Presto Raises $10M as Drive-Thru Voice AI Enters Its 'Prove It' Era

Presto Automation has closed a $10 million funding round led by Metropolitan Partners Group, with participation from Remus Capital, Link Ventures, and a group of strategic angels that includes the CEO of ElevenLabs. The raise is modest by venture standards but carries outsized significance: it arrives at a moment when drive-thru voice AI is either proving itself or washing out, and the industry is watching closely to see which companies survive the transition from demo to deployment.

The round funds Presto's push to expand from hundreds of active drive-thru locations to thousands by end of 2026. The company currently works with 12 restaurant brands including Taco John's, Fazoli's, and CKE Restaurants, the franchisor behind Carl's Jr. and Hardee's. That list represents a deliberately chosen cross-section of the QSR landscape: regional chains, multi-concept operators, and franchise-heavy systems that each require different rollout playbooks.

The McDonald's Shadow#

Any serious analysis of where drive-thru AI stands in 2026 has to start with the McDonald's/IBM experiment. McDonald's launched its Automated Order Taking (AOT) pilot with IBM in 2021 and expanded it to over 100 locations before pulling the plug in 2023. The stated reason was accuracy issues. Orders were getting garbled. Customers were frustrated. Franchisees were skeptical.

The fallout was considerable, but not fatal to the broader category. It demonstrated something important: large-scale, uncontrolled drive-thru deployments are brutal environments for AI voice systems. Ambient noise, customer accents, complex customization orders, and the sheer pace of peak-hour service create conditions that laboratory benchmarks do not replicate. McDonald's failure was not proof that voice AI doesn't work. It was proof that it doesn't work without the right stack, the right integration, and the right support infrastructure.

Presto's fundraising pitch is built in part on the lessons from that failure. The company has invested in a unified menu management system designed to reduce the friction between AI ordering logic and the actual POS, which has historically been one of the places where drive-thru AI falls apart. A voice system that takes an order correctly but can't relay it to the kitchen accurately is worse than useless.

What the CKE Results Show#

The most concrete data in Presto's current deployment story comes from CKE Restaurants. CKE ran a competitive evaluation across three AI voice vendors: Presto, OpenCity (whose system is marketed as "Tori"), and Valyant AI. The evaluation tracked the metrics that actually matter to operators: upsell offer rate, upsell acceptance rate, order accuracy, and labor efficiency.

The results from the CKE test were striking. Presto's system achieved an 88% upsell offer rate, meaning the AI presented an upsell prompt on nearly nine out of ten transactions. The upsell acceptance rate came in at 46%, which outpaces most human order-takers at comparable price points. Combined with improvements in order accuracy and measurable labor efficiency gains, the numbers were strong enough to advance Presto in the vendor selection process.

These are the kind of results that operators need to see before committing to a multi-location rollout. The upsell metric in particular is meaningful because it directly links AI ordering to revenue per transaction, not just labor cost reduction. Franchisees who have watched the McDonald's cautionary tale unfold are not going to commit capital to a technology that requires sustained troubleshooting without a clear line to payback.

ElevenLabs and the Human Sound Problem#

One of the more interesting details in Presto's current round is the strategic angle from its angel investors. The inclusion of the ElevenLabs CEO signals a specific product philosophy: the voice that greets customers at the drive-thru speaker needs to sound like a person, not a robot.

This is not a trivial problem. Early drive-thru AI systems, including the McDonald's/IBM deployment, were criticized in part because the voice quality was noticeably synthetic. Customers who interact with voice systems that sound uncanny tend to speak differently, enunciate more carefully, repeat themselves, and sometimes abandon the interaction entirely. All of that degrades the metrics.

Presto's partnership with ElevenLabs brings state-of-the-art text-to-speech synthesis into the ordering flow. ElevenLabs has built a reputation for producing voice output that passes as human in most conversational contexts. For a drive-thru deployment where the interaction is standardized, high-frequency, and relatively short, that capability is a genuine competitive differentiator. The strategic investment from ElevenLabs' CEO suggests the relationship goes beyond a vendor contract.

Taco Bell's Parallel Track#

Presto is not operating in isolation. Yum Brands has been running its own voice AI program at Taco Bell, and the numbers from that deployment have become a benchmark for the industry. Taco Bell has processed over 2 million orders through voice AI across more than 300 locations, making it the largest disclosed drive-thru AI deployment in the country.

The Taco Bell deployment is significant because it has survived long enough to generate real operational data. Two million orders across a franchise system is not a pilot. It is a production deployment that has been stress-tested through peak periods, menu changes, and the normal chaos of a high-volume drive-thru. Yum Brands has not published a detailed breakdown of its cost savings or accuracy rates, but the continued expansion of the program is itself a signal that the unit economics are holding.

Taco Bell's experience sets a floor for what operators should expect from a mature voice AI deployment. It also creates a competitive dynamic that benefits the entire category. When Taco Bell franchisees see the program working at comparable locations, it reduces the activation energy required to get other Yum franchisees to pilot similar systems.

The Vendor Shakeout Is Underway#

Drive-thru voice AI is no longer early-stage. Several companies that were active in the space three years ago have either shut down or pivoted. The capital environment for early-stage AI has tightened, and restaurant operators are no longer willing to be beta testers for systems that require six months of hand-holding before they perform reliably.

What's left is a smaller set of vendors who have working deployments, documented results, and the operational infrastructure to support multi-location rollouts. Presto is in that group. So is OpenCity. SoundHound, which has been aggressive in the drive-thru space through its acquisition of SYNQ3 and partnerships with multiple QSR chains, is another survivor. Valyant AI, which competed in the CKE evaluation, has shown staying power in the regional chain segment.

The shakeout is not complete. Presto's $10 million raise is not a war chest by any measure. Metropolitan Partners Group is not a tier-one venture firm. The round will fund operations and sales expansion, not a massive technology overhaul or a land-grab hiring spree. That constraint shapes the deployment strategy: Presto needs to convert its existing brand relationships into full-chain rollouts, not sign new logos at the expense of quality execution with current partners.

The Operator Decision Framework#

For QSR operators evaluating drive-thru voice AI in 2026, the market has clarified considerably. The early question was whether the technology worked at all. That question is largely settled. The current question is which vendors have the deployment playbooks, the integration support, and the operating history to actually get a system live across dozens or hundreds of locations without a disaster.

Operators running the due diligence process should be asking vendors for references from multi-location deployments, not pilots. They should be asking for order accuracy rates under real-world conditions, not lab benchmarks. They should be asking about the menu management workflow and how updates propagate to AI ordering logic when LTO items change or regional pricing varies.

The financial case also needs to clear a realistic bar. Drive-thru labor typically represents 25-30% of total restaurant labor, and a voice AI system that handles 60-70% of drive-thru ordering volume without human intervention can produce meaningful cost reduction. But the math only works if the system reduces errors, not adds them. A voice AI that generates a 5% increase in order errors will cost more in remakes and customer churn than it saves in labor.

Presto's CKE results, if they hold across a broader deployment, suggest the error rate problem is solvable. The 88% upsell offer rate and 46% acceptance rate suggest the revenue upside is real. The remaining question is execution at scale, which is precisely what the $10 million is intended to fund.

What 'Thousands of Locations' Actually Requires#

Presto's stated goal is to reach thousands of drive-thru locations by end of 2026. That is an ambitious target from a base of hundreds. Achieving it requires more than technology; it requires a franchise sales motion, installation infrastructure, crew training materials, and a tier-one customer success function that can handle the inevitable issues when they arise at 2 a.m. during a Friday rush.

The companies that have succeeded in QSR technology deployment at scale, whether point-of-sale systems, loyalty platforms, or kitchen display systems, have all discovered that the hardest part is not the software. It is the last-mile execution: getting systems installed correctly, crews trained adequately, and franchisees supported through the early operational friction that accompanies any technology change.

Presto's partnership network gives it some infrastructure here. Its existing brand relationships mean it has worked through at least some of the integration challenges with major POS systems. The unified menu management system is an attempt to reduce the ongoing maintenance burden that has burned other AI vendors when restaurants update their menus. Whether that is enough to support a 5-to-10x location expansion in nine months will be the real test.

The voice AI shakeout of 2025-2026 will ultimately produce two or three scaled winners and a long tail of casualties. Presto has the technology, the brand relationships, and now the capital to be in the winner's group. The question is whether it has the operational muscle to get there before the window closes.

---

QSR Pro covers technology, operations, and finance across the quick service restaurant industry. Subscribe for weekly intelligence briefings.