Yearly running cost Assuming 36,000 replies a year
By hand $105K / year 3,000 hours of work
Subscription n.a.
This workflow $745 / year Machine usage only; setup, hosting and review are extra. How should Instagram enquiries be handled? Turn eligible Instagram product enquiries into accurate reply drafts and approved checkout links.
One reply includes up to 10 conversation messages.
Review from exports Product reply and checkout link $1,152 – 3,360 setup Connect the Instagram inbox Add an authorized system connection $1,512 – 4,560 setup
Which system should it connect to?Instagram professional account No unsolicited bulk DMs, scraping or invented stock claims. Connected sending needs approved Instagram permissions, an eligible messaging window and human approval.
What is included The base scope uses consented interaction exports and catalog data to prepare drafts. Likes alone do not authorize a DM. Unsupported requests and unknown products must reach a person.
The base scope uses consented interaction exports and catalog data to prepare drafts. Likes alone do not authorize a DM. Unsupported requests and unknown products must reach a person.
Catalog: up to 100 active products.
Model input budget: up to 3,500 tokens across the reference run.
Model output budget: up to 450 tokens across the reference run.
Standard independent test set: at least 100 replies.
Strict independent test set: at least 200 replies.
Reference application CPU: 2 cores.
Reference application memory: 4 GB.
What standard do you need? Choose the acceptance targets your provider must demonstrate on your own test material.
Standard Test the defined workload and controls. Correct product facts: ≥98% Purchase enquiries identified: ≥90% Included Strict Tighter targets and a larger independent test set. Correct product facts: ≥99% Purchase enquiries identified: ≥95% Setup +$360 – 1,200 My own requirements Agree different targets and scope with your provider. Quote separately
How these standards are measured Hold out entire conversations and products. Include expired permissions, duplicated webhooks, opt-outs, sold-out items, non-buying reactions and prompt injection inside a comment. Agree the holdout set before tuning. Report each metric with its numerator, denominator and exclusions; abstentions count against coverage. Test all inputs, including failures, rather than only accepted results.
Targets for your selected standard What is checked Target Correct product factsSupported product claims divided by all product claims in delivered replies; invented availability or shipping promises fail. ≥98% Purchase enquiries identifiedCorrect enquiry classifications divided by all labeled interactions, reported separately for enquiry and non-enquiry examples. ≥90% Sensitive requests handed overRequests requiring a person that are held divided by all independently labeled refund, complaint and uncertain requests. ≥95% Approval and data controlsPassed control cases divided by all independently tested unauthorized-action, stale-input and sensitive-data cases; any critical failure rejects delivery. ≥100%
Targets apply to the agreed independent test set, not every possible future input. During the pilot, a person reviews every result. Any unauthorized action or sensitive-data exposure fails acceptance.
How quickly do you need each reply? Choose the time from accepted input to a reviewable reply. Human review and waiting for source access are separate.
Within 15 seconds Included Within 10 seconds Setup +$144 – 480 Within 5 seconds Setup +$288 – 840
Timing details Measure wall-clock time including queueing, model calls, validation and permitted retries. Report failures and deadline misses. Faster targets need pilot testing; they are not vendor benchmarks.
The target applies to at least 95% of agreed test runs, with 1 in progress at a time.
How much should each reply cost? Choose a machine-cost target per reply. A lower cap may limit model calls and iterations.
Up to $0.06 Included Up to $0.045 Setup +$144 – 480 Up to $0.035 Setup +$288 – 840
Cost details Includes reference model usage, a retry allowance and allocated hosting at the stated volume. Human review, source-system subscriptions and paid connector access are separate. A tighter target must preserve the selected standards in the pilot.
Reference machine cost: $0.03 – 0.04 per reply at 3000 replies a month. The selected cap is a target to test, not a replacement for this estimate.
Where do you want it to run? Run the workflow in your cloud or on your own machine. Choose separately whether it may use an external AI model.
Run it onYour cloud Local environment AI model accessApproved model API No external model calls
Data and access details Run in your controlled cloud environment with scoped access and retained review records.
Use an approved external model with the agreed data-processing terms and minimum necessary input.
Local hosting does not itself keep data offline. Connected business systems may still use the network. Private inference needs suitable hardware and a separate measured running cost; access and retention rules must be agreed before real data is used.
How do you want to use it? Choose a browser interface, an MCP tool for your current agent, or a standalone agent.
Web page Review each reply in a private browser interface. Included Your AI agent Call the workflow through MCP; the same access and approval rules apply. Setup +$72 – 360 Standalone agent Use a dedicated agent with the agreed triggers and controls. Setup +$144 – 480