Fluxell
Get in touch

Design value, starting with people.That is what stays in use.

Operational improvement, generative AI adoption, data analytics, AI product development.Listening to your people and the ground truth of the work, we walk with you from problem framing through implementation and operation.

Workflow automation & AI ops

Data activation & analytics

Generative AI training & adoption

AI product & MVP development

AI is never the goal.

Using AI is becoming a question few workplaces can sidestep. But AI rolled out without a purpose, or without a feel for the work itself, ends up shelved.

Fluxell sits close to the people and the work, and carefully sorts out what the AI is actually for. Then we design AI, data, and product as the right means — and stay with you until it earns its place in daily use.

Make AI and data usable on the ground.

From operational improvement and data activation to generative AI training and AI product development. We meet you wherever you are — problem framing, implementation, or rollout — and stay until it is running.

Key service areas

01

Workflow automation & AI ops

Automate manual handoffs, data entry, approvals, notifications, and reporting. We compose n8n / GAS / API / Google Sheets / Slack / LINE so the system actually survives in production.

Example conversations

  • We want to search across our internal documents with AI
  • We want to automate sales and recruiting intake conversations
  • We want to automate research and recurring report generation
02

Data activation & analytics

Take revenue, customer, and operational data and turn it into something a decision can actually be made on. KPI design, dashboards, predictive models, and reporting flows.

Example conversations

  • We want to back gut-feel decisions with data
  • We want predictive models that improve day-to-day operations
  • We want a cleaner KPI dashboard for the executive meeting
03

Generative AI training & adoption

Make ChatGPT and generative AI usable in real work. Training, workshops, prompt design, internal guidelines, and the rollout work that makes any of it stick.

Example conversations

  • We want hands-on AI training tied to actual workflows
  • We need an internal AI usage policy
  • We want AI adoption to be self-sustaining inside the team
04

AI product & MVP development

Turn ideas for new services or internal tools into working prototypes. Requirements, UX, AI feature design, and implementation direction.

Example conversations

  • We want to build an AI product on top of our domain expertise
  • We want to ship a new-business MVP in a short cycle
  • We need someone to scope AI features and direct implementation

Examples of work we have led.

Across industries and functions, from strategy through design, implementation, and operation. A few of the themes we have walked end-to-end with our partners.

Supporting people

  • AI intake chatbot

    Automate intake conversations through a LINE / WhatsApp AI chatbot so advisors can spend their time on the parts that actually need them.

  • Employee engagement survey & analytics

    Surface feedback, route recommended actions, and improve the state of the organization on a continuous loop.

Making information findable

  • Internal document search (in-house RAG)

    A single search surface across policies, meeting notes, and contracts that removes the dependency on whoever happens to remember.

  • Real-estate listing extraction pipeline

    Pull listings from multiple portals, structure them, and load them into the internal database — automatically.

  • Industry signals — weekly auto-report

    Sweep news, social, and public sources around a theme or competitor set; auto-generate a weekly report of only the signals worth acting on.

Helping decisions

  • Predictive models for manufacturing

    Build models from process parameters and historical runs, then sharpen them with daily training.

  • New-business market & competitor research

    Market size, growth, competitors, and entry risk — packaged for the first real go/no-go decision.

  • Partnership / M&A screening

    Pull candidates from enterprise databases with AI and score them across synergy, financials, and tech stack.

Speeding up creative work

  • Generative AI short-form video pipeline

    Compose ChatGPT, Imagen, Veo, and others to auto-generate social short-form video — and slash production cost.

  • Logo and brand variation generation

    Generate logo candidates from a brand’s tone and worldview, then iterate together until the chosen one is right.

  • Landing page & web design variations

    Spin up multiple LP designs across messaging, layout, and tone — to accelerate A/B testing and internal alignment.

Embedding it into the organization

  • Enterprise AI training programs

    Workflow-grounded sessions and hands-on practice that make AI skill stick inside real work.

  • AI usage guidelines & rollout

    Sort out security, copyright, and operational policy, and partner on the rollout that turns the guidelines into habit.

These are examples. Regardless of industry or scale, we propose what fits your situation.

Things we have actually shipped

A few representative examples — ideas and front-line problems turned into systems that are in use, shipped and operated as products that move outcomes.

YORISAI product imagery

AI hiring assistant

Hiring where the human can focus on the human call

YORISAI

An AI hiring assistant for small and mid-size companies. By automating resume screening, supporting interviews, and streamlining candidate communication, recruiters get to spend their time on the part of the call only a human can make.

Learn more

Scope

  • Concept framing
  • Requirements
  • UI design
  • AI feature design
  • Implementation direction

Resume screening time

60%reduced

Candidate response

2 hours30 min

Racing Oracle product imagery

Boat-race prediction AI

A predictive model that ships as daily editorial in a national outlet

Racing Oracle

Designed, built, and operated a system that generates and distributes prediction content daily for a national boat-race publication. Data acquisition, predictive modeling, and content delivery — one continuous pipeline.

Learn more

Scope

  • Data acquisition design
  • Preprocessing pipeline
  • Predictive modeling
  • Evaluation design
  • Inference operations
  • Content delivery design

Outcomes

  • Ongoing coverage on the official outlet
  • Fully automated daily prediction pipeline
  • Reliable content supply at production scale

Build with the people who know the ground.

Read the problem, build something that gets used.

Stay close to the ground, find the real shape of the problem, and connect it to a system whose value lasts. A loop of thinking together, validating together, and growing the thing together.

  1. STEP

    01

    Understand the people

    Interviews and field observation to understand context, background, and pain — properly.

    Concretely

    • Stakeholder interviews
    • Field observation and ops walkthrough
  2. STEP

    02

    Find what is really true

    Structure and analyze what we learned, expose the real problem and its cause. Why is it happening? What is keeping it stuck?

    Concretely

    • Problem structure mapping
    • Cause analysis and insight extraction
  3. STEP

    03

    Design the experience

    Start from the user’s experience and the flow of the work. Design the direction, the idea, and the concrete shape of how AI and data come in.

    Concretely

    • Experience and use-case design
    • Data activation design & prototyping
  4. STEP

    04

    Make it real

    Validate value and viability through a prototype or PoC, and grow it into a product that holds up as a system.

    Concretely

    • Prototype and PoC validation
    • Build and ship
  5. STEP

    05

    Make it stick

    Stay with you through adoption and iteration so the outcome keeps showing up — a steady cycle of operate, learn, improve.

    Concretely

    • Adoption support and user enablement
    • Measurement and continuous improvement

Not a system rollout — a system of work that keeps producing the outcome.

Notes from the field

What we are seeing when AI meets actual workflows, written down honestly.

Articles are on the way.

FUKUOKA NORIMASA

Norimasa Fukuoka

Founder, Fluxell / AI & Data Activation Director

Why this is the work

I believe technology is not the goal — it is a way to widen what a person can do.

Stay close to the work, talk until you can see the real problem, and design the value with data and AI.

That work, stacked over time, changes the trajectory of organizations and the people in them.

Creating the spark that starts that change — that is what I put my energy into.

  1. 2020–2023

    Sansan, Inc.

    Enterprise sales — finding customer problems and shaping proposals. Contributed to data-driven sales strategy.

  2. 2023–2025

    Keyence Corporation

    Sensor solution sales and implementation support. Solving problems using deep field knowledge.

  3. 2025–

    Independent / Fluxell

    Supporting value creation and growth at companies through AI and data activation.

Portrait of founder Norimasa Fukuoka

Connecting people, data, AI, and experience

The disciplines we draw on — front-line operational know-how multiplied by the power of data and AI, in service of value that matters.

01

Customer understanding

Sales / UX research / Interviews / Behavioral observation / Persona design / Problem framing

02

Data activation

Python / BI / KPI design / Visualization / Predictive models / Dashboards / Analytics design

03

AI implementation

ChatGPT / RAG / AI agents / n8n / Claude Code / Prompt design

04

Productization & experience

Product design / UI/UX / Operational flows / Adoption / Effectiveness validation / Continuous improvement

Tell us what you are looking at

Anything from problem framing to a still-forming idea. No need to have it all worked out. We will listen properly and propose the next step that fits.

contact@fluxell.net

We typically reply within 1–2 business days.