Repsmate este un startup tech multipremiat care ajută operatorii din call center (Vânzări, Customer Support) să analizeze interacțiunile cu clienții într-un mod detaliat generând rapoarte personalizate pentru fiecare client.
€381.300 din €100.000 finanțată
RepsMate is a multi-award-winning Romanian Tech Start-up that integrates AI & Data Analytics technologies to understand customers’ needs and behavior, in order to assist speech-based interactions. We aim to have a global presence by 2024!
RepsMate’s mission is to unlock business value for its clients by enhancing the interactions between them and sales and support representatives. We are developing a real-time virtual coach for agents that will be available to help increase the quality of interactions, sales performance, customer satisfaction, improve retention and reduce the overall operational costs.
“32% of customers will leave a brand that they love after one bad experience” – according to PWC
The main challenges that Organizations with Sales and Customer Support are facing can be classified into four classes:
1. For more than 40 years, quality assurance (QA) has been a foundational part of enhancing agent performance, ensuring compliance, and improving the customer experience. QA procedures suffer from inefficiencies across entire organizations:
· only 1-2% of conversations are analyzed;
· the process is entirely manual, meaning, evaluators use multiple systems and spreadsheets just to quality check a single call;
· the analysis is subjective up to the evaluator;
· the training teams cannot effectively tailor coaching programs relevant to individual agents because they cannot make data-driven decisions.
2. Destructive Agents Attrition Rate. When a company agent leaves the company, it becomes very costly to replace, the cost comprising recruiting fees and onboarding costs but also the potential lost revenue from missed deals. Recent studies indicate that onboarding a new agent cost approximately three times his/her salary. The agents typically hit top performance after their first year and their peak quota attainment typically occurs between years 2 and 3. The average sales agent tenure was 1.5 years in 2018, down from 3 years in 2010.
3. Lengthy and Expensive Onboarding/Training Processes - Today’s products and services are more and more complex and acquiring a good knowledge about their functionalities and ways to operate requires time. The average training time of new-hire agents can go from 2-3 weeks to more than 6. Even more, the sales training courses cannot cover the whole list of scenarios that agents could face. Studies showed that up to 87% of sales training content is forgotten within one month of the training and only 24% of the agents are top performers.
4. Low Customer Retention and Satisfaction conducting to behaviors such as consumers deciding to stop transacting with a business after a poor service experience. Using surveys in which only a small percentage of customers respond, they are also making a concerted effort to find out their customers' opinions and needs. However, they believe the customers who do not answer are happy, not that they will never use their service or products again.
The problems listed above are caused by the agents' issues such as call reluctances syndrome, the need for constant training, and the appreciation or even the motivation that the supervisor/manager should spread across the team. No systematic solution to address these issues directly exists. The advancements in Machine Learning and Artificial Intelligence technologies, in general, are game-changers in this field and RepsMate exploits them to provide better tools that are able to adapt to particular needs.
Q: Why do organizations ignore my opinion as a customer and I always feel like they don't care?
A: It is always their goal to improve their business processes, but they lack the tools to hear you!
Q: Why does my manager always pick the worst calls for the feedback sessions, and never trust me when I know what to do?
A: Despite being your ally, your manager lacks the right tools to find your weaknesses and help you develop!
Q: Why does my team not hit targets and they always leave?
A: While they are always striving to improve themselves and hit their targets, they lack the tools that can assist them in doing so.
Q: What causes customers to be unsatisfied and complain all the time? What is the reason for our customer-facing representatives' low attrition rate? Do we need to keep spending money on hiring and onboarding new employees permanently?
A: Because your organization lacks the proper tools to permanently improve your customer-facing representatives, they perceive their job as being harder than it actually is, so they are leaving. Having only employees with little or no experience means that your customers will always think that you don't care about them.
RepsMate aims to address the human communication limitations caused by the lack of information about the interlocutor, by developing the necessary technologies and infrastructure, using Machine Learning and Artificial Intelligence for Speech Analysis, Behavior & Emotion Segmentation, and Benchmarking & Coaching in order to deliver the best customer interaction with minimum effort and cost. Instead of just providing all the insights gathered from the interactions to the managers as our competitors do, we deliver them directly to the agents through answers and action suggestions.
RepsMate stakeholders and their benefits are:
- The Call Centre Agents will permanently increase their performance by auto-developing themselves because they will always have a trustworthy coach that will assist them in every possible situation during and after the conversations. RepsMate will also eliminate agents' reluctance by providing them with the best possible answers to help them to attain their quota.
- The Call Centre Managers will gain instant visibility across 100% of representatives’ conversations and will make informed decisions regarding customer needs.
- The Call Centre will reduce the overall cost by decreasing KPIs such as Average Handling Time, First Call Resolution, Agent Attrition Rate and will increase revenue by increasing customer satisfaction and agents' performance.
- RepsMate is developing a SaaS solution that integrates AI & Data Analytics technologies in order to understand customers’ needs and behavior, to assist companies’ agents (representatives) in their communications with customers. RepsMate aims to provide companies’ representatives with automated and tailored guidance, simulating a real-time, always-available virtual coach for agents. By using the RepsMate platform the call center managers are gaining instant visibility across 100% of representatives’ conversations and are able to make informed decisions regarding the conversations.
All those modules expose robust API interfaces and the communication is done using HTTP and GRPC communications frameworks enabling RepsMate to bring intelligence to business by leveraging the power of artificial intelligence.
We have a B2B and B2E SaaS business model, where we charge our clients, a monthly fee based on the number of agents that we process conversations – €110 cloud / €85 on-premise – price list per agent per month.
Other revenue streams:
- Initial Setup
- AI model customization
- Business & Technical Support for customized features
Business highlights & achievements
-> Business Accelerators:
· Advancing AI by Techcelerator
· Special Startup Award by Romanian Contact Center Awards 2021
· 2nd Place @X-Awards by Startup Grind Bucharest
· 1st Place @Demo Day by Commons Accel
· Pre Seed Investment Readiness @Demo Day by BeAI 2020
· Advancing AI finalist @Demo Day by Techcelerator
· Google for Startups
· 3rd Place @Spotlight by How to Web
· Microsoft for Startups
-> 1 Patent-pending application at EPO & OSIM
Investment Instrument: Agreement for Future Equity Acquisition
• Current round (Bridge): 100.000 – 350.000 Euro
• Lead investors: Cleverage VC and 10 other Angel Investors
• Valuation Cap (pre-money): 3.500.000 Euro
• Discount: 30%
• Latest date for next liquidity event: 30.09.2023
• All the relevant terms and conditions are detailed in the Key Investor Information Sheet
- 3 paying clients - €50,000 ARR – from 150 agents, scalable up to €180,000 ARR
- 5 corporate clients in the implementation process;
- other 20 potential clients, showed interest in using RepsMate after the demo meetings, and they are in different implementation stages;
- Strategic Partnership with NexTip for Romanian and Italian Market;
- €4,200 Monthly Recurring Revenue
Risk 1 – Technological
Security Data Breach
Impact: high. Probability: low
Mitigation Measures: RepsMate is using top Cloud providers, that have high-security protection measures and there is a Low to None risk of any Security Data Breach. As an extra layer of protection, we are using Firewalls, VPNs, database encryption, strong password management policies, and independent Penetration Testing firms.
Risk 2 – Technological
Impact: medium. Probability: low
Mitigation Measures: RepsMate will use API integrations, to be able to able to send or receive information from other platforms. Also, the customers are able to visualize the data in the RepsMate platform or in other platforms (ex: CRMs, VoIP Platforms) as well.
Risk 3 – Technological
Low accuracy on other AI models
Impact: high. Probability: low
Mitigation Measures: RepsMate is testing different state-of-the-art AI models before choosing which is the most suitable for a specific task. After that, we start to train an initial version of the model until we get aimed results. In order to receive high accuracy results of the AI models, we are fine-tuning each model free of charge for our customers.
Risk 4 – Commercial
The low customer adoption rate
Impact: high. Probability: low
Mitigation Measures: All potential customers that we've interacted with, were amazed about what RepsMate is capable to do. We can also pivot the technology developed in other niches or markets.
Risk 5 – Commercial
Impact: medium. Probability: high
Repsmate aims to continuously increase its features based on customers’ feedback, and we aim to be one step ahead of the competition. We will also apply for trademarks and patents for our proprietary models and processes.
Risk 6 – Commercial
Impact: medium. Probability: low
Mitigation Measures: RepsMate is protecting its organization's competitive advantages through patents, trademarks, NDAs, and Contracts.
Risk 7 – Financial
Lack of Financial Resources
Impact: high. Probability: medium
Mitigation Measures: RepsMate is already developing relationships with different Investment Funds and Private Investors. We see the EIC support as an important lever to de-risk our future A round of financing.
Risk 8 – Financial
Lack of Scaling Opportunities
Impact: high. Probability: low
Mitigation Measures: RepsMate will approach strategic clients that have multiple offices in different countries to systematically enter new markets and industries in order to guarantee strategic and organic company growth.
Risk 9 – Regulatory
Major changes in the data privacy legislation
Impact: medium. Probability: medium
Mitigation: RepsMate acts only as a ‘data processor’ and we enable our customers to record, transcribe, analyze and share the contents of their communications. Our customers act as a ‘data controller’ and is their responsibility how they use our platform. To be compliant with GDPR Regulation, we take external consulting for each country in which we operate and we are developing a framework that irreversibly anonymizes all communications.
1. What RepsMate does?
We analyze 100% of interactions between agents and customers in near real-time to identify areas where agents need to improve to increase their efficiency (ie. reducing the duration of the interactions, the repeated calls of a customer – resulting also in decreasing also the cost of the operations) in order tao increase customer satisfaction.
2. What type of interaction do you analyze?
Currently we are analyzing voice interactions, but are in the process of becoming an omni-channel platform.
3. Do we use third party API providers?
We don’t use 3rd party API providers; all the engines and features are built in-house in order to be able to provide a competitive pricing model with a high accuracy.
4. What is the Speech Recognition accuracy?
Firstly, we want to highlight that is a big difference between Contact Center audio conversations and YouTube/Podcast recordings.
The accuracy that we have achieved it so far is between 95%-97%, as our Speech Recognition Engine is having a complex infrastructure and the Speech-to-Text model being just a small part of it.
Some key aspects which are particular to call-center conversations are:
- Voice intonation, voice accents, filler sounds (such as ‘ah’ or ‘uhm’), stops or pauses in mid-sentence. These are common in telephone conversations but rare in studio recordings;
- Also, another aspect which differ from publicly available recorded audio transcripts is the noise induced by audio transmission and audio compression algorithms, such as VoIP telephony specific compression algorithms and the noise of the recording equipment;
- Background noise, background chatter which are specific to call-center conversations, such as when the rep speaks from an open-space environment with colleagues engaged in other conversations or when the customer is outside in a crowded space. Most public ASR datasets are recorded inside a relatively isolated room; therefore, they would not capture this type of noise which is common in call-center conversations;
- Conversation specific language, consisting of small phrases which are highly dependent on the conversation context. Most language models are trained on very large datasets consisting of publicly available text on the internet in a specific language, mostly captured by web-crawlers’ algorithms, such as Wikipedia articles, newsletters, transcript speeches (from parliament or news-related), product reviews etc. These domains (news, articles, public speeches, interviews) have a specific language and may not capture the dynamics and contextual information of a conversation. Therefore, labeled telephone conversations are a key factor in improving ASR’s language model performance, that’s why a labeling team focused on domain-specific data is key to an ASR system’s performance.
5. How long takes to train new languages?
Because we automate the whole process, it takes less than 2 months to build a new speech recognition engine for a new language from the scratch.
6. How you integrated with different VoIP solutions?
RepsMate is built on-top of the VoIP provider systems not to overwhelm the IT departments of our clients.
7. On-premise / Cloud?
Because we don’t use API providers, RepsMate platform is infrastructure agnostic and can be deployed on-premise or in Cloud.
8. Who are your clients?
We are targeting customer service departments, such as telesales, customer support or customer service departments, from:
- BPOs (business process outsourcing) – Contact Centers
- Organizations that have internally such departments
9. Who are your users?
Currently our platform is used by managers, supervisors and team leads and in the near future we will onboard representatives/agents.
10. What is our business model?
We have a B2B and B2E SaaS business model, where we charge our clients a monthly fee based on the number of agents from which we process conversations.
11. What clients qualify as prospects?
Currently we are targeting potential clients with a minimum annual contract value of 50,000 EUR per year.
12. How does the start of the sales process look like?
We are engaging potential customers through warm intros, events and LinkedIn outreach, by sending a short presentation about how RepsMate works and its benefits.
13. How does the development of the sales process look like?
We are engaging further interested potential customers through demos of RepsMate platform – we install the product for a 30-days pilot (on average), we personalize, train and configure the software, and the client can test the benefits that RespMate can provide. As a next step, we finalize the offer according to the customer requirements (no. of agents, use cases), quote and discuss all the technical, legal and commercial conditions that will apply as needed.