Satisloh × Midas Analytics
01 / 20
Satisloh × Midas Analytics
Service Optimization · Two Use Cases

Give your service team
its time back.

Two focused AI tools, built from what Brice raised. One answers the repetitive questions that flood in after every install. The other makes sure no existing customer is ever forgotten — so follow-ups happen and upsell moments don't slip.

2 use cases web-based · no SAP integration Satisloh Service · APAC 01 Jul 2026
Prepared for Brice — Head of Technical Services, APAC
Michele De Filippo
Prepared by Michele De Filippo, PhD
CEO · Midas Analytics
What we heard from Brice

Two things quietly eat your team's time.

Your team already carries every install and every support call — that part works. From our conversation, Brice — Head of Technical Services, APAC, and the main user of both tools, pointed to two repetitive drains on that strong operation:

After every install, the questions pour in.
Customers ask the same basic things about how the machine works. The volume is the problem — repetitive questions pull skilled people off the work that actually needs them.
Between sales, follow-ups slip.
Some customers reach out for support or new kits. Others go silent — often a sign the machine is under-used. Either way, regular check-ins get overlooked, and upsell moments pass by.
This is my understanding — correct me where I'm wrong.
What these two tools are for

Free your team from the repetitive — and make sure no customer is forgotten.

Neither tool replaces your people. Each one takes the low-value, easy-to-drop work off their plate — so the skilled hands are free for the complex, high-value cases where they matter most.

At a glance · what each tool delivers

Two tools. Two problems. Both web-based.

Use Case 01

Customer AI Servicing

After install, customers get a link to a chatbot that answers their questions — reading only that machine's manual. It handles the basic, repetitive ones for you.
70%
of routine questions answered without your team
24/7
instant answers, every time zone
Removes the flood of repetitive inbound questions
Use Case 02

Bi-Weekly Customer Report

Brice uploads a SAP export into a web app that builds a full status & delta report — what changed since last time — on every existing customer, and turns it into clear action points.
100%
of customers reviewed, every cycle
0
follow-ups left to memory
Removes the risk of a dropped follow-up

Target outcomes — and we instrument them from day one (deflection rate, follow-up coverage), so you report the real numbers, not a guess.

Use Case 01 · Customer AI Servicing for FAQs
01

An AI that answers the manual, so your team doesn't have to.

The problem
After installations, customers repeatedly ask basic questions about how their machine works. The sheer volume is time-consuming, and most answers are already in the manual.
The idea
A servicing chatbot the customer reaches by URL. It reads the manual for the exact machine or part they bought — and answers the bulk of inquiries, especially the basic, repetitive ones.
The overhead today · and where it could go

Right now, every install turns into an inbox.

Order placed
Technician dispatched
Machine installed on-site
…then the questions start
Today — answered by hand
Customer
back & forth, all day
Service team
Brice
Dozens of emails per install, back and forth. The same basic questions, answered one by one by Brice's service team — that's the overhead.
With the servicing AI
Customer
most
Servicing AIanswers the bulk
complex only
Human
Brice's team
The AI answers the bulk instantly. A human steps in only for the genuinely complex — the repetitive load, and the overhead, collapses.
How it works · step through it

From "just installed" to "already answered."

01
Private link
02
Pick the machine
03
Ask in plain language
04
Answered from the manual
05
Your team, for what's left
Step 1 / 5
Private link
The moment a machine is installed, the customer receives a private link to their servicing assistant — no login to hunt for, no portal to learn.
See it working · click a machine to switch

The customer asks. The manual answers.

servicing.satisloh.com · your machinereplies in the customer's language · EN 中文 日本語 한국어 ไทย
Machine or part:
Illustrative — the questions, codes and steps shown are examples. In the real tool the AI answers only from that item's actual manual, and hands over to a human when it should.
Why this works

Accurate because it's narrow.

Grounded in the manual — never invents
The AI only ever sees the selected machine's manual and answers from it — grounded, not made up. No mixing across products; when it's unsure, it escalates rather than guess.
Handles the bulk of inquiries
Especially the basic, repetitive ones — the exact volume that was eating your team's time.
Available any hour, any time zone
Customers get an answer the moment they're stuck — no waiting for your working hours.
Answers in the customer's language
Your APAC customers span many languages — the assistant detects and replies in theirs, so nothing's lost in translation.
Clean hand-off to a human
When something is genuinely complex, it escalates — so your team spends its time only on the cases that need them.
Statement of work · Use case 01

What we build for the servicing chatbot — and what we don't.

A private-link chatbot that answers your customers from the official manuals of your 15 key machine models — phase 1.

In scope · what we build
  • The 15 highlighted models from your classification sheet: 6 surfacing, 8 coating, 1 finishing.
  • Up to 7 manuals per model — service, software, operating, maintenance, spare parts, wiring, transport (PDF/PPT).
  • After install, your team shares the private link; the customer picks their machine and asks plainly.
  • Answers come only from that machine's manuals, in the customer's language, any hour.
  • When unsure — and for any message with a video — it hands over to Brice's team.
PRA · ART Blocker-A · ART-Deblocker-2 · VFT-Orbit2 · VFT-Orbit2i · Multiflex2 · 1200DLX · 1200DLX2 · 1200TLX · 1200MPX · MC280X · MC380X · MC380H · 1500X/MP · ES4
Out of scope · this phase
  • Models not highlighted on the sheet (e.g. Neo Orbit, ES5, Toroflex) — a later phase.
  • Pulling documents from SAP JAM — not needed at this stage; your upload is enough.
  • Sending YouTube videos to customers — excluded from the initial scope.
  • Reviewing photos or videos customers send — routed to a person by default; future development.
Inputs from Satisloh
  • Already received: the classification sheet, a sample PPT-export manual, a sample customer video.
  • The manuals for all 15 models, uploaded by your team to a shared location.
  • An NDA first, then a few hours of Brice and the team.
Project outline
W0
Prep
NDA, kickoff, manuals collected in the shared location.
W1 – W3
Index the manuals
Up to ~105 documents ingested, scoped per machine.
W3 – W6
Build & test
Chat app + machine selector; tested on questions your team already answered.
W6
Launch
Private link live for your customers; deflection measured from day one.
Done when
  • Each of the 15 models answers from its own manuals and escalates when unsure.
  • Video messages and unanswerable questions reach your team — nothing dropped.
  • Brice's team signs off the test set; the private link is live; ~70% deflection targeted.

Assumptions: Build HKD 187,000 (USD 24,000) · run & support HKD 8,100/month (USD 1,040), at 7.80 · Hosting on ISO 27001-certified, access-scoped infrastructure in a region you approve · You own the tool and the data it holds.

Use Case 02 · Bi-Weekly Report for Existing Customers
02

A standing report that makes sure nothing slips.

The problem
Regular follow-ups with existing customers get overlooked. Some reach out for kits or support; others go silent — a worrying sign of under-use. Upsell opportunities pass unnoticed.
The idea
A web app Brice feeds with a SAP export. It generates a full customer status report on a bi-weekly cadence, turning raw data into clear action points for each customer.
How it's handled today · and where it could go

Right now, the follow-ups live in one inbox.

Orders & kits
Support calls
Going quiet
…and it's all captured in SAP
Today — scattered & ad-hoc
SAP data
forgotten · overlooked · seen too late
Brice's inbox
& memory
It all lives in your SAP — but today it's read ad-hoc, out of Brice's inbox and memory. Some gets forgotten, some overlooked, some seen too late.
On one dashboard
SAP data
all in
Dashboardeverything visible
one click
Report
brief the team
The same SAP data lands on one dashboard. Brice generates the report in a click and briefs the team — nothing scattered, nothing missed.
How it works · step through it

One spreadsheet in. A meeting-ready report out.

01
Export from SAP
02
Upload it
03
Time-stamped
04
Status + delta report
05
Clear action points
Step 1 / 5
Export from SAP
Brice pulls the customer data he already has in SAP as a simple spreadsheet — no integration to build, no IT project to wait on.
See it working · click a customer

Every customer, one screen, one clear next step.

customer-report · Midas web app
Customer status & delta report Uploaded 01 Jul 2026, 09:14 · vs. report of 17 Jun
Customers tracked
48
Need follow-up
12
Gone quiet
5
Upsell signals
7
CustomerKitLast orderLast emailStatus
Meyer OptikAdvanced28 Jun28 JunActive
Optique RhôneBasic09 May28 MayQuiet
Helvetia LensPreventive02 Apr01 MayAt-risk
Nova VisionAdvanced24 Jun19 JunActive
Vista OtticaBasic15 May14 MayAt-risk
The tour runs itself — or click any customer to drill in · illustrative, fictional data.
Why this works

The value is consistency.

Nothing is forgotten
Every customer gets a clear next step, every cycle. Follow-ups stop depending on who remembers what.
Silence gets caught early
A customer who's gone quiet is flagged before under-use becomes a lost account — the delta report makes the drop-off visible.
Every check-in is a sales opening
Consistent follow-ups surface reorders and kit upgrades — better service turns directly into more opportunities.
A ready-made meeting agenda
The action points from one report become the starting point for the next — the review runs itself.
Statement of work · Use case 02

What we build for the quotation follow-up — and what we don't.

A web app that turns your SAP quotation export into follow-up reminders — so no quotation goes quiet unnoticed.

In scope · what we build
  • Brice uploads the SAP quotation export; every upload is time-stamped.
  • Each open quotation is aged from its quotation date and flagged after 60 days without feedback (configurable).
  • Per quotation: customer, value & currency, validity, probability, rejection reason — as exported.
  • Follow-up status, owner, feedback and next step are logged in the app — the export can't hold them.
  • A bi-weekly digest by email to the team; Teams optional (proposed — to confirm at kickoff).
  • Action points per customer become the agenda for the next team review.
Out of scope · this phase
  • Inferring orders or quotation status from the Status column or sales orders — they're unlinked in SAP.
  • Any connection into SAP — you upload the export; we never reach into SAP.
  • Replacing your team's judgement — the app reminds; people decide the follow-up.
Inputs from Satisloh
  • Already received: the export — 28 columns, 2,329 line items, 825 quotations, 111 customers.
  • Your follow-up conventions: owners, feedback states, reminder channel — agreed at kickoff.
  • An NDA first, then a few hours of Brice and the team.
Project outline
W0
Prep
NDA, kickoff; agree the 60-day threshold, tracking and reminder channel.
W1 – W3
Upload & mapping
Upload flow; map quotation #, customer, quotation date, validity, value.
W3 – W8
Report & reminders
Ageing, flags, feedback log, action points, digest — on Brice's real data.
W8 · Day 56
Launch
Live on Brice's real export; both tools running.
Done when
  • The real export loads unchanged; every open quotation shows its age.
  • Every 60-day case is flagged and reaches the team — 0 dropped.
  • Feedback and owners logged in the app survive the next upload.

Assumptions: Build HKD 129,500 (USD 16,600) · run & support HKD 2,500/month (USD 320), at 7.80 · The export keeps its fixed 28-column structure; cadence follows how often Brice uploads · Same hosting terms as use case 01; you own the app and the data.

Trust & security

Your data stays yours.

Both tools touch sensitive material — your manuals and your customer data. They're built so that data is protected by design and never leaves your control.

You upload — we never reach into SAP
The report app takes a file you export and upload. No live connection into your systems, nothing written back.
Encrypted & access-scoped
Encrypted in transit and at rest, in a region you approve, with access limited to two named Midas people.
NDA first, your policy throughout
An NDA before any data reaches us — aligned to your own infosec policy at kickoff.
Midas is ISO 27001-certified
ISO/IEC 27001:2022 · view PDF
The build plan

Day 56: both tools live — the repetitive handled, and no customer forgotten.

Workstream
W0W1W2W3W4W5W6W7W8
Week 0 · scoping & prep
Scope
UC#1 · Index the manuals
Ingest & scope per machine
UC#1 · Chat app, test & launch
Chat + selector → launch
UC#2 · Upload + data model
Upload flow + map SAP fields
UC#2 · Status + delta report, launch
Status & delta → action points → launch
~Week 6 — UC#1 servicing chatbot live ~Week 8 — UC#2 bi-weekly report live
What we'll need from you, to start

A short list — sorted by tool.

For Use Case 01
The machine & part manuals
The chatbot reads these — the more you share, the wider the coverage. We start with a few and grow.
For Use Case 02
A sample SAP export
One real (or anonymised) extract, so the report is shaped around the data you actually have.
For both
An NDA
We need to understand your service operation properly in order to help — the NDA lets us in.
For both
A few hours of Brice & the team
To walk us through how you work today — so both tools fit you, not the reverse.
What you're investing in

One build. Two working tools.

One-off · ~8-week build
HKD 316,500
Both tools built, tested and launched with your real inputs.
① Customer AI Servicing chatbotHKD 187,000
② Bi-weekly customer report appHKD 129,500
Total build (one-off)HKD 316,500
Then · monthly run & support
① Servicing chatbotHKD 8,100 / mo
② Report appHKD 2,500 / mo
Total runHKD 10,600 / mo

Run covers hosting, monitoring, model & content updates and support. ≈ USD 40,600 build · USD 1,360 / mo at HKD 7.80. Effective build cost after the HK R&D credit → next slide.

What's included
  • A customer servicing chatbot, scoped per machine manual, reached by private link
  • A bi-weekly report web app — upload, status & delta report, action points
  • Both launched with your real inputs — manuals and a live SAP export
  • Testing against real questions your team has already answered
  • Built on secure, access-scoped infrastructure in a region you approve

You own the tools and the data they hold; terms are set in the SOW.

Financial incentive · Inland Revenue Ordinance (Cap. 112), Schedule 45

A 300% R&D deduction — recover about half.

Contracted through a Hong Kong entity chargeable to profits tax, building with Midas — a Designated Local Research Institution — qualifies for Hong Kong's 300% enhanced R&D deduction under Schedule 45, bringing the effective cost of both tools down to roughly half.

Effective cost after the credit
≈ HKD 159,800
From a HKD 316,500 build — you recover ≈ HKD 156,700 (~50%) through the HK R&D credit.
Build fee (both tools) — paid to a DLRIHKD 316,500
300% tax deductionHKD 949,500
Tax saving @ 16.5%− HKD 156,700
Effective cost≈ HKD 159,800

Statutory basis: Inland Revenue Ordinance (Cap. 112), s.16B & Schedule 45 — payments to a Designated Local Research Institution are "Type B" R&D expenditure, deductible at 300% on the first HKD 2M, 200% above. Midas Analytics Limited is designated under s.19(1) of Sch. 45 (DLRI No. D041, renewed 11 Jul 2026). The ~50% recovery applies to the Hong Kong entity chargeable to profits tax on the spend — confirm eligibility & filing with your tax advisor. IRD guide · Cap. 112

Certified DLRI · No. D041 — click to enlarge
Get started

Three steps to kickoff.

Step 01

Green light

A short call to walk through this plan and shape the scope of both tools together.

Step 02

NDA & agreement

We circulate the NDA and a simple statement of work. You sign — about a week.

Step 03

Kickoff & Week 0

You hand over the manuals and a sample export — and the build starts.

Satisloh × Midas Analytics

Thank you.

Free your team from the repetitive — and make sure no customer is ever forgotten.

midasanalytics.ai
←→ navigate · space reveal