Comparison — Approaches to AI Integration
Several routes exist. They differ in what they require and what they produce.
This page sets out the differences between the main approaches to AI adoption in business operations — general software vendors, broad consulting engagements, and the kind of structured specialist work Shinonome does. No approach is presented as definitively correct; they suit different situations.
← Back to HomeWhen a business is considering automation, the most common mistake is not choosing the wrong tool — it is choosing before understanding which route suits the organisation's actual situation. A software vendor sells its product. A large consulting firm tends to package its existing methodology. A specialist works from the specifics of what you have.
Each approach has appropriate contexts. The comparison below tries to name those honestly rather than argue for one at the expense of the others. If a different route suits you better, that is information worth having early.
Side-by-Side
Three routes compared across eight dimensions.
| Dimension | Software Vendor | General Consulting | Shinonome |
|---|---|---|---|
| Starting Point | Their product's capabilities | Packaged methodology | Your processes and existing data |
| Scope | What their platform covers | Organisation-wide or multi-year | Specific, bounded engagements — 3 to 10 weeks |
| Output | Licence and implementation | Reports and recommendations | Ranked tables, working models, or trained staff — with handover documentation |
| Data Used | Generic or sample data | Collected during engagement | Your own historical records, prepared and validated internally |
| Ongoing Dependency | High — subscription and support | Medium — follow-on projects common | Low — handover designed to be complete |
| Staff Preparation | Vendor training on the tool | Change management module | Dedicated programme using your own material, role-specific |
| Honest Limits | Rarely stated upfront | Varies by firm | Stated in writing — including which processes we recommend leaving alone |
| Typical Cost Range | Subscription plus implementation | Higher — broad scope | ¥26,000 – ¥44,000 per engagement |
What Distinguishes This Work
Four elements that shape how we approach engagements.
The process suitability assessment exists specifically to identify what should not be automated — not only what can be. We put that list in writing alongside the recommendations, with reasons. This is not a common feature of vendor-led engagements.
Forecasting tools built on generic benchmarks or vendor sample data produce generic results. We work from your records, validate against periods your team remembers, and produce a model that reflects how your demand actually behaves — not how a comparable sector is assumed to.
Each engagement has a fixed scope and a clear end point. Handover documentation is included as standard, not as an optional extra or a paid extension. After the engagement closes, your internal team can operate what was built without returning to us.
The capability programme is not an add-on or a general awareness session. It uses the company's own material, is structured around specific roles, and includes a two-month period after delivery when questions can be submitted directly. Staff understanding is treated as part of the implementation, not a courtesy.
Realistic Outcomes
What each route tends to produce in practice.
Organisations adopting a new platform often find that the product works as described, but does not map cleanly onto their existing process. Integration takes longer than projected. Staff require ongoing re-training as the platform updates. Results depend heavily on how closely the vendor's model matches the organisation's structure.
Broad engagements produce broad recommendations. The output is often a report identifying opportunities rather than built tools or trained teams. Implementation still falls to the organisation. This route suits organisations that need strategic direction more than specific implementation.
Engagements are smaller in scope and specific in output. A process assessment produces a ranked table and a written explanation of what should be left as it is. A forecasting implementation produces a working model with validation documentation. A capability programme produces prepared staff and a written policy draft. Each can be used independently of the others.
Investment Perspective
How costs and value compare across approaches.
Known price per seat or per period
Ongoing licence cost continues indefinitely
Integration and customisation billed separately
Switching costs accumulate over time
Broad strategic perspective across the organisation
Higher total cost, longer engagement duration
Implementation often a separate subsequent project
Suits complex multi-site or multi-department situations
Fixed price per engagement, stated upfront
No recurring cost after handover
Assessment (¥26,000) can be done before committing to implementation
Narrower scope — not a replacement for broad strategic work
Working Experience
What each engagement looks like from the inside.
| Stage | Vendor / General | Shinonome |
|---|---|---|
| First Contact | Sales presentation and product demo | Direct conversation about your situation and what you are trying to understand |
| During Engagement | Project manager as primary contact, multiple handoffs | Direct access, findings shared as they emerge |
| At Close | Final presentation, proposal for next phase | Handover documentation, question channel if applicable |
| After Close | Support contract or follow-on scoping | Nothing required — tools and documentation are designed to operate independently |
Over Time
How results hold up after the engagement closes.
Tools that require the original implementer to operate, update, or interpret do not produce lasting results — they produce a new dependency. This applies equally to software platforms and to consulting arrangements where the methodology is held by the consultants rather than transferred to the organisation.
Shinonome engagements are designed around transfer. For forecasting work, the handover includes retraining documentation — so when your data accumulates another year, your team can update the model without returning to us. For capability programmes, the policy draft and exercise materials remain with the company. The question channel is time-limited for a reason: we expect it to close because the team no longer needs it.
We measure a successful engagement partly by how quickly the organisation stops needing us. Staff who understand why a model makes a particular forecast, and what to check when it seems wrong, are more durable than staff who have been told to trust an output. That is the gap we try to close.
Common Questions
Things worth clarifying before deciding.
"AI tools automate the process — staff only need to know how to use the software." +
Automated tools handle defined steps, but they produce output that still requires interpretation. When a forecast model returns an unusual result, someone has to decide whether that reflects a real shift in demand or a data quality issue. Staff without context for that decision either override outputs arbitrarily or defer to them uncritically. Both outcomes undermine the value of the tool.
"A process assessment will tell us which software to buy." +
A process suitability assessment tells you which of your processes suit automation and which do not. It does not recommend specific software — that decision depends on factors including budget, IT environment, and vendor support, which fall outside the scope of the work. The assessment informs that decision rather than making it.
"Specialist firms are for large organisations with dedicated IT teams." +
The engagements Shinonome offers are sized for organisations without dedicated AI teams — that is specifically the context they were designed for. Large organisations with internal data science departments tend to build their own models rather than engage external specialists for this kind of work. Smaller organisations benefit more from the scoped, transferable format.
"Three years of data is a long requirement — we may not qualify." +
The three-year requirement for forecasting work is a threshold for producing a model with meaningful validation — not a strict exclusion. We examine the records that exist and report what is achievable given them. In some cases, two years of consistent data is sufficient for a useful model with appropriately stated limits. We report the honest position rather than accepting a project that cannot deliver what is needed.
Summary
When this approach makes sense — and when it may not.
You have a specific process or decision area where automation has been proposed but not evaluated
You have three or more years of consistent operational records suitable for forecasting
Your team will be working with automated tools but has had no structured preparation
You want a bounded engagement with a defined end point rather than an open-ended consulting relationship
You need organisation-wide transformation strategy across multiple departments or sites
Your primary requirement is a specific software platform rather than process or forecasting work
You have less than two years of consistent historical data available for forecasting
You need ongoing managed automation rather than a one-time implementation with handover
Next Step
If the comparison raises a question worth discussing, we are straightforward to reach.
We can tell you plainly whether a Shinonome engagement fits your situation or whether a different route would serve you better. There is no obligation in making contact.