Logistics and distribution
Algorithmic pricing lawsuits: what distributors using AI pricing should know
By SourceX Editorial · Reviewed by Noah Loul ·
Short answer
Algorithmic pricing antitrust cases target one pattern above all: competitors feeding nonpublic, current pricing or sales data into a shared tool that sends price recommendations back to each of them. A distributor pricing with AI on its own transactions sits in a different position, so the first question for any vendor is whose data trains the model.
Key takeaways
- The lawsuits so far focus on shared vendors that pool competitors' nonpublic data, not on a company pricing with software on its own data.
- Courts and enforcers look closely at whether users feel bound to follow recommendations and whether pooled data is current and granular.
- Ask every pricing vendor in writing which data trains the models that serve you and whether your transactions train models used by competitors.
- Licensing historical pricing data to an AI developer raises a related question, so permitted-use terms and data age matter.
- Antitrust exposure turns on specific facts and is assessed with counsel, case by case.
What are the algorithmic pricing lawsuits about?#
The algorithmic pricing lawsuits are about whether competitors can reach an unlawful agreement through a shared software vendor instead of talking to each other. Plaintiffs describe a hub-and-spoke pattern: each company sends its nonpublic data to the same vendor, the vendor's algorithm returns recommended prices, and prices across the market move together.
The best-known cases involve rental housing software, with later suits over hotel room rates and health care claims pricing. Results have been mixed. Some complaints were dismissed where the tool did not pool confidential competitor data or users kept full freedom to set their own prices; others moved forward or settled. Federal enforcers have weighed in with court filings, and several states and cities have passed or proposed rules aimed at shared pricing algorithms, most of them focused on rent.
This is general information, not legal advice. The law in this area is moving, and the facts of each tool matter.
Why distributors should pay attention#
Distributors should pay attention because the pricing tools they buy can look structurally similar to the ones in the cases. Pricing optimization software for electrical, plumbing, industrial and building products distribution is sold to many companies in the same vertical, and some products offer market indexes or benchmarks built from data their customers contribute.
Most distributor pricing tools learn from the distributor's own invoices, quotes and win and loss history, which is a different position. The questions arise when a feature blends in other customers' nonpublic transaction data, when recommendations arrive as targets the sales team is pressed to accept, or when the vendor markets the product as a way to avoid price competition.
Buying groups, trade association surveys and supplier-run programs can raise similar information-sharing questions and deserve the same review.
Which pricing tool features raise or lower concern?#
The pricing tool features that raise or lower concern are the ones courts and commentators keep returning to: whose data is pooled, how fresh it is, how binding the recommendations are, and what users can see. No single feature decides a case, but the pattern across them shapes how a tool is viewed.
| Feature | Lower concern | Higher concern |
|---|---|---|
| Training data | Your own transactions only | Pooled nonpublic transactions from competing users |
| Data age and detail | Historical, aggregated or delayed | Current, granular, customer-level net prices |
| Recommendations | Advisory and easy to override | Acceptance tracked and pushed by the vendor or management |
| Outputs | Guidance on your own prices | Visibility into competitors' price levels or plans |
| Vendor messaging | Margin and win rate on your own book | Promises of market discipline or fewer price wars |
| Records of decisions | Documented reasons for price changes | No record of who decided and why |
Questions to ask your AI pricing vendor#
The questions to ask an AI pricing vendor should establish, in writing, what data flows in, what flows out and who decides prices. Ask them before renewal or before switching on a new feature, and file the answers with the contract.
- Which data trains or calibrates the models that generate our recommendations?
- Does any model, index or benchmark we use include nonpublic data from other customers, including competitors?
- Are our invoices, quotes or margins used to train models or build benchmarks for other customers?
- How old and how aggregated is any pooled data before it influences a recommendation?
- Can our team override every recommendation, and does the product or the vendor track acceptance?
- Can we opt out of data pooling without losing core features?
- What antitrust review has the product received, and will you describe its safeguards?
- What happens to our data if we leave or if your company is acquired?
What to keep on file about pricing decisions#
The records to keep on file about pricing decisions are the ones that show your company set its own prices. If a pricing practice is ever questioned, in a lawsuit, an investigation or an acquirer's diligence, contemporaneous records carry far more weight than recollections.
Useful records include the written pricing policy and its revision history, the approval trail for matrix and customer-specific price changes, override reasons captured in the ERP rather than in email, and the vendor's written answers about data pooling. Keep the contract versions and feature settings that were active at each point, because a tool's behavior can change with a single release.
These same records make pricing history more useful if you ever license it, since they explain why prices moved, not only that they moved.
Where data licensing fits in#
Data licensing fits in because licensing pricing history to an AI developer raises a related question: where could that data end up? If current, customer-level net prices were licensed to a developer that builds pricing tools for your direct competitors, the arrangement could start to resemble the pooled-vendor pattern.
Distributors that license pricing or quote records usually reduce that risk through both the terms and the data itself, with counsel reviewing the specifics.
Vendor contracts deserve the same reading. When a pricing vendor's terms let it train shared models on your invoices, you have in effect granted a license to your pricing data, often without payment and with fewer controls than a negotiated data license would carry.
| Safeguard | What it addresses |
|---|---|
| Age cutoff on pricing records | Keeps current prices and margins out of the dataset |
| Aggregation or removal of customer-level net prices | Limits competitively sensitive detail |
| Permitted-use terms that exclude pricing products for your vertical | Prevents reuse in tools sold to your competitors |
| Ban on combining with other distributors' pricing data for recommendations | Avoids building a pooled pricing hub |
| Removal of customer and supplier identifiers | Protects confidentiality and supplier terms |
Illustrative: an industrial distributor reviews its pricing stack#
Illustrative: a fictional industrial MRO distributor uses a pricing add-on connected to its ERP. During renewal, its general counsel reads that a new 'market index' feature blends anonymized invoice data from other subscribers, several of them direct competitors, and that reps are scored on how often they accept recommended prices.
She sends the vendor the questions above. The vendor confirms that the core models train only on the distributor's own data, but the index pools current subscriber transactions. The company declines the index feature, removes acceptance scoring from rep scorecards and adopts a short pricing policy stating that branch managers set final prices.
Later, when the distributor considers licensing older quote and win-loss history to an AI developer, counsel applies the same lens: an age cutoff, no customer identifiers, and permitted-use terms that exclude pricing products sold to distributors in its vertical.
How SourceX handles pricing records#
SourceX handles pricing records as competitively sensitive from the start. In the Rights step of the SourceX five-step transaction, pricing and quote history is reviewed with the supplier's counsel for supplier confidentiality terms and competition concerns before any sample is shared.
In Preparation, customer identifiers are removed and pricing can be aged or aggregated. The permitted use agreed with the buyer is recorded in the SourceX Evidence Packet, and the supplier approves the final scope. Records the supplier is not comfortable licensing simply stay out.
Frequently asked questions
Is using AI pricing software illegal for a distributor?
No general rule makes it illegal. Most cases challenge specific arrangements in which competitors pool nonpublic data through a common vendor and follow its recommendations. Whether a particular tool raises concern depends on its data sources, its features and how your team uses it, which is a question for antitrust counsel.
Does monitoring competitors' public prices raise the same issue?
Watching public prices, such as list prices on competitors' websites, is generally treated differently from exchanging nonpublic data through a shared tool. It is not automatically free of risk, especially if it feeds a system that coordinates with others, so describe your practices to counsel rather than assuming.
Should we document how prices are set?
Yes. A short written pricing policy, records of who approved price changes and why, and evidence that recommendations can be overridden all show independent decision-making. They also make vendor reviews and future acquisition diligence faster.
Do buying groups and benchmarking surveys raise similar questions?
They can. Information exchanges through trade associations, buying groups or third-party surveys are a long-standing antitrust topic. Common safeguards include using older, aggregated data and an independent administrator, but counsel should review the specific program before you contribute data.
Can a distributor license historical pricing data at all?
Often yes, with care. Older, aggregated pricing and quote history with customer identifiers removed is very different from a live feed of current net prices. The rights review, permitted-use terms and preparation choices decide what is appropriate, deal by deal and with counsel.
Related resources
See if your company qualifies
A short company assessment. No data uploads are needed.