# Aide vs Arbyn: Governed Agentic AI vs Flat-Rate Support and Sales > Agentic AI platforms like Aide promise "governed" automation, but this translates to per-action fees that create budget anxiety, while flat-rate models like Arbyn offer predictable costs and unlimited potential. Source: https://arbyn.app/blog/aide-vs-arbyn-governed-agentic-ai-vs-flat-rate-support-and-sales Published: 2026-08-09 --- The most consequential decision a store owner makes about agentic AI has nothing to do with technology, platforms, or feature sets. It is, at its core, a payroll decision. When you deploy an AI that can take real, consequential actions in your store, issuing refunds, changing shipping addresses, canceling orders, or starting returns, you have hired a new type of employee. Like any employee, it can be paid in one of two ways: by the task, where every completed action adds a distinct line item to an invoice, or on salary, where one flat, predictable fee covers unlimited work. A task-based worker is a freelancer you manage with a stopwatch, scrutinizing every single action for its cost versus its benefit. A salaried team member is a trusted asset you empower to solve problems creatively, proactively, and at scale. This choice defines not only your monthly support spend but the entire operational philosophy of your business. It is the difference between an AI you constantly manage and an agent that truly works for you. One model creates endless budget anxiety and cognitive overhead; the other creates a predictable, firm foundation for growth and innovation. The "Governed" Agent: Paying by the Task A new class of AI tools has emerged, branding itself with the powerful and reassuring term "agentic." These platforms move far beyond the simple, pre-scripted chatbots of the past to perform complex, multi-step actions directly within your store's backend systems. The leading philosophical example of this model comes from a company called Aide, which smartly markets itself as a "governed agentic AI platform." The term "governed" is a masterful piece of positioning, designed to inspire confidence and assuage the primary fear of any business owner considering automation: the fear of AI running rampant, making costly, unapproved mistakes, and damaging customer relationships. This messaging is prudent and, in a technical sense, it is accurate. It speaks directly to the business owner's nightmare scenarios fueled by headlines about AI gone wrong. But in a business sense, this philosophy of governance has been translated by the entire industry into a specific, and deeply punitive, billing model: you pay for what the AI does, action by action. This is not unique to Aide; it is the dominant pricing strategy for nearly every major helpdesk's AI offering, turning prudent technical guardrails into punitive financial toll booths that charge you every time a customer successfully crosses the bridge to a resolution. Consider the established players in the customer support space, who have all bolted on agentic AI with this same costly model. Intercom’s powerful Fin agent explicitly charges a fee that industry experts have documented at around $0.99 for each "outcome" it successfully produces. This fee sits on top of the already significant monthly per-seat cost for your human agents. An "outcome" can be a final resolution, but it can also include a successful handoff to a human, meaning you can pay the AI a dollar simply for doing the basic job of routing a ticket. Gorgias, a long-time staple for Shopify stores, employs a model that many store owners find particularly painful. On top of your monthly plan's ticket allowance, its AI Agent adds a per-resolution fee of about $0.90 to $1.00 for every single conversation it handles autonomously. This creates a painful "double-billing" scenario, where a single customer interaction can incur two separate charges: one for the ticket to exist in the first place, and another for the AI to make it go away. Zendesk, the enterprise standard, takes this financial complexity to another level entirely. Its advanced AI agents bill per automated resolution at a rate industry analysts place between $1.50 and $2.00. This charge kicks in after a small initial allowance is exhausted and, critically, stacks on top of both the expensive per-agent monthly seat license and a potential $50 per-agent add-on for their "Copilot" feature, which provides essential AI assistance to human agents. This per-action, per-outcome, or per-resolution pricing model is the direct financial expression of "governance." The vendor provides a powerful tool, but every single exercise of that power is metered, measured, and invoiced. The sales pitch is that you only pay for value delivered, which sounds fair and logical on a demo call. The harsh reality, as thousands of store owners discover with shock when their first end-of-month bill arrives, is that the cost of this "value" is unpredictable, uncapped, and often grows much faster than the problem it was meant to solve. A successful marketing campaign that doubles your sales or a viral TikTok that triples your traffic should be a cause for celebration, not a trigger for financial panic about the corresponding, and now uncapped, spike in your support software bill. The governance, intended as a technical safety feature, becomes a tight financial leash, tethering your store's efficiency to a constantly running, unforgiving meter. Each time the AI successfully does its job, your bill goes up, creating a fundamental misalignment of incentives that is the hidden, fatal flaw in the pay-per-task model. The Anatomy of a Per-Resolution Bill Understanding a flat-rate subscription is simple; you know the cost before you commit. Understanding a usage-based AI bill, however, requires a degree in forensic accounting and a tolerance for unpleasant surprises. The final number on the invoice is a complex composite of multiple, often overlapping, charges that bear little resemblance to the advertised monthly price that got you in the door. For a store owner trying to forecast expenses and manage a P&L, this model is a nightmare of compounding variables. The base plan fee is just the cost of admission to the park. For a platform like Gorgias, you might start with their Basic plan at $60 per month, which includes a seemingly reasonable 300 support tickets. For a seat-based platform like Zendesk, this initial cost is the per-agent license for a plan like Suite Professional at $115 per agent, per month, while for Intercom, it could be the Advanced plan at $85 per agent, per month. In all cases, this is before a single line of AI code is even activated. This is the predictable part of the bill. It is also the smallest and most misleading part, serving as a low, inviting barrier to entry before the truly significant variable costs begin to accumulate. The first layer of variability, and the first trap for a growing business, comes from overages on the core service itself. If your store on Gorgias's Basic plan has a busy month thanks to a holiday sale and receives 400 tickets instead of the allotted 300, you will pay an overage fee on those extra 100 tickets, typically around $0.40 each, immediately adding $40 to your bill. But the second, and far more financially significant, layer of cost is the AI itself. Let's say of those 400 total tickets, your newly implemented AI agent successfully and autonomously resolves 250 of them. On top of any ticket overage fees you already incurred, you will now pay for each and every one of those AI resolutions. At a conservative rate of $0.90 per resolution, that is an additional $225 charge. A seemingly modest $60 base plan has suddenly ballooned to over $325 ($60 plan + $40 overage + $225 AI usage), a more than five-fold increase before any other factors are even considered. This is the "double-billing" trap in brutal action: you pay for the ticket to exist in the helpdesk (either as part of your base plan or as an overage), and you pay a second, separate time for the AI to resolve it. This complexity is not an accidental byproduct; it is the deliberate design of a model that itemizes every component of service to maximize revenue. For platforms with per-seat pricing, the math is different but the painful effect is precisely the same. A five-person support team on Zendesk's popular Suite Professional plan at $115 per agent per month pays a steep $575 a month just for their human agents to be able to log in. Adding the AI-powered Copilot feature, which most teams need for modern agent-assist functions like summarizing conversations, can add another $50 per agent, bringing the fixed monthly total to $825 before a single customer is helped. Then, the per-resolution meter starts its relentless ticking. If their AI agents resolve 2,000 conversations that month beyond their plan's small allowance, at an industry-reported rate of $1.50 per resolution, that adds a staggering $3,000 to the bill. The total monthly cost is now over $3,800, with the AI usage component making up nearly 80% of the entire cost. The advertised seat price is functionally irrelevant to the final invoice. It is a loss leader for the much more lucrative, and volatile, business of metering AI activity. The True Cost of Governance: Operational Hesitation The most damaging and insidious consequence of per-resolution pricing is not financial; it is psychological, infecting your entire operational culture. When every autonomous action taken by your AI agent results in a direct, measurable, and uncapped cost, it fundamentally changes your relationship with automation from one of empowerment to one of fear. The very tool that was supposed to liberate your team and scale your business becomes just another resource to be stingily managed and second-guessed. Store owners, who are by their very nature obsessed with controlling costs from shipping tape and packaging to ad spend, find themselves watching their AI usage meter with the same mounting anxiety they watch a poorly performing marketing campaign. This creates a widespread phenomenon of operational hesitation. You begin to constantly second-guess whether a particular automation is "worth it." Should the AI handle all WISMO (Where Is My Order?) inquiries, or should we route them to a human to avoid the fee? Should it be allowed to process a return for a $15 item if the resolution fee is $0.99, eating into the margin? The AI, a powerful capital investment, is relegated to being treated like a costly consumable good, used as sparingly as possible. This dynamic completely undermines and reverses the core promise of efficiency that led you to AI in the first place. You have purchased a sophisticated system capable of handling a significant percentage of your support volume, but you are now financially incentivized by the vendor's pricing model to use it as little as possible. To avoid the dreaded end-of-month bill shock, support managers start to manually cap the number of resolutions the AI is allowed to perform each day or month. They create complex internal rules and workflows to disable automations for what they deem "lower-value" interactions, forcing those tickets back into the human queue just to save a dollar. This not only increases direct labor costs but also drives up customer wait times and CSAT takes a nosedive. The "governed" agent, designed for safety and control, has created a new, far more costly form of internal governance where the primary concern is not customer satisfaction or team efficiency, but simply managing the software bill. The tool's success becomes your budget's failure, a perfect and tragic example of optimizing a single metric to the detriment of the entire system. This perverse situation creates a deep and unavoidable misalignment of incentives between you and your software vendor. The vendor profits every single time your AI resolves a ticket. Their business goal, therefore, is to maximize the number of billable resolutions your account generates. Your goal as a store owner, however, should be to build a better business that generates fewer support tickets in the first place by improving product pages, clarifying shipping policies, and streamlining your fulfillment logistics. In a per-resolution model, if you successfully reduce your ticket volume by 30% through better business practices, the vendor's revenue from your account goes down significantly. The vendor is, in a very real and direct sense, betting against your ability to solve your underlying operational issues. You are paying a partner who financially benefits from your recurring problems and sees your operational improvements as a direct revenue threat. This is a fundamentally adversarial relationship, hidden beneath a pleasant veneer of "value-based" pricing. The Salaried Agent: Flat-Rate and Unlimited Action There is a different, more logical way to structure this relationship, one that completely aligns the incentives of the tool with the goals of the store owner. Instead of hiring an AI agent and paying it by the task like a temporary freelancer, you can hire one on a flat, predictable salary. This is the model Arbyn is built on from the ground up. For a single, fixed monthly fee, the AI agent works to its full capacity, 24 hours a day, 7 days a week. It can handle 150 conversations or 15,000, and the price does not change. This is not merely a pricing gimmick; it is a fundamental shift in operational philosophy. The strategic question is no longer "How can I limit my AI usage to control costs?" but rather "How can I maximize my AI usage to unlock its full, uninhibited potential?" The store owner's focus moves from defensive, anxious cost containment to offensive, exciting opportunity seeking, constantly asking "What else can I automate?" every single day to drive more value from a fixed cost. When the marginal cost of one additional AI resolution is zero, the entire dynamic of your support operation flips on its head. Every single task the AI can successfully automate becomes pure profit in the form of reclaimed time, reduced human labor costs, and faster customer response times. There is no longer any reason to hold back, hesitate, or second-guess. You can and should unleash the agent on every single automatable category of inquiry: all order status lookups, all shipping address changes, all return and exchange initiations, all product questions, and even proactive in-chat sales and upsells. The agent's performance and workload are no longer a source of budget anxiety but a powerful driver of operational leverage. The more work it does, the more valuable it becomes to your business, without ever becoming more expensive. This encourages a virtuous cycle and a mindset of continuous improvement, where teams are constantly looking for new processes to hand over to their tireless AI partner, freeing up human agents for high-value, relationship-building interactions that truly require a human touch. This model allows store owners to treat their AI agent as a true salaried partner in the business, an asset to be maximized rather than a cost to be minimized. Arbyn is designed from its core to not just answer questions but to take decisive, revenue-impacting action. It can autonomously update a customer's shipping address in your backend system, preventing a mis-shipment. With a single click of approval from you for ultimate financial control, it can cancel an order, issue a full or partial refund, or send a discretionary gift card to a frustrated customer. It performs the action; you simply provide the final oversight on sensitive transactions, which is a far more effective and strategic form of governance than a punitive, metered invoice. This is the original promise of agentic AI made real, but without the crippling billing model that discourages its use. By completely removing the per-action fee, the flat-rate model encourages the very behavior that makes AI so powerful: widespread, uninhibited, and creative automation. It aligns the tool's purpose with your own: to solve customer problems as efficiently, completely, and cost-effectively as possible, without fear of a surprise invoice. A Tale of Two Budgets: A 500-Conversation Store The philosophical and financial differences between these two models become starkly, unavoidably clear when translated into a real-world budget. Let's consider a common scenario for a growing Shopify store: 500 customer conversations in a single month. This level of volume is typical for a brand that has found its product-market fit and is beginning to scale its marketing efforts, perhaps after a successful influencer collaboration or during a seasonal sales peak. For this analysis, let's assume the store has implemented a modern AI agent capable of fully and autonomously resolving 60% of these inquiries, which amounts to 300 automated resolutions. This is a realistic, even conservative, automation rate for a well-configured agent handling common ecommerce questions. The remaining 200 conversations require human attention from one or two team members. How does the monthly bill compare between the "governed" per-resolution model and the flat-rate "salaried" model? Under the per-resolution model, the costs stack up quickly and punishingly. The store first needs a base helpdesk plan just to get started. Using Gorgias's public pricing, a store owner might hope to use the Basic plan, but 500 tickets immediately exceeds its 300-ticket limit. This forces a mandatory, and significant, upgrade to the Pro plan at $360 per month just to handle the raw ticket volume, already 6x the cost of the starter plan. On top of this steep base cost, we must add the cost of the 300 AI resolutions. At the documented rate of $0.90 per resolution for a store on an annual plan, this adds another $270. The total monthly cost is not the advertised plan price, but a staggering $630. An alternative like Intercom for a two-person team on its Advanced plan could mean $170 in seats (2 seats x $85/month on an annual plan) plus 300 resolutions at $0.99 each, for a total of $467. These are the punishing, and often surprising, economics of paying by the task. Now, let's analyze the exact same scenario with Arbyn's predictable, flat-rate model. A store with 500 conversations per month fits perfectly into the Arbyn Growth plan at $59/month. For higher volumes, Arbyn Agent offers unlimited conversations for $99/month, while the free Arbyn Starter plan covers up to 150 conversations. For this store, the Growth plan provides a generous allowance of up to 500 AI conversations for a simple, flat fee. The cost is not a complex, multi-layered calculation of base fees, ticket overages, resolution charges, and feature add-ons that requires a spreadsheet and a calculator to decipher. It is a single, predictable number that the store owner can budget for with complete confidence, month after month. This predictability is not just a convenience; it is a powerful strategic asset, allowing the business to scale its customer support capacity and sales without scaling its costs linearly. The table below illustrates the dramatic difference in total cost of ownership for our hypothetical 500-conversation store, showing the deep chasm between a volatile, unpredictable operating expense and a fixed, high-leverage investment. Billing Model Platform / Plan Base Cost AI Usage Cost (300 Resolutions) Total Monthly Cost Governed / Per-Resolution Gorgias (Pro Plan) $360 300 resolutions × $0.90 = $270 $630 Flat-Rate Arbyn (Growth Plan) $59 (includes 500 conversations) $0 $59 The numbers in the final column speak for themselves, and they are not a misprint. The per-resolution model is more than ten times more expensive for the exact same workload and the exact same outcome. This is not an unusual edge case; it is the direct, intended mathematical result of a pricing model designed to meter value rather than to simply deliver it. For a growing business, a difference of over $570 per month is not a trivial rounding error. That's over $6,800 a year, which is the budget for a new marketing channel, a professional product photoshoot, or enough inventory for a brand new SKU. The flat-rate model, in stark contrast, is designed to provide overwhelming value and absolute predictability from day one. For the store owner, this difference is profound. It is the difference between a support tool that is a volatile and rapidly growing expense, and one that is a fixed, predictable, and highly effective investment in operational scale and peace of mind. The choice between a "governed" agent and a flat-rate one is not a simple comparison of features on a checklist. It is a fundamental, strategic decision about how you want to run your business and what kind of relationship you want to have with your tools and your budget. The per-resolution model, championed by incumbent platforms like Aide, Gorgias, and Intercom, offers powerful technology but shackles it with a billing system that creates constant budget uncertainty and encourages you to throttle your own efficiency. It forces you to constantly weigh the cost of an automated action against its value, turning your support team into reluctant accountants. The flat-rate model offers a different, more empowering path. It makes your AI support agent a fully-invested, salaried partner, ready to handle unlimited work for one predictable cost. It removes the meter, eliminates the budget anxiety, and unlocks the true, uninhibited potential of automation. When it is time to hire your AI agent, ask yourself one question: do you want to pay for every task it completes, or give it a salary to work without limits? If you are ready for a true partner, you can install Arbyn free on the Shopify App Store and see the difference a salary makes. --- ## Pricing - **Arbyn Starter** - $0/month, permanently free. 150 conversations / month. Resets 1st of each month. - **Arbyn Growth** - $59/month flat. 500 conversations / month. Resets 1st of each month. Or $600/year (just under two months free, saves $108, 15% off). - **Arbyn Agent** - $99/month flat. Unlimited conversations. Or $990/year (two months free, saves $198, 17% off). - **There is no trial.** Billing starts immediately on any paid plan. The free Arbyn Starter plan is permanent. - The conversation cap is the only difference between plans. There is no feature gating. ## Channels Live today: **support email** and **on-site live chat**. That is the complete list. SMS, Instagram DMs, Facebook Messenger, WhatsApp and Voice are on the roadmap and are NOT live. Arbyn does not edit orders or change line items. Money-moving actions (cancel, refund, discount, gift card, reship, return) require the store owner's approval, and then Arbyn performs them. Running them fully autonomously is a beta authorization and is in development. Shipping address changes are already autonomous. ## What Arbyn does on a Shopify order - **Change the shipping address**: Live. Arbyn does this on its own. Arbyn updates the shipping address on the Shopify order itself, inside the conversation, and writes the change to the order timeline. - **Cancel an order**: Live. You approve it, then Arbyn cancels the order. Anything that moves money waits for the store owner's approval. That is a deliberate control, not a missing feature. Once you approve, Arbyn fires Shopify's order cancellation itself and confirms it to the customer. - **Issue a refund**: Live. You approve it, then Arbyn issues the refund. Arbyn prepares the refund against the original payment method and sends it to you. On approval it files the refund in Shopify. You can cap the value it is allowed to prepare, per channel. - **Apply a discount**: Live. Arbyn creates a real Shopify discount and applies it to the cart, handing the shopper a checkout with the code already on it. It can also issue a discount code on an order once you approve it. - **Send a gift card, or reship an order**: Live. You approve it, then Arbyn does it. Arbyn creates the gift card, or raises the replacement order, in Shopify once you approve. - **Start a return**: Live. You approve it, then Arbyn opens the return. Arbyn opens the return in Shopify on your approval. - **Look up a gift card or store-credit balance**: Live. Arbyn does this on its own. "Do I have store credit left?" is a question most support tools answer with a human. Arbyn reads the balance itself, for a verified customer or from the code they give you, and reports the masked card, the balance and the expiry. If there is no card, it says so rather than guessing. - **Handle a subscription question**: Live. You choose what it does. Arbyn knows which of your products are sold as a subscription, shows that on the product card in the conversation, and sends a subscriber to their subscription management page to pause, skip or cancel. It answers how your subscriptions work from your own knowledge, but it does not read an individual customer's contract, so it will not state their renewal date or status. Most cancels are a customer with product piling up, and the fix is getting them to the page where they can slow the cadence down. Reading the contract itself is on the roadmap. - **Answer support email and live chat**: Live. Arbyn reads every inbound support email and every chat, works out the intent, pulls the live Shopify context, and replies in your brand voice. Money-moving actions (cancel, refund, discount, gift card, reship, return) require the store owner's approval, and then Arbyn performs them. Running them fully autonomously is a beta authorization and is in development. Shipping address changes are already autonomous.