Book a Meeting →
Manufacturing IT/OT Summit Europe · Munich · 21–22 September 2026 · Stand 25 We exhibited at the Manufacturing IT/OT Summit Europe, Munich, September 2026

Most AI ideas don't survive their own business case. We'll tell you which of yours will.

Accountable, Governed & Accelerated AI. Assay, our AI consulting product, models what a use case really costs — the build, the running, the people still checking output, the governance and evidence around it — against what it saves. Six times out of eight, don't build it. The two that pass, we build in half the hands-on effort. Nothing ships ungoverned.

Book a Meeting → Or name one process and we'll say whether it's worth automating.
A number, before anyone writes code Weeks, not months Go or no-go, either way
where to find us both days
25 Stand
Event
Manufacturing IT/OT Summit Europe
Venue
Holiday Inn Munich City CentreHochstraße 3, 81669 Munich
Dates
21–22 September 2026
Also
Monday evening reception — we'll be there
Who's on the standwho you'll meet → VR VenkatCo-founder & CEO PV PrasobhCo-founder & CTO
The argument, in three beats

Each one only lands because the one before it did.

Every other stand at the summit will tell you what they can build. The most useful thing we do is tell you what not to — and then move the line so more of your ideas clear it.

One

Most AI ideas don't survive their own business case.

Assay, our AI consulting product, models what a use case really costs — build, running, the people still checking output, the governance and evidence around it — against what it saves. Six times out of eight, don't build it.

Two

We can move that line.

AI runs through our own build process, so a project takes roughly half the hands-on effort. Halve the cost and borderline ideas clear the bar. Speed only means something once there is a number for it to move.

Three

Nothing ships ungoverned — which is why the number in beat one is honest.

Governance cost is a line in the business case, so anyone quoting a return without it is guessing at the number that usually decides the answer.

Accountable is the number. Accelerated is what moves it. Governed is why you can trust it.

How we work

Accountable, Governed & Accelerated AI

Three stages, and the interesting part is that it goes round again. What a system actually does once it is live is what decides whether the next one gets funded — or stopped.

Decide → Build → Run, inside one governed band 6 of 8 opportunities stop at Don't build
DecideAssay-led accountability
  • Models cost, governance overhead and payback per opportunity
  • Risk tier caps how much autonomy each step is allowed
  • Output is go or no-go, with the number attached
BuildAccelerated by AI SDLC
  • AI applied across the lifecycle — spec, code, test, deploy
  • Eval gates and guardrails before anything goes live
  • Roughly half the hands-on effort
RunContinuous evaluation by Assay
  • Real volumes, real accuracy, real review overhead
  • Modelled against actual, reported to the sponsor
  • That result funds the next build — or stops it
Don't build6 of 8 opportunities stop here

Most AI ideas don't survive their own business case

Governed — the band around everything Decide Build Run Don't build
Decide with Assay

The line most business cases leave out is usually the one that decides it.

Pick a process shape. The bars show roughly where the cost lands over the years, and whether the governance-and-review line swamps the saving. The numbers are illustrative — the shape of the answer is not.

Assay · our AI consulting product

Assay is the assessment that puts a number on an AI idea before anyone writes code — and says go or don't, with the reason attached.

What it does

Takes one process apart and models what automating it would really cost over a realistic period — the build, the running, the people still checking output, the governance and evidence around a live system — against what it saves.

What it sets

A risk level for each step, which caps how much of the process may run without a person. That constraint changes the cost, and the cost changes the answer — which is why the number is honest.

What you get

A number and a decision. Go, with the business case attached. Or don't, with the reason. Six times out of eight it is the second one — and that is the most useful thing we will tell you.

Fixed scopeWeeks, not monthsEnds in a decision either wayRe-run after go-live
Pick a process shapefour common ones

What moves the governance line: the risk tier of each step, and how much of it may run without a person. A safety-relevant decision keeps a human on every output — which is the cost that most cases forget to count.

Drawings retyped into a system Assay · multi-year view · one opportunity
Illustrative
Buildone-off, AI-accelerated0
Runninglicences, hosting, model calls0
Governance & reviewthe people still checking output, evidence, monitoring0
Savinghours and error cost avoided0
Go — pays back in year oneHigh volume, low judgement, and the errors are visible when they happen. Review can be sampled, so the governance line stays small.
Net over the period: +€0 risk tier Low · autonomy high

Not the build. Not the licences. The evidence, the monitoring and the human checks that have to sit around a live system. It is the line most business cases omit, and it is usually what turns a confident yes into a no. We put it in — which is what makes the yes worth something.

Build with AI SDLC

Roughly half the hands-on time. Your engineers hold every gate.

Requirements, design, development, testing, code review, DevSecOps — AI does the lifting at every stage of the build, and a named engineer signs each one. Speed is not a separate selling point. It is what halves the build line in the Assay case, so the borderline ideas above clear the bar.

One release, all the way round Build  ·  Prove  ·  repeat
  • Requirementsthe process, as it really runs
  • Designoptions and the risk tier
  • Developmentin your stack, your patterns
  • Human gatean engineer signs, both ways round
  • Testinggenerates and runs the suite
  • Code reviewfirst pass on every change
  • DevSecOpsscans, SBOM, your CI/CD
AI does the lifting a person decides evidence captured as it happens
01

AI at every stage, not just the editor

Autocomplete is one stage of six. The lift is in the requirements nobody wants to write, the tests nobody wants to maintain, the review queue and the pipeline work underneath.

02

Every gate is a person

The AI proposes; a named engineer approves. Merges, releases and infrastructure changes never auto-pass. A plant system is not the place to find out otherwise.

03

It ends in a signed release

SBOM and provenance attached, deployed through your own CI/CD into your own environment. The evidence is a by-product of shipping, not a project of its own.

This is what moves the line in Decide: halve the build cost and the borderline case in the assessment above clears the bar.

Run and evaluate with Assay

After go-live, the same assessment — against real numbers.

Real volumes, real accuracy, real review overhead. Once the system is live we re-run the Assay model against what it actually did and report the result to the sponsor. That result funds the next build, or stops it.

Modelled in Decide · measured in Run the numbers are illustrative · the shape is not
Go-livereleased and monitored
Weeks 1–12real volumes accrue · outputs sampled
Re-run Assaysame model · actual inputs
Sponsor reportfund · fix · stop
Two ways it can gopick one

What this means today: we re-run the assessment against your numbers at agreed points after go-live, and report it. An always-on, automated version of the same check is on our roadmap — we don't sell it as live.

Drawings retyped — twelve weeks after go-live Assay · re-run at week 12 · modelled vs actual
Illustrative
MeasureModelled · DecideActual · RunΔ
Funds the next buildVolumes came in above the model and review overhead below it. The saving is real, and the change-request case moves up the list.
Reported to: the sponsorNext decision: fund the next build
01

Real volumes

Not the volume in the business case. The number of drawings, requests or reports the system actually handled — and what happened to the ones it couldn't.

02

Real accuracy

Sampled and checked by a person against the launch baseline, so a quiet drift shows up as a number on a report rather than a complaint from the line.

03

Real review overhead

The hours someone still spends checking output. It is the line that decided the case in Decide, so it is the line we watch hardest in Run.

Governed — the band around everything

Nine phases, eight gates — from intake to the day it is retired.

Whether it is a planning application, a drawing pipeline or an assistant over plant data, the trust layer is the same. Ordinary controls don't cover a model that drifts, an output nobody re-checks, or a prompt turned against you — so every AI system we build runs a nine-phase lifecycle with eight documented go/no-go gates, structured on the NIST AI RMF functions and mapped to the EU AI Act and ISO/IEC 42001. Governance here is a design constraint, not a countdown.

Map — what this system is, and what it touches Phases 01 — 03

Classify the system against the EU AI Act and NIST AI RMF, name what it touches on the line and who owns it, inventory the data with recorded provenance, then set the risk tier that decides how hard the rest of the lifecycle has to work — and how much may run without a person.

01Initiate & scope02Data & knowledge03Design & architecture
Every phase has a named owner on a RACI — one person Accountable, not a committee. Eight go/no-go gates sit between the phases, each a short, documented decision. Gates are cheap; drift is expensive. And the cost of all of it is a line in the Assay case, which is what makes that number honest.
Four ways AI systems go wrong in a plant — and what catches each
The unbounded change A system allowed to alter more than the task needs — a line configuration, a schedule, a threshold — with nobody able to undo it. Caught by a permission matrix and a hard human-approval step on anything irreversible. Set at Phase 03 · Design & architecture
Prompt injection & leakage Instructions smuggled in through a supplier document or a drawing the model reads. Caught by adversarial and red-team evaluation against the OWASP LLM Top 10, before launch and on every material change. Tested at Phase 05 · Evaluate & red-team
Silent quality drift A system that passed at launch and quietly stopped being right — a new part family, a changed drawing template, a new supplier. Caught by sampled online evaluation against the launch baseline, with a human-feedback loop. Watched at Phase 08 · Monitor & observe
Unaccounted data Nobody can say where a piece of grounding or training data came from — which drawing revision, which plant's numbers. Caught by a data inventory with recorded provenance and a documented basis for use. Established at Phase 02 · Data & knowledge
The governance framework — nine phases, eight gates, in full →
Assurance stack Continuously verified
The companyISO/IEC 27001:2022 · ISMS
Your dataEncrypted · minimised · deleted
Your codeSecure SDLC · gated releases
Your AIEU AI Act · NIST AI RMF · ISO/IEC 42001
13+ years200+ clients0 unanswered questionnaires
ISO/IEC 27001:2022 Information Security Management System Certified by TÜV Rheinland, an accredited certification body.
Certificate holder
Focaloid Technologies Private Limited
Standard
ISO/IEC 27001:2022 — the current revision
Certificate no.
9000041009
Issued by
TÜV Rheinland
Public register
Certipedia — TÜV Rheinland's open certificate database

Verbatim from the certificate scope “Information Security Management System covering Software Development Services encompassing AI, Data & Analytics, Cloud & DevOps and Digital Engineering including supporting functions such as Finance, IT, Administration and Human Resources.”

Verify the certificate on Certipedia ↗ Opens TÜV Rheinland's register in a new tab · no login, no gate
Vendor security review Cleared
Is the supplier ISO 27001 certified?
Where is our data stored, and for how long?
Who has access, and how is it revoked?
How is code reviewed before release?
What stops the AI taking an irreversible action?
How was it tested before it touched the line?
CAIQ · SIG-lite · your own spreadsheet Typical turnaround ≤ 2 business days
In most engagements we work inside your cloud tenancy — your account, your region, your keys. Your data never leaves an environment you control, and offboarding is a matter of revoking access rather than trusting a deletion promise.
Where this usually starts

Bring the process, not the use case.

You don't need an AI idea. You need the thing that costs you every day. These are the shapes we see most often on a plant's list.

The drawings your engineers still retype

Specifications sitting in scanned files that somebody keys into a system by hand, every day.

Usually a go

The process everyone agrees is broken

The one that has been on a list for three years because nobody could justify the fix — or prove it wasn't worth it.

Often the surprise

Change requests that take a month to assess

Where the delay is finding out what a change touches, not making it.

Usually borderline

A large OutSystems estate nobody has scoped

Where the real question is what genuinely has to be rebuilt and what can simply be lifted.

Starts with a map

Something else? That is fine. Name it and we will tell you honestly whether it is worth automating — including if it isn't. Name one process →

Proof

Thirteen years inside real plants — applications, data and connected products, still running.

Five engagements, described rather than named. Client names are shared under NDA when we meet.

Automotive · Fortune 500 Tier 1 safety supplier · Europe / Global

Six years and still running, across plants

For a Fortune 500 Tier 1 automotive safety supplier, the applications and data its European and global plants run on day to day:

  • Test-order planning across plants, scheduling around equipment capacity and site calendars.
  • Production-line configuration for process engineers, replacing the spreadsheets and email it lived in.
  • Supplier risk tracked across four operating divisions.
  • The daily reporting the plants run on.
In production · year six
Engineering & inspection · Enterprise · Europe

Safe operating limits for chemical and refinery plants

For a European engineering and inspection enterprise: sensor thresholds tracked against equipment wear, so the limit a plant runs to reflects the state the equipment is actually in.

7+ yrsworking relationship, and counting In production
Automotive components · Global group

One master data layer, and drawings that are no longer retyped

For a global automotive components group: master data management across plants and product lines, and the engineering drawings that used to be keyed in by hand now digitised, structured and searchable.

5+ yrsworking relationship, and counting In production
Automotive OEM · Connected car · IoT

A car that sets the house before it arrives

For a global automaker: a connected-car-to-smart-home bridge. The vehicle's position against a home boundary drives presets for lights, HVAC and the garage door — through a home control unit and a cloud layer — automatically or with the owner's confirmation. Two phones per home, presence resolved between them, a manual override that always wins.

Shipped · app, control unit, cloud
Industrial automation & robotics · Service platform

Service requests for CNC, robots and lasers, from the floor

For a global industrial automation maker's customers: register machines by model and serial, raise service, spare-part and repair requests against the installed base, confirm out-of-warranty charges by OTP, and track every request to closure — with an admin panel that runs the back office.

In production · end users, makers, dealers
The shape of the work above inside plants and around their products, for years
Plant applicationsPlanning, line configuration, supplier risk and the daily reporting — the systems a plant opens every morning.
Engineering & master dataDrawings digitised, master data governed, safe limits tracked — and the governed pipelines underneath all of it.
Connected productsA car that talks to a house, a service platform for an installed base — the software around the machine, not just inside the plant.

Applications and data first, AI where the number says so. Every AI idea on top of this work goes through Assay before it is built.

Databricks partner ISO/IEC 27001:2022 certified Thirteen years building software
Who you will meet

Two of us, at stand 25 on both days.

Not a sales team. The people who would do the assessment and run the build — so the conversation on the stand is the conversation you'd have on the project.

Venkat Ramakrishnan
Co-founder & CEO

Venkat Ramakrishnan

Talk to Venkat about whether an idea already on your list is worth funding, and what the honest number behind it looks like.

Ask about Business caseThe honest numberGo or no-go
LinkedIn →
Prasobh Veluthakkal
Co-founder & CTO

Prasobh Veluthakkal

Talk to Prasobh about how it actually gets built and governed: the architecture, the AI-accelerated build, and what is allowed to run without a person in the loop.

Ask about ArchitectureAI-accelerated buildAutonomy limits
LinkedIn →

If you would rather not book anything, come by stand 25. We are on the floor both days, and at the evening reception on Monday.

The ask

Name one process. We'll say whether it's worth automating.

Tell us the one thing in your operation that everyone agrees is broken and nobody has fixed. We come back with a straight view on whether automating it is worth the money — including if it isn't. It is the first question Assay asks, and that answer is the best signal either of us will get from twenty minutes.

Prefer to talk?

Book 20 minutes in Munich with either of us. Bring the process; we'll bring the questions. Slots on both days, and at the Monday evening reception.

Book a Meeting → European hours · no deck · a working conversation
Practical
Event
Manufacturing IT/OT Summit Europe
Stand
25
Venue
Holiday Inn Munich City Centre
Dates
21–22 September 2026

We are on the floor both days, and at the Monday evening reception. If you would rather not book anything, just come by.

The process you would want us to look at

Four fields. One line on the process is enough — we will come back within two working days with an initial view and a question or two.

No phone number, no job-title dropdown. We read the process line ourselves.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Thanks. We have it.

Someone will come back to you within two working days with an initial view, and a question or two if we need them. If you are at the summit in the meantime, come and find us at stand 25.