Solidroad raises $25 million to move support quality control to AI

Solidroad is going after a weak spot in customer support. Instead of manually spot-checking a handful of calls and chats, the company offers AI that scores every interaction and shows right away where a team is losing quality, money and customer loyalty. The startup, with offices in Dublin and San Francisco, raised $25 million in a Series A round led by Hedosophia. It previously received $6.5 million in seed funding from First Round Capital with participation from Y Combinator, bringing total funding to $31.5 million.

Solidroad was founded in 2023 by former Intercom employees Mark Hughes and Patrick Finlay. The company has about 20 employees, and its customers include Ryanair, Crypto.com and Oura, as well as the large outsourced contact centers PartnerHero and Tech Mahindra. It also went through the Y Combinator Winter 2025 batch, which helped it reach enterprise customers in the US faster.

The problem Solidroad is betting on has long been familiar to the market. Most support teams manually review the quality of only 1–3% of interactions: a manager listens to a few calls, reads some of the conversations and scores them against an internal checklist. The process is slow, expensive and says almost nothing about the other 97–99% of interactions.

Solidroad automates this layer of oversight. The platform analyzes calls, chats and email threads, scores them against a company's own criteria, identifies recurring mistakes, suggests who needs training and on what, and builds a complete picture of the team's quality.

Solidroad does not sell the idea of replacing support staff with bots entirely. Its position is closer to this: people will remain in support, but they need a more precise tool for feedback and training. That sets the company apart from some AI startups that bet specifically on autonomous front-line agents. For enterprise customers this is an important argument: adopting AI for quality control is usually easier to get approved than replacing agents.

Another plus is enterprise-sales readiness. Solidroad has already obtained SOC 2 and ISO 27001 certifications, which for a young company of twenty people looks like a deliberate bet on large customers and long contracts.

According to the company, after the platform was deployed at Crypto.com, the customer satisfaction score at launch rose by three percentage points and exceeded 90%. At PartnerHero and Tech Mahindra, onboarding time for new agents fell by 50%.

In the contact-center industry, where margins are usually thin and turnover is high, metrics like these matter more than any general talk about AI. If the system really does help get employees up to speed faster and reduce the number of poor-quality contacts, its value is easy to calculate in money. Against that backdrop, the idea of checking 100% of interactions instead of a standard sample looks less like polished marketing and more like an attempt to turn QA from a formal procedure into a working management tool.

The market around Solidroad has matured, too. AI already handles a significant share of customer inquiries, and as automation grows, it matters not only to answer the user but to understand how well a human or a bot is doing it. Large platforms such as NICE, Verint and Genesys already operate in this space, along with specialized players such as MaestroQA and Klaus, which Zendesk previously acquired. Solidroad's bet is that AI-native QA, built from the start to analyze all conversations and tied directly to team training, can carve out its own niche before similar features are finally built into big contact-center suites.

Against the backdrop of the huge AI rounds of recent quarters, $25 million looks modest, but for a clear B2B segment with measurable ROI it may be enough to grow sales quickly. For the market, it is a signal that the next notable layer of AI in support is not only chatbots but also the quality-control infrastructure around them. If Solidroad can prove sustained ROI at high interaction volumes, investors will get another example of how a narrow, well-understood B2B process becomes a standalone category of AI software.