Charter · Scope it
Write the one document that governs everything else: scope, explicit non-goals, the pilot boundary, decision rights, and the metrics that will actually decide whether the programme worked.
Key deliverable: Programme charter
Our method
CONVERGE: eight phases to a data- and AI-driven enterprise
CONVERGE is the Inosense method for taking an industrial organisation from fragmented plant-floor data to a governed platform shared by every site and safely usable by AI agents.
The challenge
Closing the gap between the machine and the boardroom
Most manufacturers can tell you what one machine did yesterday. Almost none can tell you, in one governed answer, what a product actually cost to make once purchasing, labour and scrap are joined to the shop-floor data that produced it, let alone let an AI agent act on that answer safely.
CONVERGE closes that gap in a structured way: every phase produces a concrete deliverable and has a measurable exit criterion, and the same sequence applies whatever the line makes, be it a stamped bracket, a composite spar or a surgical implant.
The eight phases
A clear sequence, from charter to scale
01 Foundation
Organize · Structure it
Stand up the canonical registry of assets, signals and products, and a configurable equipment hierarchy (ISA-95/88): the semantic backbone every later phase depends on, published once and reused everywhere.
Key deliverable: Asset registry and equipment hierarchy
Network · Land it
Land the cloud foundation: landing zone, private connectivity, policy-as-code guardrails and a workspace pattern that makes cost, capacity and access attributable from day one, with data residency built in.
Key deliverable: Landing zone as Infrastructure-as-Code
02 Data
Validate · Productise it
Build the medallion data platform and turn data into owned, contracted products (owner, grain, schema, service objective), including the product-centric model that traces material or units through a process.
Key deliverable: Data contracts and retention policy
Enable · Open it up
Give every persona (steward, engineer, analyst, manager) the access and applications to use the platform without filing a ticket, and join shop-floor data to purchasing, sales and HR.
Key deliverable: Access matrix and master-data alignment
03 Intelligence
Reason · Make it agentic
Design and ship the agentic layer: shop-floor copilots and autonomous recommenders that reason over governed data, with on-behalf-of identity, platform-enforced guardrails and a staged rollout.
Key deliverable: Agent safety charter and production-readiness gates
04 Scale
Govern · Run it
Run the platform as a product: cost attribution by business, plant and layer, a service catalogue, ongoing quality monitoring, and an operating model that survives the original team leaving.
Key deliverable: Service catalogue and chargeback model
Expand · Scale it
Onboard the next plant, region or business by configuration rather than rebuilding, measuring onboarding effort per plant, until the whole enterprise sits inside one governed environment.
Key deliverable: Onboarding playbook and reuse-effort tracker
The target
The Enterprise Metaverse
In CONVERGE terms, an Enterprise Metaverse is neither a headset nor a marketing graphic. It is a living, governed, queryable representation of the whole enterprise: every asset, process, product and plant, joined to the commercial and workforce data that explains why it matters. It answers in real time to anyone from a line operator to the CEO, and increasingly to AI agents acting on their behalf. Three properties set it apart from just a bigger dashboard:
01
Continuous
Sensor to boardroom KPI as one governed lineage
02
Agentic
Reasoned over and acted on by people and AI agents
03
Federated by design
Built once, rolled out to every plant the same way
Our principles
Five rules that hold it together
The Charter governs
Where any other document disagrees with the ratified Charter, the Charter wins, and the disagreement gets fixed rather than argued twice.
Prove, don't promise
Every phase has a pilot-scoped, falsifiable exit criterion. Nothing moves to the next plant on the strength of an architecture diagram alone.
The platform enables, the business builds
The platform team ships governed data and a model-serving path; business teams build their own use cases on top of it.
Configuration, not code
If onboarding a new plant, region or business requires touching code, the building block responsible is fixed before the next onboarding.
Agents read and recommend
They never write to operational technology. This architectural boundary is stated once, not renegotiated under deadline pressure.
About us
Inosense in a nutshell
Founded in 2017 and based near Paris (Val-d'Oise, France), Inosense is a consultancy specialising in data, artificial intelligence and IT architecture.
We work with executive teams, CIOs and business leaders to design and deliver information systems that make the most of data, from the shop floor to the boardroom.
All servicesOur values
01
Rigour
High-quality deliverables and commitments kept.
02
Transparency
Honest recommendations, in your interest only.
03
Proximity
Tailored support, working closely with your teams.
04
Curiosity
Continuous watch on data and AI technologies.
Let's build your data-driven organisation together
A project, a question, an idea? Let's talk about your challenges, with no commitment.