AKENI.AI · for software companies & tech teams

Plan the team after AI.
Grow the people into it.

Coding agents change what a software team spends its time on, who becomes the bottleneck and which skills matter. We build two tools for this. Team Developer plans the team. Skill Coach helps each person build the new skills inside their daily work.

Explore the Two Tools Talk to Us
Team Developer · the team plan
Skill Coach · the daily practice
scroll

The question we work on

What happens to your developers after AI? And how do they get there?

Licences alone don't change what a team delivers. Two things have to happen. The work has to be redistributed so that freed hours become output. And people have to learn a new way of working, in the middle of their real tickets.

01

AI frees hours, but output stays the same

As long as responsibilities don't change, the tools cost money and nothing more is delivered.

02

The bottleneck moves

When agents write more code, specification and acceptance become the limit, not the developers.

03

New skills decide the pace

Pair programming with agents, reviewing their output, writing specs. These skills are learned by doing.

What our insights are based on

Three Years Inside Real Development Projects

Neither tool started from a theory about AI. Both started from what we saw in the development projects we worked in, and from how the job descriptions of the clients we supported changed while we were in them.

These are observations from our own practice. They are not a study and not a representative sample.

3 years hands-on AI in development projects
Weeks to 2.5 years project durations
2 to 8 developers team sizes

Industries

ERP E-commerce Booking platforms in medical diagnostics Construction Software for recruiting and project management

Source one

From the Project Work

Tasks moved before team sizes did, and the waiting moved with them

What AI took over was rarely a whole job. It was parts of tasks, spread across several people. What changed first was who did the specifying, the reviewing and the integrating, not how many people were on the team.

Once agents wrote more of the code, the queue formed at the two ends of the work: at writing a specification precise enough to hand over, and at accepting what came back. The people who had been fastest at coding were not automatically good at either.

The new skills showed up in daily work, not in training slots

Pairing with an agent, reading a diff nobody on the team wrote, deciding what to delegate and what to keep by hand. People picked these up in their own tickets, or they did not pick them up at all.

The stumbling blocks were organisational, not technical

Licences were never the obstacle. What slowed teams down was unclear ownership of AI output, review habits built for hand-written code, and no time set aside for learning inside real work.

Source two

From Job Ads and Task Descriptions

Role profiles are being rewritten around judgement, not typing

In the job ads and task descriptions we helped clients write, lists of frameworks moved into the background. What moved forward was specifying work, reviewing output and working with agents.

The job titles usually stayed the same. The responsibilities underneath them did not. That is why our catalog describes people by the tasks they can take on rather than by the title they hold.

We treat this as the weaker of the two sources: an ad says what a company is looking for, not what its team already does. The example below shows the pattern.

The Same Role, Written Two Ways

Fictional example · not a client document and not a real job ad

The ad a few years ago

Fullstack Developer (m/f/d)

  • Strong hands-on experience with React, TypeScript and Node.js
  • Implements features from tickets, end to end
  • Writes unit and integration tests for their own code
  • Takes part in code reviews within the team
  • Experience with REST APIs and relational databases

The ad today

Fullstack Developer (m/f/d)

  • Confident with React, TypeScript and Node.js, including when an agent writes the first draft
  • Turns requirements into specifications an agent can work from
  • Reviews AI-generated changes for scope, tests, edge cases and security
  • Decides what goes to an agent and what is better written by hand
  • Experience with REST APIs and relational databases

Same title, same stack. What moved is where the judgement is needed.

The catalog behind both tools — the tasks, the roles people move into and the skills those roles need — is built on these observations. Every assumption in it stays editable, so you can replace ours with your own. See it at work in Team Developer and Skill Coach.

Two tools

One for the Team, One for Each Person

Team Developer answers what should change in the team. Skill Coach helps each person make that change in their everyday work.

For CTOs, engineering leads & leadership

Team Developer

Simulates your team before AI, after AI and after redistributing the work.

  • AI share per task: what agents take over and what comes back as review.
  • The bottleneck: whose extra hour would be worth the most, and who could relieve them.
  • Takeovers: who takes over which task, with the learning and guidance it costs and what must be proven first.
  • Future roles & skills: for every person, ranked by the hours that depend on them.
  • Monthly result: output, staff and tool cost for every scenario and team variant.
Proof of conceptREST API + notebookNo LLM in the matching
Explore Team Developer
For developers & team members

Skill Coach

A personal coach on the employee's own machine that turns a learning path into daily practice.

  • Coaching moments: notices situations on screen where a skill on the path applies.
  • Small experiments: 10 to 30 minutes inside a real ticket, with a prompt that is ready to paste.
  • Evening review: five minutes on where the time went and one workflow change for tomorrow.
  • Evidence: a reviewed diff, a spec or an automated test, shared only if the employee chooses.
  • Fully local: screen capture, language model and data stay on the laptop. No manager view.
PrototypeLocal LLM via OllamaNo team view, by design
Explore Skill Coach

How they connect

From Team Plan to Daily Practice, and Back

The two tools share one catalog of tasks, roles and skills. The team plan becomes each person's learning path, and what people can show flows back into the plan.

Team Developer

Plan the team

AI share per task, the bottleneck, takeovers, and a future role for every person.

Team Developer → Skill Coach

Learning path per person

The skills to learn, ranked by the hours that depend on them, with learning and coaching hours.

Skill Coach

Practice in real work

Nudges at the right moment, a focus skill each day and a short evening review, all on the employee's machine.

Skill Coach → Team Developer

Evidence, when shared

The employee can export proof of a skill in the format Team Developer uses for takeovers.

↺ A takeover in the plan only counts once there is evidence. Skill Coach helps people produce it in their daily work.

What both tools share

Four Principles

Both tools follow the same four principles.

1

Evidence over job titles

Records of past work contain no evidence for tasks someone has never done. That's why takeovers come with observable requirements, a work sample and a guided handover, and why the coach focuses on things people can show.

2

Honest about the limits

We show the uncomfortable parts too. AI work comes back as review, redistribution alone earns nothing, and a plan can be fragile. Both product pages list their limitations openly.

3

People have a say

In Team Developer, tasks someone wants to learn get a bonus, and tasks someone refuses are never assigned to them. In Skill Coach, the employee owns the data and decides what is shared.

4

A model you can argue with

AI shares, learning hours and values are editable assumptions in a shared catalog. They are not hidden inside a black box. Change one and you can see what follows.

AKENI team at work
AKENI team collaboration
AKENI team meeting

About us

The AKENI Team

At AKENI.AI we work on one question: how AI changes the work of teams, task by task and skill by skill, and what that means for the people doing it.

Our core team is led by founder Nils Keßler and focuses on conceptual work and project management. We work with a network of data scientists, machine learning engineers and workforce development consultants. We combine AI-based tools with statistical methods and classical machine learning.

Over the past three years we have worked inside development projects that ran from a few weeks to two and a half years, with teams of two to eight developers, in ERP, e-commerce, booking platforms for medical diagnostics, construction, and software for recruiting and project management. That track record is what the assumptions behind both tools rest on.

Our focus right now is software companies and tech teams. That's where coding agents are changing the work fastest. Team Developer and Skill Coach are our two answers: one for planning the team, one for supporting each person.

Start with a conversation

Where Does Your Team Stand?

In a free first conversation we'll talk about your team, how you use coding agents today and which of the two tools would help first.

Book a Free Conversation

Get in touch

Talk to the Person Who Built Them

No sales call. You speak directly with the founder about your team and whether Team Developer, Skill Coach or both fit your situation.

Nils Keßler, AKENI founder
Nils Keßler
Founder & AI Consultant, AKENI
Write to Us or email info@akeni.de

FAQ

From three years of working inside development projects, from a few weeks to two and a half years, with teams of two to eight developers, across ERP, e-commerce, booking platforms for medical diagnostics, construction, and software for recruiting and project management. A second source is how job ads and task descriptions changed at clients we supported. These are observations from our own practice, not a study. What our insights are based on.
No. Team Developer is a proof of concept and Skill Coach is a prototype. Both work, and both list their limitations openly on their pages.
Team Developer works on its own. Skill Coach works from a learning path, which comes from Team Developer. The Skill Coach prototype ships with an example path, so a live Team Developer is optional for trying it out.
No. The example figures, the nine-person example team and the scripted example day are all fictional. They show how the tools reason, not what your team will achieve. What is not fictional is the catalog underneath: the tasks, roles and skills the tools reason about come from real project practice. Where our insights come from.
Team Developer matches people to tasks with an embedding model that runs inside the service, without an LLM. Skill Coach runs entirely on the employee's machine with a local language model, and only what the employee chooses to share leaves it. Before Skill Coach is used with real employees, it needs a works council agreement, a DPIA and a clear opt-in.
Software companies and tech teams that already use coding agents such as Claude Code, or plan to. The catalog covers Lead Developer, Frontend and Fullstack Developer, DevOps Engineer, Project Manager, Requirements Engineer and Tester.