WAIR
WAIR is an agentic AI platform for lifestyle retailers — apparel, footwear, sport, eyewear. Its ForecastGPT engine powers three SuperAgents that forecast demand, allocate and replenish stock, plan assortments and write product content inside existing systems.
What is WAIR?
WAIR describes itself as an agentic AI retail technology company built specifically for lifestyle retail: apparel, sport, footwear, eyewear and underwear. What it sells is not a dashboard but a set of SuperAgents, digital co-workers that detect, decide and act in a closed loop inside the retailer's existing systems.
All three run on a shared engine, ForecastGPT 2.5, a deep learning model trained on pooled, anonymised data from across WAIR's client base and then retrained on each retailer's own sales, stock and promotion history. It reads more than 100 variables: demographics, geography, weather, seasonality, product images, price, promotions, in-store behaviour. Note that the site quotes two different model sizes. The ForecastGPT page advertises a billion parameters, while the Technology page gives 7.5 million for 'WAIR model 2.5'. Both figures are published as they stand.
Wallie, the allocator, handles initial distribution of new lines, daily replenishment, replenishment cut-off at size level, store-to-store and store-to-warehouse redistribution and end-of-life stock, forecasting up to 12 weeks ahead. Whoopie, the planner, covers financial planning, assortment planning, order generation and continuous replanning; the Super agents page labels it In Beta, while the FAQ still calls it Woopie and lists it as coming soon. Suzie, the content creator, enriches product attributes, writes product copy and localises it, in 120+ languages according to the FAQ and 100+ according to the Agentic AI page.
The client chooses how much autonomy to grant: suggestion only, human validation, or automatic execution within defined guardrails. Integration goes through a single REST API, and the ERP stays the system of record, with no rip-and-replace.
Claimed results are specific: lost sales down 84%, overstock cut from 50% to 4%, forecast accuracy above 90%, and named-client figures such as +2.96% revenue at Shoeby, +29.28% accuracy at Daka and publishing three times faster at OFM. Van Dal, Berden, Baukjen and DK Company are also cited as customers. In 2024-2025 WAIR took over the software, the assets and the client base of Retailisation B.V., which included VF Corporation, Ralph Lauren Corp, Wolford, 7 for All Mankind and DK Company.
What it does
- Forecast demand down to SKU, store and size, up to 12 weeks ahead
- Allocate new arrivals store by store from launch day
- Replenish daily and cut off replenishment for stores that are already saturated
- Move stock between stores, and between warehouse and store, mid-season and at end of season
- Turn top-down financial targets into weekly plans by region and store
- Generate purchase order suggestions by store, channel and delivery window
- Enrich product attributes, write titles, tags and descriptions, then localise them per market
When to use WAIR / When not to
A quick filter to help you decide if WAIR is the right fit.
When to use WAIR
- Multi-store lifestyle retailers and brands — apparel, footwear, sport, eyewear, underwear — running anywhere from 10 to 10,000 selling locations
- Merchandising, planning and buying teams still running allocation and replenishment out of ERP exports and spreadsheets
- E-commerce and product content teams that need attributes, titles and descriptions enriched and localised market by market
- Retailers determined to keep their existing ERP as the system of record, with no rip-and-replace project
- Retail organisations with no data science team of their own: the FAQ states WAIR handles the modelling and the day-to-day decisions
When not to use WAIR
- Anyone hoping to sign up and try the product alone: there is no free plan and no self-service registration, only a booked demo
- Buyers who need a published price before talking to sales, since no rate, tier or billing unit appears anywhere on the site
- Grocery, DIY and industrial B2B distributors: WAIR positions itself explicitly and only on lifestyle retail
- Young or recently launched retailers, as the integration requires at least two full calendar years of sales history
- Teams that need results within weeks: integration runs about six weeks, onboarding and optimisation about twelve, with go-live around week 10
How to use WAIR
A typical end-to-end flow, from setup to results.
- Try the Suzie content generator directly on the WAIR site: an image URL, a target language and an email address are enough to see one agent at work
- Book a discovery call, since 'Book a demo' is the only way in and there is no self-service sign-up
- Scope a pilot with the WAIR team: typically 8-10 stores over three months on NOOS lines, or 20-50 stores on a single product group
- Appoint the internal roles the programme needs: a change leader, a project champion and key users
- Open the integration development portal and the public API documentation on docs.wair.ai, with its Postman collection
- Push the five required datasets to the single REST API: items and images, stock mutations, sales covering at least two full calendar years, transfers and locations
- Avoid custom development where possible with the Business Central connector, the CSV import, or an integration partner such as Be-Terna, Qwentes, Intext or Mobi Media
- Schedule at least one daily upload, and more frequent ones for stock and sales
- Follow the published ten-week timeline: integration week 1, test data week 2, validation week 4, backtesting week 6, training week 8, go-live week 10
- Pick a governance mode per agent, then work from the customer portal at customer.wair.cloud, with support@wairforretail.com for integration questions
Pros & Cons
Pros
- Published results attributed to named clients: +2.96% revenue at Shoeby, +29.28% forecast accuracy at Daka, three times faster publishing at OFM, with measurable effects announced four to six weeks after a pilot goes live
- The ERP stays the system of record, so there is no rip-and-replace of the existing stack
- Graduated governance: the retailer decides what runs automatically and what goes through human validation
- No client-side data science team required, since WAIR states it handles the modelling and the day-to-day decisions
- No consulting or onboarding fees: the onboarding page says every hour spent on adoption, training and sparring is included
- Outcome-based commercial model, with a claimed return on investment of at least 5x and up to 50x for the best performers
- Public and detailed API documentation, with a Postman collection, a Business Central connector and a network of integration partners
Cons
- No public pricing at all: no pricing page, no range, no billing unit, so no budget can be framed before a sales conversation
- No platform terms of service; the only contractual document published covers the Wair Connector App
- The model trains on pooled customer data, with no documented opt-out
- No DPA, no product sub-processor list and no security certification such as SOC 2 or ISO 27001 mentioned anywhere
- No data hosting country or region is disclosed
- Whoopie is still In Beta, and the FAQ still calls it Woopie and lists it as coming soon; that same FAQ page carries leftover test blocks and an editorial instruction left in plain sight, and the site quotes two irreconcilable model sizes
- Long time to value: roughly 8 to 12 weeks from first integration work to go-live
Pricing & Plans
WAIR publishes no prices. There is no pricing page, no price range, no currency and no billing unit; the /pricing/ URL returns a 404 and the sitemap lists no tariff page. No free plan and no free trial are announced, and the only commercial entry point is a booked demo, after which a quotation is issued. The company describes its model as 'Outcome as a service' and states, on its onboarding page: 'We're compensated based on the outcomes we deliver together: improved margins, faster inventory turns, and measurable growth.' The same page adds that there are no consulting or onboarding fees, adoption and training hours being included in the engagement, while the FAQ claims a return on investment of at least 5x and up to 50x. A typical pilot runs three months across 8-10 stores and is not presented as free.
- the site distinguishes no editions and no licence levels
- The only visible segmentation is by SuperAgent
- Wallie
- Whoopie and Suzie being presentable separately or together
- Whoopie is announced In Beta
- with access on a waiting list ('Want first access?')
- compensation indexed on delivered outcomes rather than on a subscription tier
- a three-month pilot on 8-10 stores on NOOS lines
- or 20-50 stores or one product group according to the FAQ
Data, GDPR & hosting
A consolidated view of how WAIR handles your data.
GDPR overview
The FAQ claims 'Full GDPR compliance and data governance', a phrase that appears nowhere in the privacy policy itself. The policy lists the usual rights — access, rectification, erasure, objection, restriction and withdrawal of consent — but hedges them: 'Depending on your location, you may have the right to...'. Requests go by email to the address named in section 14, a sales alias on a separate publisher domain. International transfers are acknowledged: data collected via the site may be transferred to and processed in countries outside your residence, with a general statement that WAIR takes steps to ensure such transfers comply with applicable privacy laws. Beyond that the documentation is thin: no data protection officer is named, no Article 27 EU representative (the publisher is established in the Netherlands), no DPA is offered or mentioned, and no security certification such as SOC 2 or ISO 27001 appears anywhere.
Who owns the data?
The privacy policy covers the website. It says WAIR collects names, emails, company, phone, job title, messages, IP address, browser and pages viewed, and states plainly: 'We do not sell your personal information.' Sharing is limited to service providers (hosting, analytics), legal authorities and a possible business transfer. On the product side, the site says client outputs remain isolated, that no business data is shared or exposed, and that decisions are based solely on the client's own data. No platform terms of service are published, so ownership of the sales, stock and product data a retailer pushes to the API is nowhere set out in a public contract.
Reuse rights
Website data is used to handle demo requests, manage accounts, personalise content, measure site usage, run advertising and remarketing, and meet security and compliance obligations. Product data goes further. WAIR states that Forecast-GPT trains across all clients simultaneously and that the system is continuously learning from anonymized, aggregated data across the WAIR network; pooled data is defined by the company as anonymized data from multiple clients combined for model training. The model is then retrained on each client's own sales, stock and promotion history. Outputs are said to stay isolated per client, but no opt-out from pooled training is documented on the site, and no platform terms set out what a customer may do with the decisions returned.
Data retention & training
Hosting summary
WAIR names no hosting country and no hosting region. The privacy policy addresses the question only in the negative: 'Data collected via the Site may be transferred to and processed in countries outside your residence', followed by a general assurance that the company takes steps to make such transfers comply with applicable privacy laws. No jurisdiction, cloud provider or data centre location is given for the platform itself. The only third parties named are those serving the website — Google Analytics, Google Ads Remarketing and HubSpot — which are site vendors rather than product sub-processors, and no sub-processor list for the platform is published. The site is served behind Cloudflare, which tells you where the pages travel, not where customer data lives. The client portal runs at customer.wair.cloud and the documentation on docs.wair.ai. The publisher is established in the Netherlands, so an EU jurisdiction is plausible, but nothing on the site states it: anyone with a data residency requirement will have to get that answer from the vendor directly.
Things to keep in mind
Risks and trade-offs to weigh before adopting WAIR.
- Pooled, anonymised customer data trains the shared model and no opt-out is documented, so your sales history feeds an engine that also serves other retailers
- No DPA, no product sub-processor list and no security certification are published, leaving a compliance review with very little to work from
- No hosting country or region is disclosed; the privacy policy only says data may be transferred to and processed in countries outside your residence
- The only contractual document published covers the Wair Connector App, not the platform, so the terms you would actually sign are not public
- GDPR requests go to a sales alias on wairforretail.com, a domain separate from the site you are reading
- With no public price, no budget can be framed before entering a sales process and no cost comparison with other tools is possible
- Handing daily allocation and replenishment to an agent can dull a merchandising team's own feel for the assortment; the leftover test blocks on the FAQ page and the two contradictory model sizes are a reminder to keep reading the outputs critically
Setup & Integrations
Technical difficulty
Moderate on the engineering side, heavier on the organisational one. Integration goes through a single REST API whose endpoints are independent and can be built in parallel, with a Business Central connector and a CSV import available to avoid custom development, and a development portal provided to the client. Five datasets are required, including two full calendar years of sales, uploaded at least once a day. No data science team is needed. What costs time is the schedule and the people: integration around six weeks, onboarding and optimisation around twelve, plus a change leader, a project champion and key users.
Deployment
Integrations
Supported languages
Behind WAIR
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Who stays in control of inventory decisions?
Do we need a data science team to use WAIR?
How does WAIR integrate with our systems?
What data do we have to provide?
How long does implementation take?
When do results become visible?
How much does WAIR cost?
Does WAIR train its models on customer data?
How many languages does the product content agent cover?
Is there a mobile app, and is the API public?
Should you pick WAIR?
WAIR is an enterprise tool, not a self-service one. It targets multi-store lifestyle retailers and its strongest argument is that it acts rather than reports: the SuperAgents execute allocation, replenishment, planning and product content decisions inside the systems the retailer already runs, with the ERP left as the system of record and a governance dial that goes from suggestion only to full automation. The published results are unusually concrete for this market, attributed to named clients, and the API documentation is public and detailed.
The weaknesses are on the paperwork side. Pricing is entirely opaque: no page, no range, no billing unit, and the 'Outcome as a service' model means the budget cannot be estimated before a commercial conversation. Contractual and compliance documentation is thin, with no platform terms of service, no DPA, no product sub-processor list, no security certification and no disclosed hosting country or region.
The point that deserves the most attention is the pooled training. WAIR states plainly that its engine learns from anonymised, aggregated data across the whole client network, and no opt-out is documented anywhere. Outputs are said to remain isolated per client, but a retailer should raise the question in writing before signing.
Finally, budget the time. Integration runs about six weeks, onboarding and optimisation about twelve, with go-live around week 10, so eight to twelve weeks before the first measured effect. Add a few signs of a site moving faster than its own content: the planning agent is called Whoopie on its own page and Woopie in the FAQ, In Beta in one place and coming soon in another, and the model is credited with a billion parameters on one page and 7.5 million on another.
- Choosing a selection results in a full page refresh.
- Opens in a new window.