Short answer: Customer Service companies paying per-query cloud AI fees can eliminate that variable cost by moving inference on-device — the model runs on the user's hardware, not yours. Wednesday scopes and ships this in 4–6 weeks.
Your AI customer service features — intent classification, FAQ deflection, and ticket routing — make 4 million API calls per month. At $0.002 per call, that's $8K per month for features that a smaller on-device model can handle at near-zero marginal cost.
That $8K scales with your support volume. Every new product launch or seasonal event that drives tickets also drives the AI bill.
The Four Decisions That Determine Whether This Works
Migration priority by task type. Intent classification is the highest-value migration target — it's high volume, low complexity, and tolerates a smaller model well. FAQ deflection is second. Ticket routing is third. Starting with intent classification cuts the most API calls in the first sprint and delivers measurable cost reduction before the project is half done.
Accuracy floor for service quality. A customer service model that misclassifies intent 8% of the time instead of 3% sends 5 in 100 customers to the wrong resolution path. Your customer service team needs to agree on the acceptable misclassification rate before the model is selected — because that rate directly affects customer satisfaction scores and escalation volume.
On-device vs on-premise. If your customers use the app on their own devices, on-device processing makes sense. If your customer service agents use the app on company-managed devices, an on-premise server inside your network may be faster and more controllable. The deployment model changes the architecture, the cost structure, and the maintenance obligation.
Escalation logic. On-device AI that handles routine requests needs a reliable escalation path to a human agent when it reaches its confidence threshold. The escalation architecture has to be designed before the model ships, or customers with non-routine issues get stuck in an AI loop — which is worse for satisfaction scores than no AI at all.
Most teams spend 4-6 months discovering these decisions by building the wrong version first. A team that has shipped this before compresses that to 1 week.
React Native vs. Native vs. Hybrid: When to Use Each
| Factor | React Native | Native iOS + Android | Hybrid (WebView) |
|---|---|---|---|
| Code sharing | ~85% shared codebase | 0% — two separate codebases | 95%+ shared |
| Performance | Near-native for most interactions | Best possible | Noticeably slower |
| Development speed | 40–60% faster than native | Slowest | Fastest |
| Platform API access | Full, via native modules | Full | Limited |
| Team required | JavaScript/TypeScript engineers | iOS (Swift) + Android (Kotlin) specialists | Web engineers |
| Best for | Feature-rich apps, marketplaces, rapid iteration | Performance-critical apps, deep OS integration | Simple tools, prototypes |
For most product apps — marketplaces, fintech, edtech, consumer — React Native is the right default. Wednesday has shipped it at 500,000-user scale.
Why We Can Say That
We built Off Grid because we hit every one of these problems in production. Off Grid is the fastest-growing on-device AI application in the world, with 50,000+ users running it today.
It's open source, with 1,650+ stars on GitHub and contributors from across the world. It has been cited in peer-reviewed clinical research on offline mobile edge AI.
Every decision named above — model choice, platform, server boundary, compliance posture — we have made before, at scale, for real deployments.
How the Engagement Works
The engagement is four sprints. Each sprint is fixed-price. Each sprint has a named deliverable your team can put on a roadmap.
Discovery (Week 1, $5K): We resolve the four decisions — model, platform, server boundary, compliance posture. Deliverable: a 1-page architecture doc your CTO can take to the board and your Privacy Officer can take to Legal.
Integration (Weeks 2-3, $5K-$10K): We ship the on-device model into your app behind a feature flag. Deliverable: a working build your QA team can test against real workflows.
Optimization (Weeks 4-5, $5K-$10K): We hit the performance and compliance targets from the discovery doc. Deliverable: benchmarks signed off by your team.
Production hardening (Week 6, $5K): Edge cases, OS version coverage, app store and compliance review readiness. Deliverable: shippable build.
4-6 weeks total. $20K-$30K total.
Money back if we don't hit the benchmarks. We have not had to refund.
"They delivered the project within a short period of time and met all our expectations. They've developed a deep sense of caring and curiosity within the team." — Arpit Bansal, Co-Founder & CEO, Cohesyve
Ready to See the Numbers for Your App?
Worth 30 minutes? We'll walk you through what your current inference spend and usage volume mean for the business case, and what a realistic cost reduction target looks like.
You'll leave with enough to run a planning meeting next week. No pitch deck.
If we're not the right team, we'll tell you who is.
Book a call with the Wednesday team
Frequently Asked Questions
Q: How much can a customer service company save by moving AI on-device?
At 1M queries/month, a $0.002/query cloud API costs $2,000/month. On-device costs $0 per query after integration. At 10M queries/month: $20,000/month saved. Break-even on a $20K–$30K integration is typically 1–3 months.
Q: What's the quality trade-off between on-device and cloud AI?
For structured tasks — classification, extraction, form completion, search ranking — a 2B–7B on-device model performs comparably to cloud. For open-ended generation or broad world knowledge, cloud models have an advantage. The discovery sprint benchmarks your specific tasks against on-device candidates before committing.
Q: How long does a cloud-to-on-device migration take for customer service?
4–6 weeks. Week 1 identifies which tasks move on-device and defines quality benchmarks the on-device model must meet.
Q: What does a cloud-to-on-device AI migration cost?
$20K–$30K across four fixed-price sprints, money back if benchmarks aren't met. Typically recovered within 1–3 months of reduced API spend.
Q: What happens to AI quality when moving from GPT-4 to on-device?
Structured tasks often match cloud quality with a well-tuned 2B–7B model. Tasks requiring reasoning over long context or broad factual knowledge will show degradation. The discovery sprint benchmarks your specific tasks before any migration is committed.
Top comments (0)