Who we are
Production-grade custom AI AutomateX Bootcamp PACE OS & Scribtive Free AI Readiness Assessment
Insights Careers Book a call
Accelerating AI adoption for UK rail

Let engineers engineer.

Free your team from the X that drains them. Truly AI-enabled, finally.

Start here
RISQS Verified ICO Registered
Built by rail insiders who actually understand the problems
Form G PACE · NR/L2/RSE/02009
Drafting
Project
Station canopy refurbishment · Southern region
Cited · Project brief
Deliverable
Engineering assurance form (Form G), drafted section by section
Cited · PACE deliverable list
Compliance statement
Design complies with the approved requirements and applicable standards.
Cited · NR/L2/RSE/02009
Nothing ships without your engineer’s review. Human in the loop, by design.

Illustrative. Real output uses your own data, cited and reviewed.

Find your X
Find your X

We automate your X.

Your X is the error-prone, repetitive, time-consuming friction in your workflow, often unique to your operational context and organisation type. A few of the ones we see most often:

A few examples across 5 workflow families Assurance Design Tender Knowledge Controls See how we’d automate yours →
Who we help

For ambitious rail teams and leaders.

The pain → the prize

8 layers between your engineers and engineering.

FORMS TENDER RETURNS TENDER SCRAMBLE STANDARDS SEARCH CAT 3 COMMENTS REPORTING DRAFTING MEETINGS & COORDINATION engineering Let engineers engineer. Apply judgements Make decisions Manage risks

Hover a layer to read it. Click to automate it away. Or just scroll.

Your team, truly AI-enabled.Same team. The system runs the admin. Your engineers review, decide, and engineer.

See it in action
Forms · drafted & cited Tender returns · assembled Evidence · found & cited Standards · clause-level answers Comment responses · drafted Reporting · live, not rekeyed Drafting · grounded first drafts Meetings · actions chased
  • Forms · A/B/C/G. Senior hours lost to formatting, citation chasing and version drift. → Drafted section by section, cited clause by clause, routed for review.
  • Tender returns. Every submission reinvents the wheel. → Assembled from your win library and evidence base.
  • Tender scramble. The 11th-hour hunt for proof you know exists. → Evidence found, matched and cited in seconds.
  • Standards search. 12,000+ standards across the railway. → Clause-level answers across NR, RSSB and CSM-RA.
  • Cat 3 comments. Opinion-driven comment cycles and signature chasing. → Responses drafted against the clause, closure tracked.
  • Reporting. Half a week assembling status reports. → Live dashboards, nothing rekeyed.
  • Drafting. Every deliverable starts at a blank page. → Grounded first drafts in minutes, from your templates.
  • Meetings & coordination. Actions that die in inboxes. → Captured, assigned and chased by the system.
Common questions

Before you book a call.

The questions rail teams ask us most. If yours is not here, drop it through the chat or book a 30-minute consultation off the back of the AI Readiness Assessment.

Where should we start?

Two paths depending on where you are. If you already know your friction points ("X"), take the AI Readiness Assessment and book a 30-minute consultation off the back of it. We’ll walk through your results together and map a clear first move.

If you’re earlier in your thinking, the AutomateX Bootcamp is the structured route in. It walks your team through finding your friction points ("X"), builds the business case, and ends with a live AI-native prototype on your own data.

Are you selling a product or a service?

AutomateX is a custom-build service business, not a SaaS product. We diagnose your friction points ("X"), then build the AI-enabled solution that fits your operational context, your data, and your assurance environment. Each engagement is bespoke. Our strength is at the intersection of rail domain expertise and AI implementation, applied to your specific friction.

For teams that want a packaged AI tool rather than a custom build, we’re developing Scribtive™ and PACE OS™ as adjacent product offerings. Those are on the waitlist. The core AutomateX engagement remains custom build.

How long does a typical engagement take?

About 1 week to Diagnose. Up to 2 weeks for Alpha build and validation against retrospective data. Around 4 weeks for Beta build and independent blind evaluation. Then 3 to 6 months from Deploy to a live, fully assured workflow.

Complexity moves these. A tightly-scoped Form A automation lands sooner. A multi-discipline Operate & Evolve programme takes longer.

How is pricing determined?

Pricing depends on what you need. The structured entry point is the AutomateX Bootcamp, credited in full against your build if you take the work forward with us.

Not every team needs the Bootcamp. Many already know their friction points ("X") and want to go straight to the Build phase.

For build engagements, pricing is outcome-based, not input-based. Every build is different, sized against the friction we’re removing and the value of the solution. We quote case by case. Talk to us directly for a quote on your specific situation.

What does AutomateX do with our project data?

Your data stays in the UK. Uploaded documents are stored in a relational database and cloud object storage with our UK-based infrastructure provider. Document content is anonymised at ingest. Un-anonymised data is only ever visible from the user account that uploaded it.

AutomateX is ICO-registered and operates under ISO/IEC 42001-aligned controls.

ISO/IEC 42001 alignment, what does that mean in practice?

ISO/IEC 42001 is the new international standard for AI management systems, the AI-governance equivalent of ISO 27001 for information security. For AutomateX it means three things in practice:

1. Documented framing. Every AI-enabled workflow we deploy has a framing artefact covering scope, hazards, success criteria, and acceptance thresholds, agreed by your sponsor before build starts.

2. Audit-traced output. The model version, the original wording, and any accept/override decision are all logged per artefact.

3. Drift evaluation. We re-run evaluations on a regular cadence to catch model drift before it surfaces in production.

Rail-grade assurance discipline applied to AI delivery.

Why rail-only?

Domain knowledge is where the value sits. AI tooling is increasingly commoditised. What’s defensible is the rail expertise that knows what good looks like, where the friction actually lives, and how the assurance process really works.

We don’t want to wait for AI generalists to figure out rail. We’re rail insiders building at the implementation layer ourselves, so solutions land in workflows that engineers actually use.

What if my firm is too small or too big?

Every rail firm has friction points ("X"). Larger firms typically get most value from productivity and engineering-leverage gains, freeing senior judgement from admin. Smaller firms get most value from removing single-point-of-failure bottlenecks and reducing rework.

Firm size is not the constraint. The real question is whether you are ready to use AI responsibly and have the appetite for change. If you do, we can help.