# The Roadmap Shopify Chose Not to Build *How Keeyu becomes the post-purchase agentic operating system* *Investor memo, May 2026* --- ## The flip The previous memo answered why Shopify won't compete with Keeyu. This one answers the question that comes next: what does Keeyu become because of that? Short version: Shopify defined a strategic boundary at the buy button. Everything in front of the buy button is theirs. Everything behind it is open. The post-purchase stack today is fragmented across 8-12 point solutions that don't talk to each other. Whoever builds the orchestration layer that unifies them owns the same structural position Shopify owns for storefronts. That's the prize. Here's how Keeyu takes it. --- ## What the buy button leaves open Every D2C and mid-market e-commerce brand today runs a stack that looks something like this: **Pre-purchase (Shopify owns this):** → Storefront, checkout, payments, discovery, AI surfaces **Post-purchase (no one owns this):** → Helpdesk (Gorgias, Zendesk, Kustomer) → Order management (subset of NetSuite, CIN7, SAP) → WMS exception handling (ShipStation, ShipBob, ShipHero, Starshipit) → Carrier monitoring (AfterShip, Parcel Perform) → Returns (Loop, ReturnGo, Refundid) → Customer notifications (Klaviyo post-purchase flows) → Tracking pages (Malomo, Wonderment, AfterShip) → Subscription management (Recharge) Each of these tools sees one slice. None of them sees the whole picture. None of them can act across systems. The merchant becomes the integration layer, and the integration layer is human beings copying information between tabs. That's the structural problem. And it's the structural opening. --- ## The cannibalization sequence Keeyu doesn't try to replace this stack in one shot. The platform compounds in three stages, each one absorbing more of the post-purchase wallet. ### Stage 1: Cannibalize the ticket (Year 1-2) → Wedge: prevent complaints before they become helpdesk tickets → What gets absorbed: ticket volume in Gorgias, Zendesk, Kustomer → Customer outcome: smaller helpdesk spend, smaller CX headcount → Proof: EHP Labs went from 18 CX agents to 8. $455K saved. 10x ROI. The wedge is intentionally narrow. It builds the data infrastructure (detection across every post-purchase system) that powers everything else. ### Stage 2: Replace the helpdesk (Year 2-4) → Once detection runs across all systems, the merchant doesn't need a ticket queue. They need a brain that decides and acts. → What gets absorbed: helpdesk seats themselves. Not just ticket volume. → The structural argument: if 60-80% of e-commerce tickets are automatable (order modifications, tracking, returns, FAQs), and Keeyu prevents or auto-resolves them, the helpdesk becomes a vestigial organ. → Customer outcome: CX transforms from cost center to profit center. The remaining humans focus on revenue-generating conversations (retention, win-back, upsell). This is the labor replacement stage. The TAM expansion is significant because helpdesk SaaS plus CX labor is a much bigger budget line than helpdesk SaaS alone. ### Stage 3: Become the post-purchase operating system (Year 4-10) → With detection, decision, and action layers running across the entire post-purchase stack, Keeyu becomes the natural home for every function that touches the customer promise. → What gets absorbed: tracking, returns orchestration, customer notifications, exception management, subscription rescue, refund automation, fraud-flag triage. → The platform play: Keeyu opens APIs. Partners plug in. Specialized agents (returns specialists, fraud specialists, subscription specialists) run on top of Keeyu's data and action layer. By Stage 3, Keeyu isn't competing with the point solutions. The point solutions are running inside Keeyu. Same playbook Shopify used for storefronts. Different category. --- ## Why Keeyu specifically wins this Four structural reasons. **1. The detection layer is the wedge that requires building everything else.** To detect a complaint before it becomes a ticket, you need real-time visibility across Shopify, the WMS, the carriers, the returns platform, payments, the helpdesk, and the customer comms layer. Once you have that visibility, you are the only system that sees the whole post-purchase picture. Helpdesks see tickets. Tracking apps see shipments. Returns platforms see returns. Keeyu sees the order from promise to delivery to return, across every system that touches it. That visibility is structurally hard to replicate. It took 30+ integrations to build. **2. Action follows data.** Once you can detect across systems, action across systems is an API extension. The merchant doesn't need a separate tool to issue a refund, reroute a package, update inventory, and notify the customer. Keeyu does all of it in one orchestrated sequence. Every other tool can only act inside its own walls. Keeyu acts across them. That's the unfair advantage. **3. The data flywheel compounds.** Every issue Keeyu resolves teaches the system a pattern. Patterns become preventive rules. Preventive rules become automated playbooks. Automated playbooks become agentic workflows. This is what makes the "Proactive AI" position defensible. Every customer that ships volume through Keeyu makes the prevention engine smarter for every other customer. The category is winner-take-most. **4. CX-to-CEO reporting is the operating model.** Keeyu only works when CX owns post-purchase operations. That's a deliberate structural choice. It transforms CX from a cost center buried under the COO into a profit center that reports to the founder. Once a brand makes that shift, they don't unwind it. Switching costs become organizational, not just technical. --- ## The accelerant: agentic commerce makes this more urgent This is the timing argument. As Shopify pushes more orders through AI agent surfaces (ChatGPT, Gemini, Copilot, Perplexity), merchants lose direct control of the customer journey. More orders happen in channels merchants didn't build, with metadata they don't fully own, and customer relationships that get mediated by an AI layer they don't control. The result: behind every AI-initiated order is a brand that needs to fulfill a promise it made indirectly. Operational complexity behind the buy button grows in lockstep with discovery fragmentation in front of it. Shopify is centralizing the front door. That decision is going to accelerate post-purchase fragmentation, not reduce it. Keeyu is the calm behind that fragmentation. AI-driven orders on Shopify grew 8-15x year over year in Q1 2026. Every one of those orders needs a fulfillment promise kept. Every kept promise needs a system that detects, decides, and acts when something goes wrong. Keeyu is positioned for exactly that demand curve. --- ## The platform analogy The honest framing for the partners is this: → Shopify is the operating system for everything in front of the buy button. Their enterprise value is justified by being the de facto OS for storefronts in a fragmented merchant world. → Keeyu is the operating system for everything behind the buy button. Same playbook. Different category. The post-purchase OS for a fragmented operational world. The two systems are complementary, not competitive. Shopify's growth makes Keeyu more valuable. AI agent commerce makes Keeyu more necessary. The pre-purchase OS exists today and is publicly traded at ~$130B. The post-purchase OS doesn't exist yet. Keeyu is building it. --- ## What the partners should hold in their head When the question shifts from "Can Shopify do this?" to "How does Keeyu actually become this?", the answer is: > Keeyu starts narrow (prevent the ticket), cannibalizes the helpdesk (replace the seats), then absorbs the broader post-purchase stack (tracking, returns, notifications, exception management). Each stage builds infrastructure the next stage needs. By the time the platform is mature, Keeyu is to post-purchase what Shopify is to storefronts. The same compounding model. The same partner-ecosystem playbook. A different category that Shopify has explicitly chosen not to enter. The cannibalization sequence is the moat. The agentic commerce shift is the tailwind. The category creation is the prize. --- ## Bottom line The Shopify question and the cannibalization question are the same question viewed from two angles. Shopify's "side quest" exit from operations isn't just a defensive footnote in the Keeyu thesis. It's the strategic opening that makes Keeyu possible. Shopify drew a line at the buy button. Everything they chose not to build is what Keeyu is building. We are not on Shopify's roadmap. We are the roadmap Shopify chose not to build. And that roadmap is worth as much as the one they did.