HomeServices

Services

We offer three services: AI adoption and workflow design, AI agent development, and AI training and organizational adoption. This page explains what you can ask for, how long it takes, and what you will have when we are done. Our approach to design and quality control is described in “Technology & Approach”.

Service	What we do	For	Typical duration
01 AI adoption and workflow design	Workflow mapping, Theme selection, Pilot and rollout plan	Executives, directors, and implementation leads	Defining an adoption roadmap takes 1–2 months. Monthly advisory for executives starts from 3 months.
02 AI agent development	Internal data and RAG, Tool and API integration, Evaluation and operations design	Implementation leads, IT staff, and the owner of the target workflow	6–12 weeks per workflow. The first 45 days cover a full cycle of prototype and evaluation.
03 AI training and organizational adoption	For executives and managers, For operational staff, For implementation and development staff	Executives, operational staff, and implementation/development leads (role-specific tracks)	Half-day to 3-day workshops plus a practical assignment (4–8 weeks)

For entry points by your role — executive, implementation lead, or training coordinator — see “Common requests” below.

AI Transformation

AI adoption and workflow design

We identify which workflows should incorporate AI, evaluate impact and feasibility, and develop an implementation plan. This is the right starting point when you want to decide where to begin or see projected return on investment first.

Workflow mapping

Through interviews and process mapping, we document workloads, decision criteria, and where data lives.

Theme selection

We select implementation targets based on improvement potential, data readiness, and the risk of errors.

Pilot and rollout plan

We define success criteria and evaluation metrics, then design a plan for moving from proof of concept to live operation.

For: Executives, directors, and implementation leads
Duration: Defining an adoption roadmap takes 1–2 months. Monthly advisory for executives starts from 3 months.
Deliverables: Workflow inventory, prioritized list of implementation themes with evaluation criteria, implementation roadmap

Agent Engineering

AI agent development

We connect your internal knowledge and existing systems to build AI that carries out real work. We agree on one target workflow, then take it from prototype through evaluation, operations design, and handover.

Internal data and RAG

We organize documents, set up retrieval, and surface source references. We design information access with your access controls in place.

Tool and API integration

We build read and write connections to existing systems, including input validation and permission management.

Evaluation and operations design

We integrate quality evaluation, logging, retry logic, exception handling, and human approval into the workflow.

For: Implementation leads, IT staff, and the owner of the target workflow
Duration: 6–12 weeks per workflow. The first 45 days cover a full cycle of prototype and evaluation.
Deliverables: Business AI (from prototype to production), integration with existing systems, evaluation dataset and results, operations manual

Learning & Adoption

AI training and organizational adoption

We cover foundational understanding, role-specific practical exercises, and training for operations staff. Participants use their own real work as the material, and leave with a tool built for their own workflow.

For executives and managers

Learn how to select implementation themes, measure impact, and design an operating model.

For operational staff

Using your own workflows as material, practice prompt design, output review, and integration into daily work.

For implementation and development staff

Learn workflows, RAG, API integration, and evaluation principles to build an in-house foundation.

For: Executives, operational staff, and implementation/development leads (role-specific tracks)
Duration: Half-day to 3-day workshops plus a practical assignment (4–8 weeks)
Deliverables: Tools and operating procedures built by participants for their own workflows; materials for company-wide rollout

CONSULTATION

Common requests

We have three entry points organized by role and goal. For each engagement, the target, duration, process, and deliverables are agreed before we start. Durations are estimates and are adjusted based on the number of target workflows and internal capacity.

Consultation 01

Clarifying management and adoption priorities

You want to decide what to delegate to AI, where to start, and how much to invest.

For: Executives and directors. Companies that do not yet have an internal AI lead.
Duration: Monthly, starting from 3 months.
How it works: A monthly session (60–90 min) to work through investment decisions, prioritization of target workflows, and internal structure for moving forward. Questions between sessions are handled by chat.
Deliverables: AI adoption policy document, quarterly action plan and retrospective
When contacting us: Please include Your current business and the challenge you want AI to solve.

About the AI Advisory Service ↗

Consultation 02

Prototyping and implementing a business AI

You want to pick one workflow, build something that works, and put it into operation.

For: Implementation leads, IT staff, and the owner of the target workflow
Duration: The first 45 days plus stabilization (3–6 months)
How it works: We proceed in order: inventory, decomposition, prototype, evaluation, operations design, and handover. Weekly sessions and working side by side aim to leave your team able to fix things themselves. The first 45 days cover a full prototype-and-evaluation cycle, and the result determines whether to move to production.
Deliverables: Workflow inventory, prototype and evaluation results, operations manual, handover materials
When contacting us: Please include The workflow you want to target and the names of the systems you use (it is fine if the workflow is not yet decided).

How the first 45 days work (article, Japanese) ↗

Consultation 03

Developing internal AI leads

You want your team to be able to build and fix things themselves.

For: Executives, operational staff, and implementation leads (role-specific tracks)
Duration: Half-day to 3-day workshops plus a practical assignment (4–8 weeks)
How it works: Lectures are minimal. Participants map their own workflows, decide what to hand off to AI, and build something that actually works. The program is aligned with Japan's Digital Skill Standard ver. 2.0 (METI/IPA) under the categories "AI implementation and operations" and "AI governance".
Deliverables: Tools and operating procedures built by participants for their own workflows; materials for company-wide rollout
When contacting us: Please include Number of participants and their roles, preferred timing.

Subsidies and grants available for this training ↗

Examples of participants’ work

**Audio archive viewer**Transcribes meeting recordings and displays them alongside the waveform. Includes a silence-skipping playback feature. Built with Whisper and HTML/JavaScript.

**AI agent status dashboard**A single view for monitoring the status (waiting for input, running, stopped) of multiple AI agents.

**Advisory AI for an education service**Answers operational questions from participants using course materials as the source. Holds a record for each client to keep response context consistent. Built with retrieval (RAG) and a chat interface.

Automated accounting system using Google Apps Script and spreadsheets
Automated generation of monthly report presentation slides
In-house chatbot for new employee onboarding

Some screens have been edited to protect personal information.

THEMES

Example themes

Examples of workflows we have addressed or can address through Agent Engineering and business AI prototyping. If your workflow is not listed, please get in touch. The common design principles behind these are described in “Technology & Approach”.

Handling inquiriesDraft responses to email and chat inquiries; a staff member reviews and sends.
Processing forms and documentsExtract fields from invoices and applications; create candidate entries for existing systems.
Meeting minutes and reportsFrom a meeting recording, produce minutes, decisions, and tasks with assignees.
Searching internal documentsSearch policies, procedures, and past case files with access controls intact; return answers with cited sources.
Accounting and back-office routinesInvoice processing, expense reimbursement, payment reconciliation, and contract management. Separate what AI handles from what people verify.
Sales and marketingProposal drafting, customer communication, and ad operations. Build templates and quality standards that the team can run themselves.

Read about the basic architecture of business AI ↗

TECHNOLOGIES

Supported AI models and platforms

OpenAI

Claude

Gemini

Microsoft 365 Copilot

Azure OpenAI Service

Amazon Bedrock

Dify

n8n

LangChain

Hugging Face

Notion AI

The names and logos shown are trademarks of their respective companies. This list shows representative examples of supported tools and does not indicate any partnership or endorsement. The role of each tool and our criteria for selecting them are described in Technology & Approach.

CONTACT

How to start

Please contact us using the inquiry form. Including the items listed under “Common requests” above will allow us to have a concrete discussion from the first meeting. It is fine if the target workflow or budget is not yet decided.

Go to the inquiry form ↗