Question 1

What is botBrains ServiceLens?

Accepted Answer

ServiceLens delivers insights for operational excellence and helps you listen more closely to your customers. It is part of the botBrains platform and analyses every customer conversation, whether by email, chat or phone. You see topics, resolution and AI involvement, grade quality with scorecards, find the conversations that need attention with monitors and get alerted to incidents.

Question 2

Does ServiceLens also analyze conversations handled only by our team?

Accepted Answer

Yes. Conversations handled only by your team get a topic a few minutes after they end, and you can analyse them with attributes and monitors to understand customer issues and processes. Whether you may also grade your team with them is covered in the Scorecards question below.

Question 3

Which channels and helpdesks are supported?

Accepted Answer

Web chat, Slack, WhatsApp, email via Zendesk and Salesforce Service Cloud, and voice calls. Each channel gets the metrics that fit it: resolution rate for chat, and for ticketing the share the AI contributed to or handled on its own.

Question 4

Do we have to maintain tags or a topic list?

Accepted Answer

No. You can start from an industry template, seed a few topics yourself or let ServiceLens find all of them on its own. New topics are marked as new, and ServiceLens never overwrites topics you set by hand.

Question 5

Does the AI learn from our agents' replies automatically?

Accepted Answer

No. ServiceLens shows you which questions the AI couldn't answer and puts your agents' replies next to them. The AI only uses the new knowledge once you approve it.

Question 6

Is ServiceLens GDPR compliant?

Accepted Answer

Yes. Conversation data stays in the EU, application data is stored in Germany, and your conversations are never used to train AI models. The Engineer runs inside the same environment and adds no new data processor.

Question 7

Can I use Scorecards with AI attributes to grade my human team too?

Accepted Answer

Not currently. AI attributes, Monitors and Scorecards built on AI attributes evaluate conversations automatically: an AI model answers a prompt for every conversation. If that evaluation targets the performance or behaviour of an identifiable person on your team, it is AI-based employee evaluation and therefore a high-risk use case under Annex III point 4(b) of Regulation (EU) 2024/1689 (EU AI Act). Allowed: grading your AI agent's conversations: This is what Scorecards are for. You are grading software, not a person, and you find wrong answers, bumpy handoffs or missing knowledge to fix in your content, procedures and integrations. Allowed: analysing team conversations to understand customers and processes: This works as long as you analyse issues and processes rather than the person. '30% of requests are about delivery delays' or 'customers are confused by the returns policy' analyses customer issues. 'Anna explains the returns policy poorly' or 'Anna follows the guidelines in 82% of her tickets' evaluates Anna. Currently not allowed: AI grading your team: Don't use AI attributes, Monitors or Scorecards with AI attributes to evaluate or compare individual employees, to select them for coaching or to prepare HR decisions. Because this is profiling, the exceptions in Art. 6(3) do not apply. 'Objective' criteria such as response time or policy adherence don't change that. Nor is it enough for a person to review the result, to add a single manual item or to hide names while the ticket, timestamps or the shift plan still allow attribution. Allowed: a person enters every value: If you want to grade your team, the scorecard must not use a single AI attribute. All attributes are maintained manually, and people enter every value themselves, for example in the Review Queue. Monitors then only select which conversations get reviewed, for example by topic or channel, and do not judge how well someone worked. That way no AI evaluates the person, and ServiceLens only calculates the weighting. Data protection and works councils: Manual evaluation still needs a legal basis under the GDPR. If you have a works council, Section 87(1) no. 6 of the German Works Constitution Act (BetrVG) applies. Co-determination only requires that the system is objectively able to monitor performance or behaviour. No intent to monitor is required (Federal Labour Court, decision of 13 December 2016, 1 ABR 7/15). Why this is currently not allowed: The requirements for a high-risk AI system under Article 6 and Annex III of the EU AI Act are currently not in place. For employee evaluation, botBrains has not carried out a conformity assessment, set up a quality management system under Article 17, drawn up technical documentation under Annex IV, issued an EU declaration of conformity, affixed CE marking or registered the system in the EU database. Under section 11(4) of our Terms of Service, any high-risk use also requires a prior written agreement. If you're interested in a high-risk use case, please contact legal@botbrains.io. This overview is not legal advice.

Question 8

How fast does ServiceLens detect incidents?

Accepted Answer

Every five minutes, ServiceLens compares each topic against the same weekday and time over the last four weeks. A confirmed incident typically shows up about 30 minutes after it starts. This needs topics and four weeks of history.

Listen to every customer.

Most conversations are never read.

What are your customers asking?

Topics

Custom attributes

Incidents

How well does your AI answer?

Resolution

Scorecards

Monitors

Find me all customer conversations where …

What should you fix next?

For everything else, ask the Engineer.

Common questions

Make better decisions with data.