# Keeyu — Supporting Pitches *Strategic memos and bear-case decks for IC and partner Q&A* *Investor memo, May 2026* --- ## What's in this document Four supporting materials live alongside the main investor deck. This file consolidates the two written memos into one reference; the two interactive decks are linked below for live walkthrough. ### Decks (interactive, browser-based) 1. **Why Shopify Won't Build Keeyu** — 9-slide deck (~5 min) → https://keeyupitch.com/shopify/ 2. **The US Market Opportunity** — 9-slide deck (~6 min) → https://keeyupitch.com/us-market/ ### Memos (full written, included below) 3. **Why Shopify Won't Build Keeyu** — defensive thesis 4. **The Roadmap Shopify Chose Not to Build** — offensive thesis --- --- # Memo 1 — Why Shopify Won't Build Keeyu *Investor memo, May 2026* --- ## The question, and what's underneath it "Can Shopify do this?" sounds like one question. It's three. → Could Shopify technically build it? Yes. → Will Shopify decide to build it? No. → If they did, could they win? No. This memo focuses on the second question, because it's the one that matters. The bear case here isn't whether Shopify has the engineering talent. It's whether they'll point that talent at post-purchase operations. Every public signal says they won't. Here's why. --- ## What Shopify is actually focused on Look at the 2025-2026 product roadmap. Every major investment sits in front of the buy button. Not one sits behind it. **Universal Commerce Protocol (UCP).** Open standard co-developed with Google. Defines how AI agents transact with merchants. Backed by Amazon, Meta, Microsoft, Stripe, Visa, Mastercard, Walmart, Target. Pre-purchase infrastructure. **Agentic Storefronts.** Merchants now sell directly inside ChatGPT, Microsoft Copilot, Google AI Mode, Gemini. AI-driven traffic to Shopify stores up 8x YoY. AI-initiated orders up 15x. Pre-purchase distribution. **Shopify Catalog.** Opened to every brand (even non-Shopify merchants) via the new Agentic Plan. The universal product feed for AI agents. Pre-purchase discovery. **Sidekick (Winter '26 RenAIssance Edition).** Re-architected as a "proactive AI coworker" for merchants. Writes product descriptions. Builds Shopify Flow automations. Sets up taxes and payments. Theme design. It's a merchant admin productivity layer that lives inside Shopify Admin. It has no visibility into WMS, carriers, returns platforms, or helpdesks. **Shop Pay, Shop Promise, Shop App.** Checkout, delivery badge, customer-facing tracking. All pre-purchase trust signals or notification layers. Harley Finkelstein, Upfront Summit, March 16, 2026: > "AI agents will become the new front door for e-commerce." Front door. Not the entire house. --- ## They already tried operations. They quit. This is the part most investors miss, and it's the strongest evidence in the file. → 2019: Launched Shopify Fulfillment Network. Committed $1B over five years. → 2022: Acquired Deliverr for $2.1B. Largest acquisition in Shopify history. → May 2023: Sold all of it to Flexport for a 13% equity stake. Cut 20% of staff. Tobi Lütke's letter to employees, May 2023: > "Building the logistics infrastructure was a side quest. We've been subtracting everything that's in the way of making Shopify the go-to platform for merchants to build their e-commerce stores." Harley Finkelstein on CNBC the same week: > "This allows Flexport to do what they do best, and allows Shopify to go back to doing what we do best, which is building incredible software for e-commerce." Shopify told the market plainly: we are not the operations company. If they wouldn't own warehouses, they're not going to build the orchestration layer that sits above 50+ disconnected operational systems they don't own. --- ## The "arm the rebels" doctrine Tobi's most quoted strategy line, ripped from Bill Gates' platform playbook: > "Amazon is trying to build an empire. Shopify is trying to arm the rebels." The Shopify App Store exists specifically so partners build everything Shopify doesn't want to. → 6,600+ apps in the App Store → 80%+ of Shopify merchants use 3rd party apps → Partner ecosystem generates more revenue than Shopify itself → 2.1M+ jobs in the partner economy vs ~5,000 Shopify employees For customer service specifically, Shopify's own help docs route merchants to Gorgias, Zendesk, Tidio, Rep AI. They built Shopify Inbox (basic chat) and stopped there. In Winter '26, Shopify launched **Sidekick App Extensions**, letting third-party apps plug their data and actions directly into Sidekick. This is the opposite of competing with apps. It's deeper integration. The partner ecosystem isn't an accident. It's the strategy. --- ## The structural reason it doesn't fit For Shopify to build Keeyu, they would have to: → Build deep integrations with WMS systems (NetSuite, CIN7, SAP, ShipStation, ShipBob, ShipHero) that compete with each other → Build deep integrations with carriers across geographies (Australia Post, USPS, UPS, DHL, Aramex, AfterShip) → Build deep integrations with returns platforms (Loop, ReturnGo, Refundid) → Build deep integrations with helpdesks (Gorgias, Zendesk, Kustomer) → Build real-time detection and decision logic across all of them → Build action orchestration that touches every system, all asset-light, all low-margin It also cannibalizes the highest-grossing partners in their app store. Gorgias, Loop, Klaviyo, ShipStation each pay Shopify millions in app store revenue. Building a Gorgias killer means killing a partner. This is why Lütke called operations a "side quest." It's not just a strategic preference. It violates the platform model. --- ## Common pushbacks, and how to handle them **"Couldn't Sidekick eventually do this?"** Sidekick is a merchant admin tool that lives inside Shopify Admin. It has no visibility into WMS, carriers, returns, or helpdesks. Its purpose is to make merchants more productive at Shopify-internal tasks (content, theme, segments, flows). The Winter '26 launch of Sidekick App Extensions explicitly tells partners to build on top of Sidekick, not the other way around. **"What about Shop Promise and Shop App? Aren't those post-purchase?"** Shop Promise is a delivery badge displayed at checkout, based on historical fulfillment data. It's a pre-purchase trust signal, not an operational layer. Shop App is customer-facing tracking. Neither orchestrates across the merchant's operational stack. Neither prevents a complaint. Neither replaces a helpdesk. **"Couldn't Shopify just acquire a Keeyu competitor?"** They could. But the 2023 Deliverr precedent argues against it. Shopify wrote down a $2.1B operational acquisition inside 18 months because the model didn't fit. Public statements consistently emphasize software focus, partner leverage, and capital return ($2B share buyback announced in Q4 2025). They are not signaling appetite for more operational M&A. **"They have billions in cash. Why wouldn't they?"** Cash isn't the constraint. Strategic focus is. They generate ~$2B in annual FCF. They authorized a $2B share buyback. They are returning capital to shareholders, not deploying it into new operational moats. If they wanted to be in operations, they would have kept Deliverr. **"What if AI changes everything and Shopify pivots to agentic ops?"** This is the strongest pushback, and the answer is: their AI strategy is the inverse of agentic operations. UCP, Agentic Storefronts, and Catalog are all about centralizing the storefront and decentralizing operations to partners. Sidekick App Extensions is the same pattern. Every AI bet they've made in the last 24 months extends the platform model, not contracts it. --- ## What would have to change To become a credible threat to Keeyu, three things would need to happen at once: 1. Shopify reverses Lütke's 2023 "side quest" decision and re-enters operations. 2. Shopify cannibalizes their app store by competing with their highest-grossing partners. 3. Shopify pivots their AI investment away from agentic discovery and into agentic operations. None of these are signaled in any public statement, earnings call, or product roadmap. The opposite is signaled in all three. If any of these signals appear, the thesis breaks and we should re-evaluate. Until then, the structural position holds. --- ## The position for the partners When the question comes up again, the answer is: > Shopify owns the front door of commerce. Keeyu owns the operating system behind it. Shopify tried operations once and sold it for a stake in Flexport. Their entire 2025-2026 roadmap is pre-purchase. Their platform model is built on app partners doing what they won't. Keeyu isn't competing with Shopify. Keeyu is one of the apps they're arming. If pushed further, ask: which announcement in the last 24 months suggests Shopify is going post-purchase? There isn't one. Every signal is the other direction. --- ## Bottom line The Shopify question is the right question to ask. It's also the wrong question to be afraid of. Shopify's success makes Keeyu more valuable, not less. More AI-driven orders means more channels merchants don't fully control, which means more operational chaos behind the buy button, which means more demand for an orchestration layer that prevents the failures. We are not on Shopify's roadmap. We are on the roadmap Shopify chose not to build. --- --- # Memo 2 — 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. --- ## Live decks for IC presentation → **Shopify Deck:** https://keeyupitch.com/shopify/ → **US Market Deck:** https://keeyupitch.com/us-market/ → **Supporting Pitches Hub:** https://keeyupitch.com/supporting/