Sponsor intake
A 60-minute session with the CEO or sponsor to define the objective, boundaries and business context, and agree the business area and candidate workflows to explore.
Claes Parflo Advisory
Find where AI creates enough value to change the work.
I work with CEOs and leadership teams to identify where AI can materially improve decisions and workflows, where it cannot, and what should actually be built.
My background is in running businesses, turnarounds, international sales and operations. Today I also work hands-on with AI systems and agents, building real products and operational workflows rather than stopping at recommendations.
The distinction
Business judgment first. AI capability second. Both matter.
AI makes it increasingly easy to demonstrate something impressive. That does not mean it is useful, reliable or worth implementing.
The difficult questions are business questions:
- What decision are we trying to improve?
- What information does it actually depend on?
- Where is human judgment still essential?
- What happens when the model is wrong?
- What would have to change in the surrounding workflow?
- And is the resulting improvement worth the effort?
I spent much of my career answering questions like these as an executive, with responsibility for businesses, customers, organisations and results across multiple international markets.
The difference today is that I can also test the answer. When an AI-enabled workflow appears promising, I can move beyond a slide describing it and build enough of the system to discover whether the idea survives contact with reality.
Fixed-scope engagement
AI Decision & Workflow Sprint
From “we should do something with AI” to one tested workflow and a clear decision.
€6,500 fixed + VAT where applicable
Approximately 2 weeks
A two-week remote sprint combining structured customer workshops, independent analysis and practical testing of one priority workflow, with a working proof of concept where appropriate. We start with one agreed business area, understand the work with the people who do it, and jointly select the workflow worth testing. The result is a decision grounded in your operation.
Workflow workshop
A two-hour workshop with the people who do the work to map decisions, information sources, pain points, exceptions and workarounds. Understand how the work actually happens and where human judgment matters.
Opportunity mapping
Independent analysis of the workshop findings and agreed materials to identify approximately 5 to 10 concrete AI opportunities within the agreed business area, assessing business value, feasibility, information quality, effort and operating risk.
Prioritisation workshop
A 90-minute workshop to review the opportunities and jointly select one workflow for deeper testing. Define how the future workflow should operate between people, information, software and AI.
Build enough to learn
Where technically feasible and appropriate, create a working prototype or proof of concept using representative or client-approved information. The purpose is not production software. It is to expose assumptions, limitations, failure modes and real implementation requirements before a larger commitment is made.
Prototype checkpoint
A 45–60-minute session to review what testing has revealed, validate assumptions with workflow participants and clarify data gaps. Where a prototype is appropriate, use it to make the findings concrete.
Executive decision review
A 90-minute review of the findings and any prototype, with a concise executive assessment and a practical 90-day plan covering next steps, ownership, dependencies, controls and success criteria. Make the decision explicit using four categories:
DO NOWTESTLATERDON'T DO
Customer participation: five structured sessions, approximately seven hours of live collaboration. CEO or executive sponsor, with up to five relevant leadership or workflow participants; attendance varies by role.
Plan for around four hours for the sponsor and up to six hours for core workflow participants, depending on attendance, plus preparation. An internal coordinator should allow a further two to three hours to gather the agreed representative materials.
The outcome
At the end of the sprint
The objective is not a larger consulting programme. It is decision clarity.
The leadership team receives
- a prioritized AI opportunity map
- one redesigned workflow
- a working proof of concept where appropriate
- documented limitations and failure modes
- an executive decision brief
- a practical 90-day implementation plan
- a final management review session
The result should make it possible to decide what deserves investment and, equally importantly, what does not.
Delivery and scope
Engagements are delivered remotely. The stated fee covers the Sprint and deliverables described above. On-site work, travel, production implementation, live-system integration, substantial data preparation, third-party software/API costs, and work outside the agreed scope are not included. Any additional work is separately agreed and priced before it begins.
Each engagement starts with a short written scope confirming the business area, candidate workflows, participants and a bounded set of representative inputs, including their format and readiness. One workflow is selected together for deeper testing. The two-week timetable assumes the agreed materials and required participants are available at the start.
Why Claes
Operator first.
Before working with AI systems, I spent decades running and growing businesses.
My career included senior international roles at Ericsson and other technology companies, covering general management, sales, services, turnarounds and new-business development across Europe, Asia, Africa and the Middle East.
I have managed organisations of several hundred people, held responsibility for businesses measured in hundreds of millions of dollars, led major international customer relationships, launched new service businesses and worked through both growth and turnaround situations.
I later returned to building.
At Klarhet Studio I use AI and agentic development workflows hands-on to design, develop and operate focused software products. LocalBrief turns fragmented local-language information into structured situational briefings. Trendwell turns continuous health data into consistent longitudinal signal.
Different domains, same underlying discipline: start with the source, impose structure, define the rules and boundaries, and make the result useful for a decision.
Ongoing advisory
Executive AI Partner
€1,800/month + VAT where applicable
Monthly engagement. Cancel monthly.
For leadership teams that want experienced outside judgment without creating another full-time role or consulting programme.
two 60-minute executive sessions per month
review of proposed AI initiatives
challenge of vendor claims and internal business cases
workflow and decision-system design
review of important AI outputs or approaches
up to one small exploratory prototype per month, where appropriate, for validation rather than production implementation
assessment of relevant new AI capabilities
The scope is deliberately bounded. This is an advisory relationship, not outsourced implementation or unlimited support.
Remote advisory. The monthly engagement includes the two executive sessions and bounded review or exploratory work described above. It does not include production implementation, operational ownership, unlimited asynchronous support, travel or on-site work. Additional work is separately scoped and priced.
Discuss fitBusiness pricing. VAT treatment depends on customer location and VAT status.
Fit
A good fit when
- Management knows AI matters but has not identified where it creates enough value to justify changing the operation.
- There are recurring management or knowledge workflows involving fragmented information, manual synthesis or repetitive analysis.
- The company wants to test an idea before committing to a large implementation.
- Leadership wants an independent view that is not tied to selling a particular software platform.
- The problem crosses the boundary between business process, management judgment and technology.
Boundaries
Not the right fit when
- The requirement is simply AI training for a broad workforce.
- The company has already selected a system and only needs implementation resources.
- The requirement is commodity software development.
- The objective is to deploy AI everywhere rather than determine where it is useful.
- The organisation wants confident answers without examining uncertainty, limitations or failure modes.
Start here
Start with one real workflow.
If there is a decision, management process or recurring piece of knowledge work where AI appears promising, describe it briefly.
I can usually determine quickly whether it is suitable for the Sprint.
Contact Claes