STEN.AI

AI business innovator · Sparring partner for boards

Using AI is easy. Turning it into results is leadership.

Sten Franke Founder & CEO, ETHORITY AI Transformation Architect

AI is changing how companies work, compete and grow. I work with boards and executive teams as an expert and sparring partner in this transformation: together we decide where AI creates value, choose the solutions and partners that hold up, and change the organisation so that it adopts AI, stays resilient and grows.

Sten Franke, portrait

Sten Franke and ETHORITY have worked for

  • BMW
  • Mercedes-Benz
  • Audi
  • Bosch
  • Siemens
  • Bayer
  • Roche
  • Boehringer Ingelheim
  • Merck
  • ZEISS
  • T-Mobile
  • Disney
  • Coca-Cola
  • DHL
  • E.ON
  • European Central Bank
  • Continental
  • Henkel
  • Telefónica O2
  • RWE
  • Honeywell
  • Sennheiser
  • Aldi Süd

Board agenda

Six AI topics that land on the board’s table.

AI has moved from the IT budget to the board agenda. It changes costs, risks, the operating model and the position in the market at the same time. In my work with executive teams, six topics come up again and again. For each one you see the market evidence, how to work on it in your company, a use case and where to start.

AI spending grows by half. Measured returns lag.

ETHORITY Insight on ethority.ai

37 %

of surveyed organisations attribute some effect on profit (EBIT) to AI, while about nine in ten use it regularly.

McKinsey, The state of AI, Aug 2026 · Good AI Prism Market Study 2026/27

01 Value

The AI budget grows. Its contribution to profit is unclear.

Most companies already use AI somewhere. Far fewer can show where it improves the result. The board question is no longer whether to use AI, but where it creates measurable value, from when, and which investment is worth scaling.

How to use it in your business

  1. Start from the strategy and the P&L: which costs, revenues or risks should change?
  2. Separate capacity from savings. Freed hours become money only when costs fall or revenue grows.
  3. Set a threshold before you invest: the value at which you expand, redesign or stop.

Use case · Scenario

A service organisation frees 1,000 hours a month with AI and uses 600 of them productively. At €50 an hour that is a modelled €30,000. After €12,000 for AI, review and rework, €18,000 remain. The board approves the expansion once lower costs or extra contribution can be measured.

Capacity value calculator

Productive hours per month
–
Modelled value per month
–
After costs, per month
–
After costs, per year
–
Set-up recovered after
–

Scenario calculator based on the method of the Good AI Prism keynote. The result is a modelled value. It becomes a financial result when you can measure lower costs or an extra contribution. More AI calculators on ETHORITY.ai

Where to start

AI Opportunity CaseKnow what would make an AI investment of about €200,000 or more pay.

AI spending grows by half. Measured returns lag. Read the analysis

961 AI offerings. Start with the job, not the tool.

ETHORITY Insight on ethority.ai

43 of 107

customer-service AI offerings publish a price. Prices appear in at least seven different units.

Good AI Prism, vendor information checked 19–20 Oct 2026

02 Choice

Every vendor promises the same. The board needs a choice it can defend.

The market moves faster than any buying process: new offerings, renamed products and prices in units that do not compare. A choice made on demos and slide decks is hard to defend later. A defensible choice starts with the task and compares offerings on dated facts with sources.

How to use it in your business

  1. Name the task and the result you expect before anyone names a tool.
  2. Open the task in the Good AI Prism and shortlist offerings with sources: capability, price model, data region and evidence.
  3. Compare the shortlist on the same volume and the same cases, and record why one option wins.

Use case · Scenario

A retailer has three offers for AI in customer service: one charges per user, one per conversation, one per resolved case. Converted to the same annual volume, the ranking changes. The board signs off on the offer with the lowest cost per accepted answer and an EU hosting option.

Where to start

AI Decision BriefCompare up to three AI offerings on dated facts before you buy.

961 AI offerings. Start with the job, not the tool. Read the analysis

Tool access is not authority.

ETHORITY Insight on ethority.ai

2 Dec 2027

is when EU duties for listed high-risk uses apply. Transparency duties have applied since 2 Aug 2026.

Good AI Prism Market Study 2026/27, exhibit 2.1 · EU AI Act, as amended

03 Accountability

AI agents act in your company’s name. Who answers for them?

Agents answer customers, change records and trigger payments. Access to a tool does not give an agent the right to act. Supervisory boards, auditors and regulators will ask who decided, what the system may do on its own and how you know that it works.

How to use it in your business

  1. List every AI system that acts or decides, each with a named owner.
  2. Scope each agent by action, limit, time and consequence.
  3. Keep the evidence your supervisory board can check: decisions, tests, incidents and results.

Use case · Scenario

An AI agent answers customer requests and triggers refunds. Five questions make it safe and worth it: who decides and carries the consequences, what may the agent do alone and up to which amount, which answers need proof, when does a person step in, and does the result really improve?

Five-question check for one AI agent

Where to start

AI Trust RecordGet a clear answer on one AI system: pilot, redesign, expand or stop.

Tool access is not authority. Read the analysis

AI-first is an operating model, not a tool count.

ETHORITY Insight on ethority.ai

1.6–4.3×

is how much more often large companies use AI than small ones. Missing expertise is the most cited barrier.

National studies compared in the Good AI Prism Market Study 2026/27

04 Organisation

You are hiring AI roles. The organisation still works the old way.

Many companies answer AI with new roles and new tools. Value appears when strategy, work, authority, knowledge, evidence and leadership change together. AI-first is an operating model, not a tool count.

How to use it in your business

  1. Find out where you stand on strategy, data, skills, governance and operations, in one workshop day.
  2. Redesign the work around a few objectives, with clear decision rights between people and AI.
  3. Build the skills your teams need to judge AI for themselves.

Use case · Scenario

A new CEO wants an AI agenda within 100 days. Instead of a tool list, the executive team agrees on three objectives, names an owner for each and decides which tasks people keep, which AI prepares and which run automatically. Hiring follows the operating model.

Where to start

AI Readiness & Maturity CheckOne workshop day, then a written assessment of where you stand.

AI-first is an operating model, not a tool count. Read the analysis

One model is not enough. Large enterprises already run three or more.

ETHORITY Insight on ethority.ai

1 in 5

organisations say operating costs already limit their use of AI. 26 % of large US firms see their AI costs in real time.

McKinsey, The state of AI, Aug 2026 · KPMG AI Pulse Q2 2026

05 Resilience

Models retire, prices change, rules tighten. Can you still act?

AI is becoming a running cost and a dependency. Model versions retire on fixed dates, price lists change within months, and one provider can sit behind many processes. Resilience means knowing which tasks must keep running and which alternatives are realistic.

How to use it in your business

  1. Map your AI estate: systems, providers, models, costs and owners.
  2. Mark the tasks that must keep running and the points where you depend on one provider.
  3. Keep a tested alternative for each critical task and weigh its cost against the cost of a disruption.

Use case · Scenario

A group runs customer service, contract review and reporting on one model family. A retirement date is announced. Because the AI estate is mapped, the CIO knows within a day which processes are affected, what a switch costs and which alternative is already tested.

Where to start

AI Estate Audit & MappingThe AI you use, what it costs and what to keep.

One model is not enough. Large enterprises already run three or more. Read the analysis

AI does not scale by itself. Redesign the work.

ETHORITY Insight on ethority.ai

44 %

of surveyed organisations report that they are scaling AI.

McKinsey, The state of AI, Aug 2026

06 Growth

A pilot proves little. Growth starts when you scale what works.

Many organisations run pilots. Fewer take them into daily operation and across teams, sites and countries. Scale follows the design of the work: one objective, the simplest mode per task and the cost of an accepted result. That is where AI starts to change growth, not only efficiency.

How to use it in your business

  1. Choose one use case with a clear objective and a measured starting point.
  2. Test it under real conditions for six weeks, with stop rules agreed upfront.
  3. Scale only what reaches the agreed threshold, to more teams, sites or countries.

Use case · Scenario

A machine builder tests an AI assistant for service technicians at one site. After six weeks, the time to the right solution and the number of follow-up questions are measured against the baseline. The threshold is met, and the board approves the roll-out to all sites with the same measures.

Where to start

AI Trust PilotRun one AI use case under real conditions, with stop rules agreed upfront.

AI does not scale by itself. Redesign the work. Read the analysis

The Good AI Prism by Sten Franke and ETHORITY: AI offerings by task, industry, data type and technology layer
Good AI Prism v3.12 · Sten Franke & ETHORITY

Why the choice decides

More money, more offerings, few comparable facts.

  1. +49.5 %

    worldwide AI spending forecast for 2026, to US$2.67tn

    Gartner, Sep 2026

  2. 37 %

    of surveyed organisations attribute some effect on profit to AI

    McKinsey, Aug 2026

  3. 43 of 107

    customer-service AI offerings publish a price

    Good AI Prism, Oct 2026

So the lever is the choice: the right task, the right solution, the right partner and clear accountability. The Good AI Prism maps the market by task and industry, each offering with a source and a date. I turn that map into decisions your board can stand behind.

Three questions lead to the AI that fits

  1. Where does AI help in my work?

    Open the Prism
  2. What does AI change in my industry?

    Industry studies
  3. Which AI fits our decision?

    Compare offerings

Read the evidence

How I work

From orientation to results, in five steps.

This is the operating loop of ETHORITY, the AI Operating System I am building: it connects AI, data, people, decisions and execution.

  1. 1

    Discover

    Find where AI can help your business and what the market offers for it.

  2. 2

    Create

    Choose the option and build the solution, including its costs, data and responsibilities.

  3. 3

    Run

    Operate the system day to day, with named owners and agreed checks.

  4. 4

    Improve

    Measure the result against the starting point and adjust what does not work.

  5. 5

    Grow

    Extend what works to more teams, sites or countries.

Good AI is

  • Effective
  • Efficient
  • Economic
  • Responsible

“I help you make sense of the AI market, take sound decisions and shape the implementation — combining market intelligence, systems architecture and economic evaluation.”

Sten Franke · Founder & CEO · AI Transformation Architect

ETHORITY

Where to start

From a 15-minute conversation to a test in your operation. Each ETHORITY.ai page lists scope and terms.

15 minutes with Sten

You are deciding where to begin. We clarify your question and discuss one next step.

Request 15 minutes
AI Decision Brief

AI Decision Brief

You are comparing vendors. Get a written recommendation covering up to three options, with sources.

See the sample and offer
AI Business Blueprint

AI Business Blueprint

You have a goal and need a plan. Get a roadmap for 30, 60 and 90 days.

See the roadmap

AI Opportunity Case

A larger investment is on the table. Clarify the economics, risks and the evidence you need to release budget.

See the business case

AI Trust Record

A system is nearing deployment. Document the evidence, the open questions and the accountable decision.

See the Trust Record
AI Trust Pilot

AI Trust Pilot

You need to know if an application works in your operation. Test one use case with explicit acceptance criteria.

See the pilot
Sten Franke on stage at DFKI in Berlin, 2018
DFKI, Berlin, 2018

Keynotes, panels, interviews

On stage, Sten Franke turns market evidence into board decisions.

Founder and CEO of ETHORITY and creator of the Good AI Prism. Talks in German and English, for boards, leadership offsites, conferences and the media.

  1. Keynote: What boards must decide about AI

    Three questions for the next AI investment: what must change, who decides, and what result justifies the spend?

  2. Interview: Why a good AI pilot gets stuck

    What business and IT teams should measure on real cases before moving into production.

  3. Panel: Who has authority when AI agents act?

    Practical limits on what agents can do, and who answers for errors and results.

Track record

Twenty-five years, told from today.

Market intelligence, digital strategy and innovation since the early web. The full story is on About Sten.

  1. Today · ETHORITY

    Sten Franke · Studio, 2024

    2024

    Studio, 2024

  2. Sten Franke · brightonSEO, San Diego

    2023

    brightonSEO, San Diego

  3. 2005–2021 · ethority

    Sten Franke · DFKI, Berlin

    2018

    DFKI, Berlin

  4. Sten Franke · Portrait, 2017

    2017

    Portrait, 2017

  5. Sten Franke · Portrait, 2015

    2015

    Portrait, 2015

  6. Sten Franke · ethority, Hamburg

    2013

    ethority, Hamburg

  7. A cake for the Prisma, and the office dog

    2012

    A cake for the Prisma, and the office dog

  8. Sten Franke · MarKom, Munich

    2008

    MarKom, Munich

  9. Sten Franke · BITKOM CE Forum, Sony Center Berlin

    2007

    BITKOM CE Forum, Sony Center Berlin

  10. 1997–2005 · mediatime

    Sten Franke · mediatime, early 2000s

    2000s

    mediatime, early 2000s

Why leaders work with Sten Franke

  1. 01

    He knows the market.

    He has mapped technology markets since 2009 – today 961 AI offerings in the Good AI Prism.

  2. 02

    He builds systems.

    He founded software platforms such as Gridpatrol and today designs AI systems that run in production.

  3. 03

    He knows large organisations.

    Earlier projects for BMW, Deutsche Telekom, Disney, Daimler and the ECB show experience working with large organisations.

  4. 04

    He answers personally.

    He personally leads selected projects and brings in specialists from his network when needed.

Mission and principles

My mission: AI that works for people. Responsible, effective and human-centred, with evidence, sound economics and a person who answers for the result.

Ten principles

Ten convictions on people and AI

For me, keeping the whole in view means judging value, people and the ability to act together. That is where these ten convictions on people and AI come from. I want to develop them further with others.

Sten Franke · Founder & CEO · AI Transformation Architect · November 2026

  1. 01

    People define the purpose.

  2. 02

    Responsibility needs a name.

  3. 03

    Experience belongs at the decision table.

  4. 04

    People who are affected need to be heard.

  5. 05

    A sound future needs capabilities of your own.

  6. 06

    Productivity must benefit people.

  7. 07

    Trust needs reasons that can be checked.

  8. 08

    Resilience needs real choices.

  9. 09

    Resource use counts too.

  10. 10

    Good leadership revisits its decisions.

Find Sten Franke

One person, one network of work.

Follow Sten Franke on LinkedIn and read the work behind him on the ETHORITY sites. Each address keeps one part of the story.

The ETHORITY Ledger: AI. Social. Search. Made by people.
  • LinkedIn · Sten FrankePosts, talks and news, newest first.
  • ethority.aiWork with ETHORITY: the Good AI Prism, research and offers.
  • ethority.orgMission, principles and the public ledger.
  • The ETHORITY LedgerAI. Social. Search. Twenty entries from today back to 1997.
  • ethority.netThe archive of articles and case studies.
  • media-time.deWhere it began in 1997.

Bring the decision that is on your table.

Tell me what you have to decide, who is affected and what a good result would look like.

For journalists: the newsroom with press kit.