SPS Commerce Is Using Chat as the Gateway to Sell Autonomous AI Agents

SPS Commerce reported Q2 2026 revenue of $198 million on July 30, up 6% year over year, with the core business growing in high single digits. They also used the earnings call to disclose a very specific agentic AI monetization roadmap

SPS said the rollout starts with MAX Chat, a conversational interface that is now included with all SPS Fulfillment subscriptions at no additional cost. The company expects the chat interface to become the primary entry point for customers using the platform. 

The second layer is autonomous agents, systems that can detect supply chain anomalies and in many cases resolve them without human intervention. SPS plans to sell the agents as premium subscription bundles beginning in late Q4 2026.

"Users of MAX through the chat interface are more likely to explore advanced MAX features," CEO Chad Collins told analysts. "We expect this usage trend will define the path to monetization of our AI solutions."

What MAX Has Already Proven in Beta

The MAX beta program produced specific, named, and verified commercial outcomes, according to the press release. Branch Furniture, a wholesale brand selling to Williams-Sonoma, Lumens, and Office Depot, achieved 90% weekly time savings managing overdue orders with a key retail partner by using MAX to resolve order issues in minutes rather than days. 

The company said the improvement significantly reduced the time employees spent managing that retailer relationship. 

Beyond operational efficiency, MAX demonstrated anomaly detection that prevented direct revenue loss. During the beta phase, MAX caught a $290,000 invoice failure caused by an incorrect UPC code, a compliance error that would typically surface only after a retailer has rejected or deducted the invoice. 

The same capability flagged $70,000 in unacknowledged purchase orders for a food manufacturer and identified 100 stalled dropship orders for an outdoor brand before those issues escalated into retailer penalties.

The pattern across these examples is consistent. MAX is finding problems that existing systems were not designed to catch proactively, and finding them before the financial consequence arrives rather than after.

The Proprietary Data Moat Behind MAX

SPS Commerce has spent 25 years building a network that connects more than 3,500 buying organizations with tens of thousands of suppliers, according to the press release. 

In the process, it has accumulated proprietary databases of supply chain compliance expectations, trading partner requirements, and transaction patterns that are not available in any downloadable vendor guide and in many cases are not published anywhere. 

Retailers do not disclose their specific compliance rules publicly. SPS has them, built from decades of operational activity on its network.

"No other company can match the unique combination of AI capabilities, 25 years of proprietary data, deep domain expertise, and an expansive network access to drive this kind of tangible value," Collins said. 

The agents being built on top of MAX are trained on this proprietary compliance database, which is what allows MAX to guide suppliers toward compliant execution rather than simply flagging that something went wrong.

The customer evidence supports the claim. In a recent SPS customer survey, 83% said the SPS network data improved their own AI readiness, 87% cited improved scalability, and 100% said that without the SPS network they would need more headcount, more tools, more time.

The Path to Autonomous Agents

Collins described the specific mechanism through which MAX Chat leads to autonomous agent adoption. Customers currently using MAX Chat to investigate supply chain issues may require 20 prompts to reach an answer and resolution. 

SPS has observed that the workflows customers are completing through those 20-prompt sessions are candidates for full automation. The same outcome, delivered by an agent without requiring any user interaction.

"We are developing those types of agents on top of this MAX technology now," Collins said. "We believe that those agents that can do things more autonomously, identifying these anomalies and in many cases resolving them, not only finds the hard ROI in the supply chain savings, but also is going to be a very favorable headcount and efficiency impact for our customers."

The autonomous agent bundles will be tiered, with the first set of agents targeting the most complex customers with the largest number of trading relationships before expanding to medium and smaller customers over time, according to the press release. 

Collins confirmed that SPS expects to be in a position to sell agents starting late Q4 2026, with revenue contribution building from there. Q2 revenue of $198 million came in on 6% growth, with adjusted EBITDA of $66.6 million and trailing 12-month free cash flow up 40% to $198.7 million.