SOPHIA

Free analysis — For the Mittelstand

Order to cash
Your competitor now uses AI

Keeping pace is easier than you think. On your real ERP and inbox data, we calculate where your money sits idle between inquiry and payment — and what fixing it is worth.

Book a call See a sample analysis
Free of charge No obligation No changes to your systems Read-only access

01 — Why now

For twenty years, improvement meant making people faster.
Now machines finish entire processes on their own

The new question: which processes should you still touch at all? Three numbers that appear in no report, because they are scattered across ERP and inbox:

F/01

How many of your inquiries sit longer than 24 hours?

Counted, felt. Whatever sits for a day competes with someone else’s quote.

F/02

How many hours a year does your team retype orders?

Always the same, usually by email and PDF. This work shows up in no cost center.

F/03

How many days pass between delivery and invoice?

Those days come before any payment term. Nobody but you can win them back.

You get these three numbers from us.

02 — What sets us apart

AI agents on the
systems you already run

No system change, no migration, no year-long project.

U/01

We bring together what belongs together

One route from the email into the ERP, instead of an island next to your systems.

U/02

No system change. Live in weeks.

Your team keeps working in email and ERP.

U/03

We charge for work done only

Only successfully completed processes count. If the work is not done, it is free.

U/04

Sovereign intelligence

Model-agnostic, hosted in Germany. Your data stays here — GDPR and EU AI Act included.

U/05

No lock-in

What we build belongs to your company. It keeps working if you swap the system underneath.

U/06

Start small, grow large

One process after another, in your order. Pace and budget stay with you.

03 — Your first step

On your data.
Calculated, in the concrete

Based on our experience, your master data and your customer correspondence, we look together at what profit AI can add for your order desk and back office — along the whole route:

01

Inquiry

How many arrive, how fast they get answered, how often one slips through.

02

Quote

How long until the quote, and how many email rounds it takes.

03

Order

How long until order entry, and how much of it is typing.

04

Delivery

How often customers had to ask where things stand.

05

Invoice

How many days from delivery to invoice, and what they cost.

06

More revenue

Which existing customers are going quiet, and what demand sits unanswered in the inbox.

Two sources, both read-only

Your ERP. The document chain from inquiry to invoice, with timestamps. We write nothing back.

Your inbox. The correspondence around the same cases. The ERP alone only shows what was entered after the fact.

What you do

Have read-only access set up. Sign an NDA and a data processing agreement before we see anything.

If read access is off the table, we run the same route on your estimates. Weaker, but honest.

04 — Our commitment

A free analysis of your
email and ERP history

01 · Potential
Where time and revenue sit idle from inquiry to cash — from twelve months of data.
02 · Next steps
A ranking instead of a wish list: which process first, and which does not work today.
03 · ROI estimate
One estimate per process with your volumes, times and full costs. Clearly labeled as an estimate.
04 · Rollout
One process after another, on your ERP. No project that stalls your day-to-day business.

05 — What you get

A sample analysis,
stage by stage

This is what the result looks like. The assumptions are printed next to every number, because a number without them is worth nothing.

Order-to-cash potential analysis

Sample company — €8.0M revenue · 42 employees · ERP grown over the years

Example · invented numbers

68.000 €

Costs per year you already carry today

121.000 €

Revenue opportunity per year, unrealized today

118.000 €

Liquidity, freed up once

Three different things, so no grand total: saved costs, possible revenue and freed-up capital cannot be added together.

01 Inquiry — How fast do you reply, and what never gets a reply at all? Revenue opportunity€48,000
38 %of inquiries sit longer than 24 hoursOn your ERP today

What we see

Inquiries run through three mailboxes nobody sees side by side. The fastest reply went out after 20 minutes, the slowest after nine days. Eleven inquiries were never answered.

What would be possible

Instant confirmation with a case number, routing to the right person, and a list of what sits idle today.

How we get there

1.800inquiries per yearfrom your data
38 %of them slower than 24 hours, i.e. 684from your data
3.500 €average order valuefrom your data
2 pointshigher win rate when replying within 24 hoursassumption

684 inquiries × €3,500 × 2% = around €48,000 in order value per year

Base assumptions here

The two percentage points are the only value here we cannot measure. We set them together with you in conversation, and we tell you how sensitive the result is to them. Only the slow path is counted — inquiries you already answer quickly do not appear in this number.

02 Quote — How many emails go back and forth before you can write a quote? Cost savings€23,000
3,1email rounds until all details for a quote are inOn your ERP today

What we see

In 62 percent of cases the same three details are missing: quantity, requested date, delivery address. The follow-up question goes out, on average, only the next working day.

What would be possible

Spot the missing detail and ask right away instead of the next morning. One round stays with a person; the two before it do not.

How we get there

1.800inquiries per yearfrom your data
2,1avoidable email rounds per inquiryfrom your data
6 minhandling time per roundassumption
62 €full cost per hourfrom your data

1,800 × 2.1 × 6 minutes = 378 hours × €62 = around €23,000 per year

Base assumptions here

The six minutes per round are an estimate: reading, understanding, replying, finding your way back in when the answer arrives. We remove 2.1 of 3.1 rounds — one round always remains, because a person looks at the quote in the end.

03 Order — How much of your order intake is pure typing? Cost savings€33,000
4.100orders per year arrive by email and PDF rather than EDIOn your ERP today

What we see

71 percent of them are repeat orders from existing customers, mostly the same items in the same structure. They are transferred into the order by hand.

What would be possible

Order from email and PDF straight into the system. Approval stays with a person; the retyping does not.

How we get there

4.100orders per yearfrom your data
11 minentry time per orderfrom your data
71 %share of structurally identical repeat ordersfrom your data
62 €full cost per hourfrom your data

4,100 × 11 minutes = 752 hours, of which 71% = 534 hours × €62 = around €33,000 per year

Base assumptions here

The 11 minutes are derived straight from your timestamps: time between email arrival and order entry, cleaned of nights and weekends. The 29 percent of special cases deliberately stay with a person — we do not calculate them away.

04 Delivery — How often do your customers have to ask where their order stands? Cost savings€12,000
2.900status inquiries per year — questions about something the customer could not seeOn your ERP today

What we see

Every fourth customer email is a status question. On 340 orders the question arrived before your own company knew the date had slipped.

What would be possible

The customer hears about the delay from you — before they have to ask. That saves the inquiry, and it is the moment customers gain or lose trust.

How we get there

2.900status inquiries per yearfrom your data
4 minhandling time per inquiryassumption
62 €full cost per hourfrom your data

2,900 × 4 minutes = 193 hours × €62 = around €12,000 per year

Base assumptions here

The four minutes include looking things up in the system. The real value is not in this number: what a customer thinks of having to ask twice cannot be priced in euros — and we do not pretend it can.

05 Invoice — How many days pass between delivery and invoice? Liquidity€118,000
6.4 dayson average from delivery to invoice — your in-house best is 1 dayAfter connection

Freed up once, plus around €7,000 in interest effect per year.

What we see

These days come before any payment term. They are not customer behavior and not a dunning topic — they arise in your own house. In 90 cases the invoice sat longer than three weeks.

What would be possible

Invoice right after the delivery note, with a human check only on deviations. Shortens your capital lock-up without a single customer changing anything.

How we get there

5.4 daysreduction from the current 6.4 to your own best of 1from your data
€8.0Mannual revenuefrom your data
6 %cost of capitalassumption

€8.0M ÷ 365 × 5.4 days = around €118,000 less capital tied up, × 6% = around €7,000 in interest effect per year

Base assumptions here

We calculate against your own best value instead of an industry benchmark — one day is demonstrably possible in your house, because you already do it. The 6 percent cost of capital is a set value; name your real rate and we calculate with that. What happens afterwards — payment terms and dunning — is not our field. This calculation ends where your invoice goes out.

06 More revenue — What demand sits in your emails that never turned into a quote? Revenue opportunity€73,000
210candidates in twelve months: asked for, never quotedAfter connection

What we see

Customers asked about items, quantities or dates that never made it into a quote. For 34 existing customers the order rhythm changed and nobody followed up.

What would be possible

The demand sits in your inbox, in so many words. Nobody reads it, because nobody reads 40,000 emails a year.

How we get there

210candidates in twelve monthsfrom your data
3.500 €average order valuefrom your data
10 %share that converts to an order when followed upassumption

210 × €3,500 × 10% = around €73,000 in order value per year

Base assumptions here

The 10 percent is deliberately set low. With warm existing customers, follow-up conversion tends to run higher, but we have no evidence for that in your data. This is the only place where we claim revenue that does not exist today — every other number is a cost you already carry.

07 Pricing — And what does not work today?  not quantifiable
400special cases per year where the price is not in the systemNot today

What we see

The costing hangs on the experience of two people. It is written in no rule, no table and no field.

What would be possible

Nothing we could seriously promise today. As long as the rule is written down nowhere, a machine cannot apply it — and a system that guesses here costs you more than it saves.

Base assumptions here

This stage is in the analysis on purpose. A potential analysis in which everything works is a sales brochure. If you want to write the rules down, that is its own undertaking — we tell you which cases it would concern, but we do not sell it to you here.

This is a sample company.
Yours looks different

The same seven stages with your cases, your times, and assumptions we set together.

Run it on your numbers

What the analysis is, and what it is not: it belongs to you even if you never speak to us again, and every assumption is printed in plain sight. It is no offer, no price sheet and no assessment of individual employees. We see where time sits idle — not why.

06 — What else sits in the same data

More revenue.
Beyond the savings

Order to cash ends at the money; your correspondence does not. The same data shows four revenue levers.

H/A

Repeat and add-on demand

Asked for, never quoted. Usually it sits in an email, word for word.

H/B

Campaign triggers

What each customer group is asking about more often, before the revenue statistics show it.

H/C

Customers going quiet

Existing customers whose order rhythm has changed. Visible in the pattern long before annual revenue shows it.

H/D

The same question, every time

Where your team keeps rewriting the same answer. The template for your own knowledge base.

07 — The platform behind it

The analysis is the beginning,
the platform is the product

What comes after the analysis runs on the same platform: agentic processes on the systems you already have.

P/01

An agentic platform for the whole company

The order desk is the first process. The same platform carries the next areas of your company.

P/02

One hub for ERP, DMS and CRM

One access point instead of three. The processes run directly through your existing systems.

P/03

MCP server

Sophia speaks the protocol AI systems use to dock onto business applications. Including your own.

More on the platform

08 — The next step

Thirty minutes — and the first step
is done

No sales call with a follow-up appointment. You speak directly with the person running the analysis. Bring a real customer inquiry.

Simon Tüchelmann

Simon Tüchelmann

Founder and CEO

Founder and CEO of Sophia. Over 14 years in manufacturing, raised in the German Mittelstand. Has been building technology for companies running on grown ERP systems for years.

DF

David Felsmann

Head of GTM

PLACEHOLDER, to be provided by David: role and years, stations that speak to the audience — and why he can run this analysis.

→/01

Book a call

Thirty minutes, free of charge. Which of us takes it, we sort out internally.

Book a call

→/02

Prefer to write first

Two sentences: which ERP, and where things jam between inquiry and invoice. You get an assessment back, no appointment needed.

Send a short email

→/03

Not there yet

Four times a year, what we learned from the analyses: which process pays off, what works on which ERP. No advertising, no sales call.

Join the report