In production at XL-BYG
From 30 minutes a product to 2
XL-BYG, a Danish builders' merchant, runs Speedtrain directly on its Perfion PIM. Product onboarding that used to take half an hour now takes two minutes – and almost everything passes the quality checks on its own.
XL-BYG, in production. Results reflect their own workflow and data quality, not a guarantee.
95% done from the very first run
Even before the rules are tuned to your data, a first run handles 100 products in under ten minutes – and 95% need nothing from your team. From there, the share grows as you refine the rules. XL-BYG is at 99%.
- 60%Approved immediately
- 35%Pass a quality check, no action
- 4%Quick retry
- 1%Human guidance
Sound familiar?
Manual doesn't scale. Generic AI breaks your brand.
We see it in retail catalogues all the time, before teams bring in Speedtrain. The product is in the ERP, the price is set, the campaign is booked – and the content isn't ready.
Your supplier files arrive half-finished
Abbreviations, ALL CAPS, and empty fields where your market needs attributes. Someone on your team fixes every record by hand.
Every new market multiplies your work
Copy written in one language gets re-keyed into six others. When a spec changes, someone corrects all six.
Generic AI writes fast – and wrong
A general-purpose chat tool doesn't know your termbase, your legal phrasing, or what you're not allowed to claim. So everything gets checked twice, or not at all.
Either way, launches wait for content, and content waits for people. That's not a content problem. It's an operating-model problem – and it's the one Speedtrain is built to solve.
Product enrichment
From a messy feed to a product page you'd publish
One supplier record, before and after Speedtrain. The same product facts – written the way your brand writes, in every field your channels need.
- Upper: Breathable knit
- Midsole: Foam
- Outsole: Rubber
- Heel: Reinforced
- Size: EU 45 (UK 10.5)
- Colour: Black
How it works
Configured once. Run for every product.
You decide how your products should be described. Speedtrain does the writing and the checking – and your team only steps in where a human decision actually matters.
- Set
You set the rules
Your brandbook, termbase, legal phrasing, forbidden words, and required fields – per attribute and per market. Loaded as hard constraints, not good intentions.
- Generate
Speedtrain writes and checks
Every field is generated from your source data, then runs through quality checks before anyone sees it.
- Approve
Your team handles the exceptions
Output that passes is approved automatically. The rest lands with a specialist, who validates instead of writing from scratch.
- Publish
Live in every channel
Approved content goes back into your PIM or CMS, and on to your webshop, marketplaces, and AI agents via API.
More than enrichment
Enrichment is where Speedtrain starts – not where it stops
Plenty of tools can write a product description. The hard part comes after: combining it with price, stock, and everything else, and getting it live in every channel. Speedtrain runs on Enterspeed Core, so it keeps going where enrichment tools hand over.
Write and check the content
Speedtrain enriches every field against your rules and quality checks.
Speedtrain keeps going, on Enterspeed Core
Join it with the rest of your data
Approved content is combined with price, stock, and content from your PIM, ERP, CMS, and other systems.
Live in every channel
Your webshop, apps, search, and marketplaces read it via Delivery API. Sub-second, via a global CDN.
Ready for AI agents
AI agents read the same approved data via MCP server – with the right price for each customer, and without opening your core systems.
One source of truth
Approved once, live everywhere. No sync jobs between your enrichment tool, your PIM, and your webshop.
Your mappings stay yours
You own how your data is mapped, on your own or with us. When your systems change, the setup follows along – it isn't locked inside a vendor's platform.
Room to grow
Start with product content. Add sources, markets, channels, and AI agents to the same data layer when you're ready – without starting over.
With humans in control
AI does the work. Your people decide what goes live.
Speedtrain isn't a black box that publishes on its own. The rules are yours, every output is checked, and every decision is recorded.
- Rules
- Forbidden words
- Legal phrasing
- Termbase
- Brandbook
- Taxonomy
- Sign-off flow
Rules you own and version
Change a rule, test it on a sample, and roll it out – or roll it back with one click. A/B test tone and wording by market.
Checks nobody can skip
Required fields, character limits, and taxonomy are hard checks. If output fails, it doesn't go anywhere.
Judgement calls stay human
Tone, claims, and regulated language get flagged for review. Your specialists can overrule – and the reason is recorded.
An audit trail on every field
See which rules fired, which source data was used, and which checks passed. For every product, every time.
Corrections that follow the facts
When a product fact changes, every language regenerates from it. No hunting down six translations by hand.
Inside Speedtrain
What running Speedtrain actually looks like
Four screens from a normal day: setting a rule, tracing an output, clearing the exceptions, and watching it go live.
Set the rules – no code required
Every attribute gets its own rules: length, tone, what it must mention, what it must never say, and which locale it's written for.
Rules are versioned, so you can test a change on a sample before it touches the whole catalogue – and roll it back with one click.
Every output, traceable
See exactly why a description came out the way it did: which rules fired, which source fields were used, and which checks it passed.
Audit trail by design – not something you piece together afterwards.
Your team reviews the exceptions
Anything that needs a human decision lands in a review queue, assigned to the right person, with the reason it was flagged.
Speedtrain handles the 99%. Your team owns the 1% that matters.
Regulated language: clinically proven comfort – unsupported claim
SEO score 5.2 / 10 – suggested rewrite available
Forbidden term: best-in-class – superlative
Approved once. Live everywhere.
Approved content is written back into your PIM or CMS and made available to every channel – no copy and paste, no manual exports.
Speedtrain runs on Enterspeed Core, so your webshop, marketplaces, and AI agents all read from the same source via Query API, Delivery API, and MCP server.
What teams use it for
Start where the pressure is highest
Most teams start with one problem and add the next once the first one is solved.
Product onboarding
Supplier file in, live in every market. Thirty minutes a product becomes two.
Product onboarding →New markets and languages
Open a market in local tone, without multiplying the content budget.
New markets and locales →SEO, category, and campaign copy
AI drafts, your rules decide what goes live, and every field carries its source.
Governed AI content →Visibility in AI answers
Complete, consistent product content that answer engines describe correctly – and credit to you.
Show up in AI answers →Works with what you have
No new system of record. No migration.
Speedtrain works on top of the systems you already run. It writes into them – not around them.
From your sources
- PIM
- ERP
- DAM
- Supplier feeds
- Spreadsheets
- GenerateWrites every field against your per-attribute rules
- CheckRuns quality checks and keeps the audit trail
- RouteSends exceptions to the right person before anything is written back
- RegenerateUpdates every language when a product fact changes
Back into
- PIM or CMS
- Webshop
- Marketplaces
- Search
- AI agents
What stays exactly where it is
Keeps the master product record. Speedtrain reads from it and writes approved content back.
Keeps your published content and your editorial process.
Keeps your images and assets, and manages them the way it does today.
Your source data decides what can be claimed. Your rules decide how it's said.
Speedtrain vs. the alternatives
But can't we just use our regular AI tool?
You can – for one product at a time. The difference shows up at catalogue scale, when brand, facts, and approvals all have to hold, and the content has to reach every channel.
| Writing it by hand | Generic AI tools | Speedtrain | |
|---|---|---|---|
| Speed at catalogue scale | Limited by how many people you have | Fast – one prompt at a time | Whole batches at once – 100 products in under ten minutes |
| Brand and tone | Depends on who writes it | Depends on who writes the prompt | Your brandbook and termbase, applied to every field |
| Facts and claims | Checked by hand, when there's time | Can invent features and claims | Grounded in your source data, with regulated language flagged |
| Approval | Email threads and spreadsheets | None built in | Automatic for output that passes, specialist review for the rest – all recorded |
| Data it works with | Whatever's in front of you | Whatever you paste in | Product, price, stock, and content data from all your systems |
| Getting it live | Copy and paste into each system | Copy and paste into each system | Written back to your PIM or CMS, and live via API in your webshop, search, marketplaces, and AI agents |
| When a fact changes | Find and fix every version by hand | Prompt it all again | Every language regenerates from the corrected fact |
| AI model | – | Whatever the tool runs on | Your choice per task and per quality check – no lock-in to one AI vendor |
Built for enterprise
Made for many brands, many markets, and a lot of systems
Speedtrain is built for the setups enterprises actually run – and for a short conversation with IT, security, and procurement.
Many brands, many markets
Separate rules for every brand, market, and language – all running from one catalogue.
Your choice of AI model
Choose the model per task – and even per quality check – so you control quality and cost, without being locked to one AI vendor.
Security on record
ISO/IEC 27001 ISMS – Enterspeed's Information Security Management System is certified by Insight Assurance. GDPR-compliant. See the Trust Centre.
Your core systems stay closed
Channels and AI agents get only the data they're allowed to see – like the right price for each customer – without direct access to your ERP or PIM.
Uptime you can plan around
An uptime SLA on the Delivery API, depending on your agreement – with live status at enterspeedstatus.com.
People who've done it before
Our own team or Gold Partners like JAKALA and Novicell handle the implementation, with a Customer Success Manager as your day-to-day contact after launch.
Who it's for
One system. A different win for every seat at the table.
Built for enterprises in retail and e-commerce with big catalogues, several markets, and a brand worth protecting.
Heads of e-commerce
Products go live when they're ready to sell – not when the copy finally is. Launches stop waiting on content.
Content and marketing teams
Validate instead of writing the same product text for the hundredth time, and keep control of tone in every market.
Master data and PIM teams
Fill the gaps in supplier data without growing the team in step with the catalogue. Every change traceable to its source.
IT and tech leads
No new system of record and no point-to-point integrations. Speedtrain writes into the stack you already run.
Frequently asked questions
How is Speedtrain different from other product enrichment tools?
Enrichment is where Speedtrain starts. Because it runs on Enterspeed Core, the same approved content is combined with price, stock, and data from all your systems, and delivered live to your webshop, search, marketplaces, and AI agents. Your mappings stay yours, so you're never locked inside one vendor's platform.
How is Speedtrain different from ChatGPT?
Speedtrain is a governed workflow, not a chat window. It works with your rules, termbase, and brand guidelines, has quality checks and human approval built in, and writes approved content back into your systems and out to every channel via API.
How fast can we see value?
You see real output on your own products in the first demo. From there, typical time to value is weeks, not months, depending on how many systems and markets are involved.
Can we control tone of voice and compliance?
Yes – that's the point. You give Speedtrain your brandbook, termbase, legal phrasing, and forbidden words, and every output is checked against them before it's approved.
Does it work across languages and markets?
Yes. Rules are set per locale, so each market gets its own tone, terminology, and required mentions – all from the same product facts.
Which data sources can we use?
Your PIM, ERP, and DAM, supplier feeds in XML or JSON, spreadsheets, and custom integrations via API. XL-BYG, for example, runs Speedtrain directly on its Perfion PIM.
Do we need to replace our PIM?
No. Your PIM keeps the master product record. Speedtrain reads from it and writes approved content back – nothing is migrated or replaced.
Is our data secure?
ISO/IEC 27001 ISMS – Enterspeed's Information Security Management System is certified by Insight Assurance. We're GDPR-compliant, and you'll find our security documentation in the Trust Centre.
How much does it cost?
Every project is different, so we price based on your use case rather than fixed price points. Our pricing page shows examples based on real use cases, and you start with a free demo on a sample of your own data.
See Speedtrain on your own products
Send us an export of your product data, and we'll show you what Speedtrain does with it – real output, not sample content.
