TalkPlug and Play Summit
Your enterprise, as it really runs
Heard us speak? This is the page for the talk you saw.
Video to come
Placeholder until OQ-21 is answered
The argumentWhat the talk says, in short
Leaders are asked to change a company they cannot see.
Interviews and workshops show the company as people describe it. The systems show it as it is built.
AIx2 reads the records a company already keeps, read-only. It maps the process, the data supply chain and the AI infrastructure as built, measures the gap to best practice, designs the better way, and checks live that it holds.
Five Step MethodFive steps, from the records to a result that holds
Five steps, in this order.
Collecting Data
Read-only, from ERP, CRM, spreadsheets, e-mail, chat and cameras.
Mapping Processes and Data Flow
The processes, the strategy and the data supply chain, as built.
Gap Analysis
The distance to best practice, and the red flags.
Optimal Design
Processes, data flow and a playbook.
Implementation Live Assurance
A live score against the optimal design.
The case in the talkThe same path, in one manufacturer
Twelve weeks at an industrial manufacturer in Istanbul.
An explosion-protected equipment maker. The whole path, from first connection to a forecast, took twelve weeks; the work then continued as a managed service.
Step 01: Collecting Data
Twelve sources were read, including three years of paper forms.
Step 02: Mapping Processes and Data Flow
The official order process existed in three variants, not one. 26% of engineering changes were agreed outside the change register.
Step 03: Check against the standards
The work was set against ISO 9001, ATEX and IECEx. Major audit findings went from 3 to 0.
Step 04: Design the better way
A release check and a late-order forecast.
Full story, with the sample behind each number, on the case page.
ResultsWhat the client measured, before and after
What the client measured.
75% → 95%
on-time delivery
Measured by the client: May–Oct 2026 vs the 12 months to Dec 2025
15 → 1
working days to assemble audit evidence
Measured by the client: May–Oct 2026 vs the 12 months to Dec 2025
4.8 → 0.9
danger-zone entries per 1,000 hours
Safety against a four-week shadow baseline
3 → 0
major audit findings
Measured by the client: May–Oct 2026 vs the 12 months to Dec 2025
One engine, two buyersWho the same method serves
One engine serves operators and investors.
How we workThe people and the setup behind the method
Engineers and strategy consultants, inside your organization.
Forward-deployed engineers and forward-deployed strategy consultants work together inside your organization. Everything runs on-premises, with open-source models. The platform comes with expert support when needed; the experts widen the data sources and stay through the implementation.
How it comparesFour kinds of provider, by category
What each kind of provider covers.
| Works from | Covers | |
|---|---|---|
| Consulting firms | Interviews, not real data | Analysis and design |
| Large platform vendors | Their own product | A piece each |
| Process-mining newcomers | Real data | Mapping; they stop short of design |
| AIx2 | Real data, kept live | Mapping, analysis and design: all five steps |
Compared by category, not by company.
Go furtherPages for the cases and industries in the talk
Where to read more.
About the numbersWhere the figures on this page come from
Every figure is from the case page.
The numbers here are the ones the client measured, with their window and sample on the case page. The slides shown on stage are not the source of the numbers.