AI consulting, engineering and delivery

AI systems that survive contact with real work.

We design and build software, AI knowledge systems and agent workflows around how your business actually works. We can deliver the complete solution, strengthen your existing team, or help you make the right technical decisions.

Discuss your projectExplore our services

Three ways to work with us

01

Build

From discovery and architecture to a working system, release and handover.

02

Extend

Hands-on engineering, architecture and delivery support inside your team.

03

Advise

Clear recommendations on AI opportunities, knowledge quality and technical direction.

See what each engagement includes →

What we work on

01Business and product clarity

Define the problem, priorities and measures of success.

02Knowledge and context

Make trusted information available to people and AI systems.

03Agents and workflows

Connect AI to tools and processes within clear operating limits.

04Delivery and infrastructure

Build, evaluate and operate systems your team can maintain.

See our knowledge systems approach →

Our Agentic Development Life Cycle

Agentic delivery. Human direction.

We organise delivery into small, testable tasks. Agents support implementation and review; people own intent, architecture and release decisions. Feedback improves the next iteration.

  1. 01Define intent
  2. 02Shape architecture
  3. 03Slice the work
  4. 04Build and integrate
  5. 05Test and challenge
  6. 06Release and learn

06 returns to 01. Small slices move through the cycle repeatedly.

ArchitectureQualityGovernanceFlow
Explore the ADLC
Founder

Gunnar Kristoffer Wold

Founder, Last Requirement · Working with clients globally

Gunnar Kristoffer WoldMore about Gunnar

Led by Gunnar Kristoffer Wold, who helps organisations turn AI capabilities into useful, dependable systems.

More than 20 years across software engineering, product development, enterprise data management and technology leadership, combined with hands-on work in AI agents, knowledge retrieval and business software.

The last requirement is often the one that reveals what the system really needs.

Labs

Practical experiments for better AI delivery.

Practical experiments in agentic development, knowledge systems and AI evaluation. Explore the methods and tools we are developing.

Explore Labs
  • Clearer task specifications for agents
  • Knowledge retrieval with source traceability and access controls

Bring us the problem before you bring us the solution.

Tell us what is slow, difficult or ready for change. We will help you identify a useful next step.

Discuss your project