TrainThis Gives Sales Reps Real-Time Feedback on Every Call

01 /The Concept

Feedback is the bottleneck in almost every training system – not content, not instruction, but the loop that tells someone whether what they just did was right. TrainThis is an early-stage startup, barely months old and still in pilot deployments, that's trying to automate that loop for customer-facing communication.

The platform integrates with a company's email, voice, and video conferencing systems and analyzes every client interaction – calls, meetings, messages – in real time. Employees receive immediate feedback on each conversation: not just a critique of what went wrong, but explicit acknowledgment of what they did well. That last part matters more than it might seem. Research consistently shows that effective feedback requires roughly six positive signals for every critical one. Training systems that only flag errors tend to demoralize employees before improving them.

Managers see aggregate results on a dashboard: who is improving, who needs support, where specific skills are falling short. The feedback loop also serves as a diagnostic tool for training priorities rather than just individual performance review.

The platform's most distinctive feature is its fine-tuning mechanism. Companies can customize the AI's evaluation framework in two ways: by uploading examples of high-quality communication they consider exemplary, and by explicitly defining the communication rules that matter to them. Crucially, this fine-tuning applies not just to what the AI evaluates, but to how it delivers feedback – tone, phrasing, even stylistic register can be adjusted to match company culture.

As the LinkedIn post from the founders puts it: if an employee performs a communication action 50 times a week and improves by even 0.1% each time from targeted feedback, their communication quality nearly doubles by the end of a quarter.

TrainThis launched in January and recently closed a £200,000 pre-seed round – roughly $270,000 – despite still running pilot engagements without a public product.

02 /Why It Matters

Three startups are currently working variations of the same underlying technology – AI analysis of sales and customer communication – but packaging it as three different products.

Oliv, [covered in a spring review](/review/on-vyrastet-eshhjo-bolshe), leads with "clone your best salespeople": capture what high performers do and replicate it across the team. Zenarate, [reviewed in summer](/review/bylo-7-7-milliardov-dollarov-budet-eshhjo-bolshe), built a simulation environment where trainees practice against AI-generated customer scenarios before touching live calls – it raised $18 million. TrainThis positions on real-time feedback during actual work rather than pre-deployment simulation.

Same technology, three distinct product frames, three different buying contexts. The lesson is worth naming clearly: the product is not what you build, it's what you sell. Buyers evaluate what they're offered, not the underlying mechanics. Leading with "real-time feedback" versus "best-rep cloning" versus "pre-hire simulation" routes each product to a different procurement conversation – training and development, sales ops, or HR – with different budgets and different success metrics.

TrainThis's fine-tuning loop has a structural advantage worth noting: it creates lock-in through organizational specificity. Each time a company uploads an example or adds a communication rule, the platform becomes better calibrated to that company's specific context. After six months of fine-tuning, the evaluation model reflects nuances no out-of-the-box AI would know. That institutional memory is hard to transfer to a competitor.

Upduo, [previously reviewed](/review/uchit-mnogo-i-bystro-luchshe-tak), built peer-to-peer sales coaching – pairing experienced reps with newer ones to create structured learning without scaling the trainer headcount – and earned strong endorsements from clients including Motorola before raising $4 million. TrainThis takes that peer learning model one step further: when a star performer's examples train the AI, that knowledge propagates to every employee simultaneously rather than one pairing at a time.

03 /Opportunities

Feedback – not content, not curriculum – is what actually drives learning. You can find information on virtually any topic for free. People enroll in courses not for the material but for the corrective signal that tells them whether they understood it correctly. That signal is expensive to produce at scale: a lecture reaches fifty students in an hour, but meaningful individual feedback on fifty assignments takes fifty hours.

The AI unlock is making feedback economically scalable. TrainThis applies this to communication skills in a professional context. The same logic works in online education, corporate onboarding, skills credentialing, and professional certification – any domain where the current bottleneck is reviewer time rather than learner access to material.

The market implications are large: eventually, no education or training platform will be able to skip an automated feedback layer and stay competitive. The question is which specific context to enter first. Professional communication has the advantage of being high-frequency and high-stakes, which creates both rich training data and clear ROI metrics. But the same architecture could apply to code review feedback, writing coaching, or management skill development.

The strongest entry angle isn't the most obvious one – large enterprise sales training – but rather the contexts where feedback quality is most demonstrably broken today and where AI calibration to company-specific standards is the primary differentiator.