About us
Small team. Serious system.
Heramind builds business software for entrepreneurs: a modular ERP with everything around it. You work with a dedicated team of two or three people who know everything about your system. You have at most two fixed points of contact. No ticket desk, no chain of account managers: the people you talk to build it with you.
01How we work
The same people, from start to finish.
01Dedicated team
Two or three people, from start to finish.
The same people for setup, go-live and maintenance. They know your processes and your figures.
02Flat organisation
Your question reaches the people who build.
No layers in between. A question or request is picked up by someone who knows the system.
03Custom work is regular work
Your process, not ours.
The system is built in modules. An adjustment for your company is regular work for us, not a separate project.
02Built, not promised
What is inside. Counted, not estimated.
- 14
- packages in the modular ERP
- 5
- languages the ERP works in
- 2
- servers of our own in Europe, with encrypted backup
- 7
- guardrails on every AI, a person decides
Built on experience in finance since 2014, and in software since 2021.
Every change is made under four eyes, with a trail of who did what and when. Every new version is measured first, before it goes live.
03The team
The people you talk to build it with you.
We did the work ourselves. That is why we build it differently.

Emanuel Boateng
Founder and architect
Conceived and builds Heramind. Has worked in finance at large organisations since 2014: credit management, financial control and finance transformation. In every assignment he took the manual work out, first with macros and scripts, later with his own software, and he worked in and on the large ERP systems. Has been writing full-scale software since 2021. Lean Six Sigma Black Belt. He designs the system and is your first point of contact.

Mujeeb Ur Rehman
Software engineer
Full-stack developer with a bachelor’s degree in software engineering. Works on the web screens, the integrations and the automation of workflows.

Jerry Asamoah
Onboarding and training
Helps your team find its way in the system. Years of experience onboarding and supporting colleagues and explaining rules and processes clearly. Legally trained.
04How we work
Ten principles. The same every day.
- 01
Measure first, then build.
We start from a real entry, invoice or step and record the starting point.
- 02
One owner per part.
Every product and every file has one owner. Work is claimed before it starts.
- 03
Every release has a way back.
Before anything goes live, we record which version returns if it goes wrong.
- 04
A second look at everything that matters.
Changes to access, money, privacy or data get an independent review.
- 05
AI speeds up, a person decides.
Paying, sending, publishing and going live only happen after a person says go.
- 06
What we promise is on a list.
Every promise shows where it runs. What we show is running.
- 07
Real figures or no figures.
No invented clients or logos. Every number on this site has a source.
- 08
You talk to the team that builds it.
No salesperson in between. Whoever hears your question also builds it.
- 09
Export is always free.
Your data is yours: everything out, in an open format, at any time.
- 10
Calm in the work.
Short lines, a fixed order and every mistake only once: what we learn, we write down.
05The harness
One platform. Everything around it measured.
This is how it fits together, as it really runs. Every change follows the same line, from building to the way back.
- Two own serversin Europe, products and backup apart
- AI with seven guardrailsproposals with sources, a person decides
- Arvectand what comes with it
- Access with rolesfour eyes and a hash chain
- Nightly backupencrypted, with a restore test
- Lessons registerread before every task
The release line
- 01Build
- 02Measure
- 03Independent review
- 04Deploy
- 05Way back
06The guardrails
Seven layers. Each one testable.
01 · Purpose and limits
What may AI be used for here, and what not?
- What you get
- A register of AI use cases with an owner, a purpose and the tasks that deliberately stay outside AI.
- How we work ourselves
- A language model never takes a financial or legal decision for us and never sends an automatic payment reminder.
- Reference frame
- EU AI Act (risk classes), ISO/IEC 42001
FOLLOW ONE REQUEST
How the guardrails work in practice.
Pick an example and press Start. Fictional data.
AI drafts a reminder for an open invoice.
Log
AI use is running ahead of the rules.
Your people already use AI. Supervision is getting stricter. Set the limits now.
- ± 2/3of surveyed office professionals used AI tools their employer had not approved. PagerDuty, June 2026 (international)
- 79%of surveyed organisations have no tested kill switch for AI. Kiteworks, 2026 (international)
- 2 Dec 2027New date for EU obligations on stand-alone high-risk AI after the Digital Omnibus. AI literacy has applied since 2025. Gibson Dunn analysis, 2026
Figures from public research. They indicate a direction and are not legal advice.
We apply this to our own work first.
Our own rules, every day.
- 01Several models, each with its own task Claude, OpenAI, Grok, Cursor and four local models on our own hardware. Sensitive material stays local.
- 02One workspace with sources Pandora OS brings knowledge, decisions and handovers together, so context does not disappear into separate chats.
- 03One writer per component Two AIs never work on the same file at the same time. It stays clear who changed what.
- 04A go for each action Publishing, sending, paying and deploying only happen after explicit permission, every time.
- 05Secrets stay out of sight Passwords, keys and codes never appear in logs, chats or handovers. Private data is encrypted.
- 06Independent review Work with material risk gets a separate review before it moves on.
This describes our own way of working. It is not a client result and not a certification.
What we deliberately do not promise Honest about limits.
- 01No 100% safety. We design so that a mistake is caught before it does harm.
- 02No certificate or legal declaration of conformity. Standards are our reference; legal review is for a lawyer.
- 03No fully autonomous AI for decisions about money, people or rights.
- 04No vague results: you see what has been tested and what has not.
Start small, grow with evidence.
Scope and price are agreed after an introduction.
- 01AI guardrail scan In one to two weeks: where AI is used, where the risk is and what to fix first.
- 02Setting up guardrails We set it up, test it and hand it over to your team.
- 03AI literacy by role Short sessions per team, with your own examples.
The film
The family, with a purpose.

Watch the film (with narration)
A SHARED PURPOSE
What we build should make room.
For work that matters. For your talent. For the people who count on you.
For Emanuel, building a business is a way to serve people, rooted in his Christian faith. We work towards better working lives, fair opportunities and loving care for children. Each brand expresses that purpose in its own way, welcoming people of every background.
Room to build a business.
Software, process knowledge and automation that make everyday work simpler.
Room for your talent.
An honest picture of who you are, with direct contact around suitable work.
Room for parenthood.
Less organising around nappies, more attention for life with your child.
Room to be a child.
A safe foundation and caring adults. With Shelters of Grace.
Show us where things get stuck.We show you how it runs.
A conversation with the team that builds it, about your company and what you need.