Questions
The questions you are probably asking too.
Short answers. If yours is not here, ask it in the 30 minutes.
What it isWhat exactly is TheAICalls: a calling tool, analytics, or a chatbot?
One platform for customer communication. It analyses every call and email, not a sample, and turns it into coaching and management data. On top of that you can call through it and automate emails and tickets. Most clients start with analytics on data they already have, so nothing changes for the team on day one.
What it isDo we have to switch our phone system or helpdesk?
No. Analytics runs on the recordings and emails you already have. Moving calls or tickets into our platform is optional and comes later, only if it makes sense.
What it isDoes it integrate with our phone system, CRM or helpdesk?
Yes. We can build any integration, and several large tools are already integrated, such as 3CX, Zoho and Daktela. For most setups it is a small piece of extra work. Tell us what you run and we'll say exactly what it takes.
What it isOur phone system already has speech analytics. Why would we need you?
Built-in analytics usually transcribes, tags and counts. We score every call against your own criteria, turn it into specific coaching per person with the evidence attached, and cover email in the same view. The easiest way to judge is to put both side by side on the same calls.
What it isWhich languages do you support?
All of them. Live today we run across European languages, the US and the Middle East, in 23+ countries. There is no language the model has trouble with; the difference between languages is a few percent in transcription at most.
What it isHow accurate is it?
Transcription is typically around 97 to 98% accurate. When a word is missed, the model still understands it from context, so it makes no difference to the analysis. The scoring criteria are yours, and in the first step your QA can score a sample of calls, we score the same calls, and you compare. Worth knowing: two human reviewers also often disagree on the same call.
Getting startedWhat does the first step look like, and how long does it take?
The first step is Insight: you give us calls and emails, ideally 6 to 24 months, we set the platform up on your data and analyse all of it: what is in it, how each person performs, what could be automated, what it would take and what it would save. Your operation does not change. From data to results it takes 2 to 4 weeks if you already have your scoring criteria and know what you want to track, 4 to 6 weeks if we build them together.
Getting startedWhat do you need from us?
Two things. First the data: an export of recordings and/or emails with basic information (date, agent, direction). Second, exactly what you want to evaluate and track, in other words how your managers assess the quality of your agents today. That has to be written down somewhere; if it isn't, we put it on paper together, and then we apply it in the platform for the AI. Plus one contact person. Later, read access to your CRM or helpdesk; for automation, access to the order or customer system.
Data and securityWhere is our data stored? Is it GDPR compliant?
We process your data as your processor under a data processing agreement, hosted in the EU or the US as you choose. If you need it on your own infrastructure, it can be done: you give us the requirements, we build the right setup, and it has its own deployment cost. The best option for most clients is our managed cloud, which we run and maintain. Your legal team gets the DPA and the list of subprocessors before any data moves.
Data and securityDo you train AI models on our data?
No. We do not train AI models on your data.
Data and securityIs analysing call recordings legal? Do we need extra consent?
It depends on the country, and the rules are the same as for recording today. If your legislation requires you to inform callers, you inform them; that stays with you as the data controller. We analyse what you already record, as your processor.
Data and securityWho can see what?
You set the roles. Managers see their teams. In Coach every agent sees their own calls, scores and coaching, not colleagues'.
Data and securityDo you have ISO 27001 or SOC 2?
Not today. Send us your security questionnaire: we fill it in and walk your team through how we handle hosting, access and encryption.
PeopleWon't our agents feel watched?
They are already assessed today, on a few calls someone happened to pick. This makes it fairer: the same criteria for everyone, on every call. In Coach agents see their own results and exactly what to improve. Introduce it as coaching, not control, and it lands well.
PeopleWhat about the works council? (Germany, Austria)
In Germany and Austria the works council has a say in any system that can evaluate employee performance. We provide what they need: what is scored, who sees what, how long it is kept. And the key point for them: a score is a recommendation. A manager decides what to do with it; it is not tied automatically to pay or any other consequence. Involve them early and it goes faster.
PeopleIs this going to replace our people?
Customer care: the AI takes the repetitive tickets, so you hire less, especially for the season, and your team handles the cases that need a person. Sales: the goal is better reps, not fewer; new hires become productive in weeks.
AutomationWhat if the AI answers wrong or makes something up?
It answers only from your approved knowledge and your systems. When it is not sure, it hands the ticket to a person with a summary. Sensitive topics you choose (prices, legal questions, refunds above a limit) always go to a person by rule. Every reply is logged and can be reviewed.
AutomationWhich tickets can it actually close?
It starts with simple, high-volume questions: order status, invoices, refund status, how to return. With access to your systems it takes more. The first step shows exactly which of your ticket types it can close, which it can prepare for a person, and what each connection unlocks.
AutomationWho maintains it when our rules change?
Rules and texts, like a threshold or a reply template, your team edits itself without code. Flows are configuration. Integrations are ours. One person on your side reviews the handovers weekly; that is what keeps it improving.
AutomationWhat if a customer wants a human?
They get one, with the whole context passed on, so nobody has to repeat themselves.
PriceWhat does it cost?
We usually work with teams of 100+ agents, 300,000+ call minutes or 15,000+ tickets a month. The first step is a paid Insight on data you already have, priced by the size of your archive; automation is priced on what the AI handles. You get the number in the first call.
PriceWhat counts as a resolved ticket? Every vendor says it resolves.
Closed means the customer did not write back about it within five working days and nobody from your team touched it. If the customer comes back, it was not closed.
PriceHow do we know it actually saved anything?
We measure it in your inbox. First a baseline month: how many minutes your people spend per ticket, by type. Then the same measurement with the AI on. You see the hours saved from your own data, not our estimate.
PriceWhat if it doesn't deliver?
It works as well as it is set up. Track the wrong indicators and the results will be poor; spend the time on defining what you really want to follow, and the data will surface it and score it correctly. We bring our expertise and recommendations and help with the setup, but the choice of what matters is yours. And you start with the first step on your own data, so you see it working before you commit.
CompetitionWhy you and not one of the big established vendors?
The big players have been on the market a long time and have added AI to platforms built before it existed. We are the other way round: an AI platform with everything else built around it. Everything runs on AI, new models go in once they pass our tests on your data, and we are far more flexible in adapting to how you work. With AI moving every month, that is how you avoid being stuck on last year's technology.
CompetitionWe already use Gong or Chorus.
If it already works for your team, we are probably not your first priority. We fit teams that sell in several European languages, call from a phone system rather than video meetings, or want coaching run as a weekly routine instead of a library of recordings.
CompetitionWhy not Intercom Fin or Zendesk AI?
Ask how they define a resolved ticket and what you pay for. We bill on a strict definition, analyse 100% of calls and emails, and connect to your order systems. And we can work on top of the tools you already have: integrations are usually a small piece of extra work.
CompetitionWe could build this ourselves with ChatGPT.
A demo, yes. The hard part is the rest: connecting to your systems, measuring what it saved, scoring every call consistently across languages, and keeping it working when your rules change. That is the product.
Next
Thirty minutes that save you a lot of guessing.
Thirty minutes, live, on our model data or straight on your own calls: we show you what is in them and what it would find across your team. If it is not useful, we will not chase you.