Process Optimization and Process Automation in Mid-Market Companies

Process optimization, automation, data ownership and user experience as the foundation for scalable companies

A company today can achieve a remarkable amount with very few people.

CRM, marketing automation, ERP, project management, scheduling, helpdesk, newsletters, forms, cloud storage, analytics and AI tools make it possible to digitize processes that used to require entire departments.

But at some point, the advantage tips over.

Ten specialized tools become twenty. Twenty become thirty.

And suddenly something paradoxical happens:

The company is digitized — but not integrated.

Employees copy information from the CRM into Excel. Sales manually transfers data from forms. Purchasing waits for information from sales. Project managers ask for status updates by email. Customer data exists multiple times, sometimes contradicting itself.

And every new feature brings another SaaS subscription.

The problem is no longer missing digitization.

The problem is missing process architecture.

The real challenge for mid-sized companies isn’t digitization — it’s connection

Germany has invested heavily in digitization in recent years. At the same time, execution remains difficult.

According to Bitkom, 53 percent of German companies reported problems managing digital transformation in 2025. Companies cited a shortage of skilled workers, lack of time, lack of financial resources, and technical security requirements among the central obstacles.

The picture among mid-sized companies is equally mixed.

The KfW digitization report shows that in 2025, only around 30 percent of mid-sized companies carried out digitization projects. At the same time, mid-market digitization spending reached roughly €23.8 billion in 2024.

That means:

Companies keep investing significant sums in technology.

But technology alone doesn’t solve a process problem.

The decisive question isn’t how many digital tools a company owns. What matters is how well those tools, data, processes and people are actually connected.

The real problem is called friction

In many growing companies, an invisible extra layer of work forms.

It doesn’t show up in any org chart or job description.

It’s the work that happens between systems.

A lead comes in through the website.

Then something like this might happen:

1. The lead lands in a form.
2. An employee reviews the inquiry.
3. The data is transferred into the CRM.
4. Sales adds more information.
5. A meeting is scheduled.
6. A note is created after the call.
7. Information is transferred into a project management system.
8. Purchasing receives a message.
9. A quote is created.
10. The quote later ends up in the ERP.
11. After the order is placed, data is transferred again.
12. The project team needs the same information.
13. Accounting needs yet other information.
14. In the end, the same customer data lives in five or six systems.

Each individual step looks small.

Together, they add up to a significant operational burden.

The most expensive processes are often not the visible ones. They’re the handoffs in between.

Internal processes eventually become the customer experience

Customers don’t see CRM systems, ERP landscapes or n8n workflows.

They only see the effects.

If an employee has to search three systems before answering a question, the customer waits longer.

If sales doesn’t know what the project manager already discussed with the customer, the customer has to repeat information.

If an order gets lost between sales, purchasing and project management, the customer experiences delays.

If a company internally holds five different versions of the truth about a customer, that customer eventually gets five different answers.

Customer experience therefore doesn’t begin at the surface. It begins in internal processes.

A good customer experience is often the result of good internal collaboration.

That applies to mid-sized companies just as much as startups: the more steadily a company grows, the more visible the weaknesses in its internal communication become.

What does “just copy it over quickly” actually cost?

Let’s take a deliberately conservative example.

On average, an employee only needs to spend 15 minutes a day transferring data between systems, searching for information, or correcting duplicate records.

Over 220 working days, that’s:

55 hours per year, per employee.

With ten employees, that’s already:

550 hours per year.

At an assumed fully-loaded internal hourly rate of €50, that equals:

€27,500 per year.

And that doesn’t yet account for errors, follow-up questions, delays, and lost value creation.

With 30 employees, under the same assumptions, it would already be:

€82,500 per year.

This isn’t a general statistic — it’s a simple example calculation.

And that’s exactly why it’s worth taking process friction seriously instead of dismissing it as a “small inefficiency.”

What costs 15 minutes a day can quickly turn into a five-figure sum at company level.

The hidden cost structure of manual processes

Manual data transfer doesn’t just create labor costs.

It creates at least five further cost categories:

1. Time costs

Employees spend time on transferring data instead of value-creating work.

2. Error costs

An incorrect data record can lead to wrong quotes, deliveries, invoices or decisions.

3. Coordination costs

The more systems are involved, the more coordination is required.

4. Opportunity costs

Sales reps sell less, project managers coordinate more, and executives search for information instead of making decisions.

5. Scaling costs

Effort increases with every additional employee.

This is the decisive point.

A poorly integrated process doesn’t automatically scale with the company — its complexity often grows faster than its actual benefit.

The SaaS problem: when digitization itself becomes a cost factor

The second development concerns the software landscape.

Cloud software and SaaS have democratized digitization.

A company can today set up a CRM, a helpdesk, a scheduling tool or a project management system within hours.

That’s great.

But SaaS has a downside:

Every tool initially optimizes its own slice of the company.

The CRM wants to manage leads and customers.

The ERP wants to manage goods, purchasing and finance.

The project management tool wants to manage tasks.

The marketing tool wants to manage campaigns.

The meeting tool wants to manage meetings.

The calendar wants to manage appointments.

And at some point, someone has to make sure these systems actually talk to each other.

On top of that, there are often:

– monthly base fees
– per-user costs
– additional costs for automations
– API limits
– premium integrations
– storage fees
– extra seats for external staff
– costs for reporting and analytics
– costs for AI features

German mid-sized companies already use cloud services broadly. In 2025, 65 percent of companies with 50 to 249 employees used paid cloud services. Among companies with 10 to 49 employees, it was 51 percent.

The cloud is long since part of the corporate infrastructure.

So the question isn’t:

“Cloud or no cloud?”

It’s:

“Which data belongs where — and who controls the flow of information?”

Data ownership doesn’t mean hosting everything yourself

Data ownership is often misunderstood.

It’s not about forcibly running every application on your own server.

It’s about deliberately deciding:

– What data is stored?
– Where is it stored?
– Who has access?
– Which systems are allowed to process it?
– How is it synchronized?
– How long is it retained?
– How is it deleted?
– How can I export it?
– What happens if I want to switch providers?
– What happens if a provider changes its pricing or business model?

For personal data specifically, GDPR requirements are clear: data must be processed for a defined purpose, limited to what’s necessary, accurate, and adequately protected, among other principles.

That doesn’t automatically mean self-hosting is compliant.

But self-hosting can give companies additional control over infrastructure, data storage and access models.

What matters is the architecture.

One central data foundation instead of five truths

The goal of a modern company architecture therefore shouldn’t be to connect as many systems as possible.

The goal should be:

To create one reliable source for business-critical information.

A so-called single source of truth.

For example:

A customer doesn’t exist five times.

A customer exists once.

CRM, ERP, marketing, support and internal processes access the same relevant information — each with clearly defined ownership.

That fundamentally changes the architecture.

Instead of:

CRM → Excel → ERP → email → project management → accounting

you get:

Central data foundation → connected systems → automated processes

The systems stay specialized.

But the data no longer gets duplicated arbitrarily.

Industry-specific process automation

Every industry struggles with its own media breaks and interface problems. See how intelligent process architecture and automation give you the decisive edge in your industry:

From sales to value creation: thinking about processes end to end

Good process optimization therefore doesn’t start with a single tool.

It starts with the company.

A typical end-to-end process might look like this:

Marketing → lead generation → website / landing page → form → CRM → lead scoring → sales → scheduling → call → quote → order → purchasing → project / production → delivery → invoice → payment → support → retention → repeat sale

If every department only optimizes its own section, interface problems emerge.

If the entire process is considered as a whole, it becomes visible:

– Where does information originate?
– Where does information get changed?
– Where does information get copied?
– Where does information get lost?
– Where does one person wait on another person?
– Where does a person decide something a machine could actually decide?

And above all:

Where does unnecessary friction occur?

A company is only as strong as its communication

A company isn’t just org charts, roles and departments.

It consists of relationships and information flows.

Sales talks to project management.

Project management talks to purchasing.

Purchasing talks to suppliers.

Marketing talks to sales.

Leadership needs information from every area.

The better this communication works, the easier it is for a company to make decisions and create value.

And this is where an interesting feedback loop appears:

A company’s structure often reproduces itself in the software it uses.

When departments work in strong isolation, separate systems tend to emerge too.

When communication mostly runs over email, email-based processes emerge.

When information is spread across spreadsheets, spreadsheet-based workflows emerge.

When nobody owns master data, duplicates emerge.

When decisions run informally through individual people, those people become human interfaces in the process.

Software, in other words, often mirrors the existing organizational structure — weaknesses included.

That’s why optimizing software alone isn’t enough.

You have to look at organization, communication, processes and technology together.

That’s why automation doesn’t start with n8n

A common mistake sounds like this:

“We need automation. Let’s install n8n.”

That’s roughly like building a conveyor belt before knowing what should actually be produced.

Tools aren’t the first step.

Understanding the process is the first step.

That’s why meaningful automation usually begins with a workshop.

Making processes visible

In a process workshop, the following typically becomes visible together:

– What processes exist?
– Who is involved?
– Which stakeholders are affected?
– What systems are used?
– What information is created?
– Where is data stored?
– What handoffs exist?
– What decisions get made?
– Where do waiting times occur?
– Where do errors occur?
– Which steps are manual?
– Which steps are redundant?
– Which steps could be automated?
– What information do employees actually need to do their job well?

This produces process maps, data flows and system landscapes.

The goal isn’t to run a nice workshop.

The goal is:

To create a shared picture of how the company actually works.

Because that picture often doesn’t exist yet.

Sales knows its processes.

Purchasing knows its processes.

Project leadership knows its processes.

Leadership knows its processes.

But nobody knows the whole process.

UX design: understanding the people behind the processes

Process optimization is especially successful when it doesn’t only look at the organizational and technical perspective.

It also has to understand the people who work with these processes.

This is exactly where user experience design comes in.

Over the years, employees build their own mental models of how their work actually functions.

They know:

– which information is really relevant,
– which exceptions come up,
– which shortcuts work,
– where systems slow them down,
– which workarounds they’ve developed,
– which information they’re missing,
– and where they regularly need to communicate with other teams.

This reality often isn’t reflected in process diagrams.

It becomes visible when you listen.

That’s why ThinkForm works with, among other things:

– Interviews
– Stakeholder interviews
– Team workshops
– Observation
– Process mapping
– Journey mapping
– Usability analysis
– Prototyping
– Feedback loops
– Coaching and enablement

The goal isn’t to convince employees of a finished solution after the fact.

They’re brought into the change process early.

A technically perfect solution can still fail

New software can work technically perfectly and still fail.

Not because the API doesn’t work.

But because the people who are supposed to use it don’t understand or accept it.

When a new process simply gets imposed over the existing reality of work, resistance follows.

And that resistance isn’t automatically a sign of unwillingness to change.

Often it’s a sign that the system doesn’t match the actual work.

That’s why human-centered design doesn’t mean asking employees for their opinion only at the end.

People need to be heard at the beginning.

Good digital transformation therefore considers:

– Business needs
– User needs
– Customer needs
– Technology
– Organizational constraints

Only then does a solution emerge that works in the long run.

Usability is an economic factor

Usability is often treated as a pure design question.

For companies, though, it’s directly economically relevant.

If an employee needs five steps instead of two for a task, that costs time.

If information is hard to find, follow-up questions pile up.

If a system is confusing, errors happen.

If employees build workarounds, shadow processes emerge.

If new software isn’t accepted, part of the investment goes unused.

Good usability doesn’t just reduce friction for users. It reduces operational costs.

That applies to internal tools just as much as to products and services facing customers.

The automation layer

Only after that comes the technical implementation.

This is where tools like n8n, webhooks and APIs can play a central role.

n8n, for example, can serve as a connecting automation layer between existing systems.

One example:

Website → webhook → n8n → validate lead → check for duplicates → update CRM → identify the right sales rep → create task → send notification → prepare scheduling → log information

Not every system needs to be replaced in the process.

It’s often more sensible to first connect the existing system landscape intelligently.

APIs instead of copy and paste

The most important technology here is often surprisingly unspectacular:

the API.

An API lets systems exchange information in a structured way.

That turns:

“Peter copies the customer from HubSpot into Excel.”

into, for example:

“A new customer in HubSpot automatically triggers a defined workflow.”

That’s a fundamental difference.

The first process needs a person.

The second process needs a rule.

HubSpot doesn’t have to be the center of the universe

HubSpot is actually a good example here.

CRM systems are extremely capable.

But they don’t automatically have to be the central database for every business problem.

A company can, for example, use HubSpot for sales and marketing while other information lives in ERP, project management, accounting or its own database.

What matters is the architecture:

Which information system is authoritative for which data type?

For example:

– Leads → CRM
– Customer relationships → CRM
– Products → ERP / inventory management
– Stock levels → ERP
– Orders → ERP
– Projects → project management
– Documents → DMS / storage
– Invoices → accounting / ERP
– Central master data → central data foundation

That prevents any one system from trying to be everything.

Open source as a strategic option

Another option is reducing SaaS dependency.

Not every piece of software has to be sourced as SaaS.

For certain use cases, self-hosted open-source solutions can be attractive.

For example, for:

– Scheduling
– CRM
– ERP
– Content management
– Document management
– Knowledge management
– Analytics
– Automation
– Communication
– Project management

This isn’t about using as many open-source products as possible.

That would just be the next form of tool chaos.

The decisive question is:

Where does self-hosting actually create an economic or strategic advantage?

Self-hosting isn’t an end in itself

A self-hosted system comes with additional responsibility:

– Updates
– Backups
– Monitoring
– Security
– Access control
– Encryption
– Resilience
– Disaster recovery

That’s why a self-hosting decision should always be considered economically and technically.

A company shouldn’t replace 20 SaaS subscriptions with 20 self-operated applications.

That wouldn’t be progress.

The better strategy is a deliberately designed hybrid architecture.

Cloud, SaaS, open source and in-house data storage each get used where they actually make sense.

The next step: the company brain

Once processes are properly integrated and data is structured, another possibility opens up.

A company brain.

This doesn’t mean some magical AI.

It’s really a central information architecture in which business-relevant information gets brought together and made accessible to people and machines.

For example:

– Customers
– Contacts
– Quotes
– Orders
– Projects
– Products
– Documents
– Processes
– Knowledge
– Communication
– Decisions
– Metrics

AI systems can later build on top of that.

At that point, an AI can not only ask:

“What does this document say?”

But, for example:

“Which open quotes do we have for customers whose projects are supposed to start in the next 30 days?”

Or:

“Which customers haven’t had contact in six months despite open potential?”

Or:

“Which processes currently cause the most manual handoffs?”

That’s the difference between an AI chatbot and an intelligent company architecture.

A company brain also needs a human interface

A company brain therefore can’t just be a data project.

People need to be able to understand where information comes from, why a recommendation gets made, and what action is expected next.

This transparency is becoming increasingly important, especially with AI systems.

An intelligent system that nobody understands or trusts doesn’t create real relief.

Technology doesn’t just have to work. It has to be understandable, accessible and usable for people.

That closes the loop between data architecture, UX design and artificial intelligence.

Automation first. AI second.

Many companies currently want to deploy AI immediately.

But AI can’t automatically fix bad processes.

If data quality is poor, the AI will work with poor data.

If information is stored redundantly, contradictory answers emerge.

If processes aren’t defined, AI can’t reliably decide what should actually happen.

That’s why the more sensible order is:

1. Understand processes
2. Structure data
3. Integrate systems
4. Automate routine work
5. Improve data quality
6. Apply AI on top of the existing company structure

Only then does a resilient company brain emerge.

Automation isn’t a job killer — it’s a scaling mechanism

Automation shouldn’t primarily be seen as a way to replace people.

It should free people from bad work.

McKinsey estimates that by 2030, generative AI could make around 27 percent of hours currently worked in Europe automatable. That doesn’t mean 27 percent of jobs disappear — rather, tasks and workflows change.

Especially relevant here are activities around:

– Data entry
– Information processing
– Documentation
– Standardized communication
– Data transfer
– Classification
– Research
– Administrative processes

That’s exactly where the invisible friction lives in many companies today.

A company shouldn’t copy faster

It should have to copy less.

That’s a decisive difference.

The answer to inefficient processes isn’t:

“We need better Excel templates.”

Nor is it:

“We need another CRM.”

Nor is it:

“We need another automation.”

The better question is:

“Why does this information have to be entered again at all?”

What makes a good company architecture

A modern, scalable company architecture should have at least six properties.

1. Clarity

Every important process is visible and understandable.

2. Data ownership

The company knows what data lives where and who has access to it.

3. Interoperability

Systems can communicate with each other via APIs and standardized interfaces.

4. Automation

Recurring tasks run reliably, based on defined rules.

5. Human-centered design

Systems are designed so employees and customers understand, accept and meaningfully use them.

6. Extensibility

New tools, AI systems or business models can be integrated without rebuilding the entire infrastructure.

The ROI often sits between the systems

The most interesting optimization potential therefore isn’t inside any single tool.

It sits between the tools.

Between:

– Marketing ↔ Sales
– Sales ↔ Purchasing
– Purchasing ↔ Project management
– Project management ↔ ERP
– ERP ↔ Accounting
– CRM ↔ Support
– Employees ↔ Information

That’s exactly where media breaks, follow-up questions, duplicate data and manual work emerge.

And that’s exactly where automation can have the biggest impact.

Process optimization is also organizational development

A process workshop can therefore reveal more than just technical workflows.

It often shows:

– where responsibilities are unclear,
– where teams work past each other,
– where information is missing,
– where decisions are unnecessarily centralized,
– where individual employees have become human interfaces,
– where departments hold different mental models of the same process,
– and where the actual organization diverges from the official org structure.

This is a decisive point.

A company’s digital transformation is always also a transformation of how it collaborates.

Changing only the software doesn’t automatically change the organization.

ThinkForm: designing processes for people and systems

ThinkForm connects user experience, customer experience, process design, technology, automation and AI.

With more than 15 years of experience at the intersection of UX, technology, innovation and digital product development, ThinkForm works with companies of different sizes and industries — from large organizations like Mercedes-Benz and REWE Digital to mid-sized companies and startups.

The work doesn’t start with the question:

“Which tool do we need?”

It starts with:

“How does your company actually work today — and where is friction happening?”

Through interviews, stakeholder interviews, team workshops and process analysis, the actual workflows, needs and mental models of employees become visible.

From there, processes get rethought, systems get connected, and automations get developed.

That can mean, for example:

– making existing processes visible
– restructuring processes
– optimizing CRM systems
– connecting HubSpot with other systems
– integrating APIs and webhooks
– building n8n-based automations
– building in-house data storage
– centralizing data flows
– reducing redundant SaaS solutions
– evaluating open-source alternatives
– developing self-hosting architectures
– connecting ERP, CRM and marketing systems
– improving data quality
– preparing digital processes for future AI applications
– improving usability and UX of internal tools
– involving employees in change processes through workshops and coaching

The goal isn’t the most complex technical solution possible.

The goal is the simplest architecture that makes the company scalable long-term — and that people actually enjoy and trust using.

When organizational structure gets mirrored in software

That’s why ThinkForm always looks at two levels at once:

The organization

– How do people communicate with each other?
– How do decisions get made?
– Where does responsibility sit?
– What information does a team need?
– Where do handoffs occur?

The technology

– How do systems communicate with each other?
– Where does data live?
– Which systems are authoritative?
– What information gets synchronized?
– Which processes are automated?

These two levels influence each other.

A poor communication structure can lead to poor software architecture.

A poor software architecture can, in turn, reinforce poor communication.

Both can stabilize each other.

That’s why ThinkForm’s job is to look at people, processes, data and technology together.

From process analysis to the company brain

A typical approach might look like this:

Phase 1 – Understand

Process workshops, interviews and analysis of the existing system landscape.

Phase 2 – Structure

Process modeling, data model, ownership and system architecture.

Phase 3 – Integrate

APIs, webhooks and automation workflows connect the existing systems.

Phase 4 – Centralize

A suitable data foundation creates a resilient basis for central company information.

Phase 5 – Automate

Recurring processes get automated.

Phase 6 – Optimize

Processes get continuously improved based on real data.

Phase 7 – Extend intelligently

AI and a future company brain can build on this structured foundation.

Digitization doesn’t start with software

The most important insight is perhaps also the simplest:

A company doesn’t become digital by buying more software.

It becomes digital when information flows reliably.

When a lead automatically reaches the right sales rep.

When an order automatically triggers the next processes.

When purchasing knows what needs to be ordered.

When project managers no longer have to hunt for information.

When accounting automatically receives the relevant data.

When leadership gets a reliable view of the company.

When employees understand and accept their tools.

And when the data belongs to the company — not to the logic of individual tools.

The real digital transformation is a question of architecture

The companies that will be able to grow especially efficiently in the coming years won’t necessarily be the ones using the most tools.

They’ll be the ones who understand their processes best.

Who own their data.

Who connect systems meaningfully.

Who improve communication.

Who put human work where it actually creates value.

And who design technology to support people instead of creating additional friction.

Not more software. Less friction.

Not more data copies. A reliable data foundation.

Not more SaaS subscriptions. A deliberate technology architecture.

Not AI for AI’s sake. A company structure AI can actually work on.

Not change imposed over employees’ heads. Digital transformation together with the people who live it every day.

Sources and further reading

Bitkom Research – Digitalisierung der Wirtschaft 2025: Representative survey of 603 companies with 20+ employees.

KfW Research – KfW-Digitalisierungsbericht Mittelstand 2025: Current data on digitization activity and spending among German mid-sized companies.

Destatis – ICT in companies / cloud computing 2025: Data on paid cloud service usage by company size.

Bitkom – Digital Office Index 2024: Data on the use of digital solutions and business process automation.

McKinsey Global Institute – A New Future of Work: Estimates on the automation potential of generative AI.

European Union – General Data Protection Regulation, Art. 5: Principles for processing personal data.