The useful output is not a list of tools. It is a better process and, when implementation is included, a working system that produces a defined result.
1. Define the business problem
A consultant should begin with the work: what happens, who does it, which information is used, where decisions wait, what exceptions occur, and what the problem costs. This prevents the project from forcing AI into a situation better solved by a simpler process change, integration, or existing software feature.
2. Assess readiness and risk
The consultant reviews process consistency, data availability, system access, ownership, security, legal or industry constraints, team capacity, and the ability to measure an outcome. A responsible answer can be “not yet” or “start smaller.”
3. Prioritize use cases
Potential use cases should be compared by business impact, feasibility, risk, time to evidence, and operating burden. The best first project is usually narrow enough to test but important enough that improvement matters.
4. Design and implement the system
Implementation may include workflow design, prompts or model configuration, retrieval from approved information, software integrations, permissions, human review, exception handling, logging, testing, and a user interface. The consultant should be clear about which parts they deliver and which require another partner.
5. Support adoption and governance
A system is not implemented merely because the technology works. People need to understand when to use it, when not to use it, how to review outputs, what to do when it fails, and who approves changes. Documentation and ownership are part of the product.
6. Measure results
Useful measures depend on the process: time per case, response time, error rate, backlog, cost, throughput, conversion, customer experience, or employee capacity. The consultant should help establish a baseline before the system changes the work.
Questions to ask before hiring
- How will you determine whether AI is the right solution?
- What will be working at the end of the engagement?
- Which systems and information must be accessed?
- How will security, permissions, and human review work?
- How will we test quality and exceptions?
- Who owns the system after launch?
- What result will we measure?
Losung Consult combines operational assessment with focused implementation across workflow automation, systems integration, internal AI knowledge, and AI tool governance.
Start with the work that is stuck
Describe the repeated task, broken handoff, scattered information, or AI investment that needs a clearer answer.
Talk with an AI consultant