Framing
To make a decision before committing a team or significant budget.
Current-state analysis, interviews, constraints, options, trade-offs and recommendation.
Deliverable: framing material that lets you decide and start the next step.
arkalabs · Advisory
Product Management & AI systems
I work with product, tech and business teams to design a product, frame its evolution or build an AI system integrated into their operations.
Two areas of work: Product Management and AI system design. They can be used separately or together when a product includes agents, models or automation.
01
Product Management
I work as a Product Manager on an existing product, a new offer or a transformation phase.
I take care of the work needed to move from a problem to a framed, prioritised and buildable product.
- 01
Strategy & framing
Clarifying the problem, users and objectives. Analysing business and technical constraints. Defining scope and key trade-offs.
- 02
Discovery
Interviews, usage analysis, solution exploration and hypothesis validation before committing delivery effort.
- 03
Prioritisation & roadmap
Structuring opportunities, arbitrating between value, cost and risk, building the roadmap and preparing product decisions.
- 04
Delivery
Driving implementation with design, engineering and business teams. Slicing, specification, trade-offs and follow-up until production.
- 05
Product evolution
Analysing usage and outcomes, identifying friction points and preparing the next iterations.
Outcome
A clear product direction, explicit decisions and a team able to move on a controlled scope.
02
AI systems
I support companies that want to integrate AI into a product or their operations without reducing the topic to choosing a model or building a prototype.
The work covers the whole system: models, agents, data, tools, memory, orchestration, controls and IS integration.
- 01
Framing
Identifying use cases where AI creates measurable value. Analysing existing processes, available data, constraints and acceptable autonomy levels.
- 02
Architecture
Designing the system: agent responsibilities, models, accessible tools, memory, component exchanges and user interaction.
- 03
Integration
Connecting applications, APIs, knowledge bases and existing processes. Defining rights, interfaces and control mechanisms.
- 04
Evaluation & reliability
Defining evaluation criteria, testing behaviours, managing errors and edge cases, observability and human supervision.
- 05
Production
Moving from prototype to usable system: target architecture, security, costs, operations and evolution.
Outcome
A documented architecture and an AI system whose behaviour, limits and operating conditions are explicit.
01 + 02
When Product and AI meet
Some missions require both disciplines.
This is the case when creating a product based on AI agents, introducing an important AI feature into an existing product or turning a business process into semi-autonomous software.
In these situations, I work on both:
the product
users, value proposition, journeys, priorities, roadmap;
the system
models, agents, data, tools, architecture and controls.
This avoids separating a product decision from its technical consequences.
Contracting
Engagement formats
The format depends on the problem to solve. I favour missions with a clear scope, expected outcome and responsibility level.
Targeted mission
To solve a precise problem: discovery, roadmap redesign, new product definition, AI system framing or architecture design.
I take responsibility for the topic until the agreed result is reached.
Product Management
To reinforce an existing team or take product responsibility for a scope.
I work directly with business, design and engineering teams on discovery, trade-offs, roadmap and delivery.
AI architecture & support
For a team building or evolving an AI system and needing someone responsible for the overall design.
Architecture, structural choices, technical reviews, evaluation and production support.
Deliverables
What you get
I do not deliver a pile of recommendations.
Depending on the mission, the work produces elements directly usable by teams: product decisions, scope, roadmap, specifications, architecture, prototypes, evaluation criteria or implementation plan.
Deliverables are there to build, decide or arbitrate.
About
Jeremy
Product Manager & AI systems architect
I work at the interface between product, business and technology.
In Product Management, I work from discovery to delivery, with specific attention to framing and trade-offs.
On AI systems, I work on their overall design: agents, models, data, tools, memory, orchestration, evaluation and integration with existing systems.
These two expertises are independent. I can join a team for a Product Management mission without AI, design an AI system without product responsibility, or take care of a topic requiring both.
Contact
Do you have a topic to frame?
A first conversation usually makes it possible to identify the problem, mission scope and suitable engagement format.