days.
Preliminary check based on your own assessment. How good the data really is only shows in the backtest.
Request a backtest
A few details are enough for the initial call.
So we can assess in advance whether your machine type and case data are suitable for a backtest.
- We review your detailsand get back to you personally, no newsletter automation.
- 30-minute initial callMachine type, data situation, contact person on your side.
- Fixed-price offercreditable toward a later pilot.
Thank you, your request has arrived.
We'll get back to you personally to schedule the initial call.
About SDI
SDI by logicline is the AI platform for the technical service of machine and equipment manufacturers. It reads data where it lives: machine file, CRM, ERP, IoT telemetry and documentation. The result is diagnoses with source and confidence, context on the equipment, and prepared work steps. Specialized agents handle triage, visit preparation or warranty review. People decide, every step remains traceable. Single-tenant, language model of your choice, open via MCP.
SDI is a logicline product. For more than 30 years we have worked with machine and equipment manufacturers on their service, today as a product company on Salesforce.
Where your Salesforce partner is already on the project, we work with them: they build the framework, SDI delivers the diagnosis.
FAQ
Frequently asked questions
Do we need Salesforce?
SDI delivers the most value with Salesforce Service Cloud or Field Service, because diagnosis and next steps land directly in the case. You can connect other AI front ends via the open MCP protocol.
What data does the backtest need?
An export of closed service cases for one machine type from the last 12 to 24 months, with resolution and parts used, the asset master data, the fault code list and the service documentation for the type. Telemetry is optional.
Where does our data run and which language model is used?
SDI runs as a single-tenant service in your cloud environment. You choose the language model. For the backtest we work in an isolated environment in the EU with pseudonymized data.
How do we measure whether it works?
The backtest provides the baseline on your own cases: hit rate of causes, parts hits and visits that could have been avoided. In the pilot we measure against these figures from the first week.
How does SDI fit with our current AI program and our Salesforce partner?
SDI is a building block, not an entire program. Your AI assistant or Agentforce calls SDI for service diagnoses. Your implementation partner builds the framework, SDI delivers the diagnosis.
How is SDI different from knowledge management and document AI?
Document AI finds matching passages. SDI also knows the specific machine with its build status, history and current signals, and delivers a reasoned decision proposal with confidence that lands in the service case.
What would SDI have delivered last year?
Find out on your own cases. One machine type, 3–4 weeks, fixed price.