FAQ
Questions about Avara
Common questions from agency directors, account team leads, and operations managers evaluating account health tooling.
Data and sources
Avara analyzes three categories of qualitative data: meeting notes and call summaries, survey verbatims (open-ended text responses from client satisfaction or pulse surveys), and email thread content. The platform extracts sentiment, topic, engagement frequency, and tone signals from all three sources and combines them into a single health score per account.
Quantitative data such as invoice history, ticket volume, or NPS scores is not currently analyzed. The premise of Avara is that relationship cooling appears first in qualitative text, before it appears in any quantitative metric.
There are four ways to bring data in: CSV upload (works for meeting notes, survey exports, or any structured text you already have in spreadsheets), Google Forms integration (connects directly to your existing client survey responses), paste-in (copy-paste text directly into the interface for ad-hoc or one-off ingestion), and REST API access (available on the Radar tier for teams that want to push data from their own CRM or survey platform programmatically).
Email integration with read-only Gmail and Outlook access is on the roadmap. It is not available in the current release.
You can begin with whatever data you have. The platform generates a baseline score from the first batch of data you import, whether that is two months of meeting notes or twelve months of survey verbatims. More historical data gives you a longer trend line, which makes the signals more meaningful, but the health score and risk flags are functional from the first import.
For new accounts where no historical data exists, scores are marked as "initializing" until sufficient data has accumulated, typically after three imported documents or interactions.
The current release is optimized for English-language text and produces reliable results for English meeting notes and survey verbatims. Swedish, German, and Dutch analysis is in testing and will be available in a later release.
If your agency operates in multiple languages and non-English analysis is a requirement, let us know via the contact form. We prioritize language expansion based on actual demand from our user base.
How scoring works
The health score is a number from 0 to 100 that represents the strength and positivity of the relationship as evidenced in recent qualitative interactions. Scores above 70 indicate a healthy, engaged relationship. Scores between 50 and 70 suggest a relationship worth monitoring. Scores below 50 indicate cooling and warrant proactive outreach.
The score incorporates four signals: sentiment drift (whether the emotional tone of interactions is improving or declining over time), topic clustering (whether conversations are narrowing around problems or expanding into new opportunities), engagement frequency (whether the pace of interaction is increasing or decreasing), and verbatim intensity scoring (the strength of language used in open-text responses).
The score updates every time new data is imported for a given account, whether via CSV, Google Forms sync, paste-in, or API push. The platform processes the new data and recalculates the score within a few minutes of ingestion.
There is no fixed refresh schedule. The score is current as of the most recent data import. If you want daily score updates, you need to push new data at least daily for each account you want to monitor at that cadence.
Scoring customization is on the product roadmap. In the current release, the scoring algorithm uses default weights that are calibrated to agency and consultancy relationship data from our pilot program.
If you have a strong view that one signal type (for example, survey verbatim intensity) should be weighted more heavily for your specific business, note that in your early access request and we will discuss it during onboarding. Custom weighting is available for annual Radar tier accounts on request.
Security and data handling
Data is stored in EU-based infrastructure (Frankfurt region). Each agency's account data is logically isolated at the tenant level. Avara staff have no access to the content of your meeting notes, survey verbatims, or email threads. Access to metadata (account counts, score ranges) is limited to a small engineering team for infrastructure monitoring purposes.
All data in transit uses TLS 1.3. Data at rest is encrypted using AES-256. We do not use customer data to train or improve our models.
Avara Insights AB is a Swedish company operating under GDPR as both a controller (for your account and subscription data) and a processor (for the meeting notes and survey text you import). We operate under Data Processing Agreements with all agencies that process personal data of their clients' employees or representatives within the platform.
You can request data deletion, export, and DPA agreements via the contact form or directly from your account settings. Data deletion requests are processed within 30 days.
No. The content of your meeting notes, survey verbatims, and any other imported text is not used to train, fine-tune, or improve Avara's NLP models. Your data is used only to generate scores and signals for your own account portfolio.
Model improvement at Avara relies on synthetic training data and, for scoring calibration, on aggregate anonymized score distributions shared voluntarily by pilot participants under explicit data sharing agreements.
Pricing and getting started
An account is a single client relationship that you are actively monitoring in Avara. If you work with 30 clients, that is 30 accounts. Accounts are counted at the point of activation, which means creating the account record and importing at least one document for it.
Accounts you have created but not yet imported data for, or accounts marked as inactive, do not count toward your billing total. You can archive accounts that have churned or been closed without deleting their data, and archived accounts are not billed.
Yes. You can upgrade at any time from within your account settings. The change takes effect at the start of the next billing period for monthly plans, or immediately (prorated) for annual plans. There is no penalty or setup fee for upgrading.
Downgrading is also possible, but takes effect at the end of your current billing period. If you are on a monthly plan, you can downgrade with 30 days notice. On an annual plan, downgrade takes effect at renewal.
The Signal tier includes a 14-day trial. During the trial you have full access to all Signal tier features for up to 50 accounts. No payment information is required to start the trial. At the end of the 14 days, you can convert to a paid plan or your account will automatically deactivate.
We do not offer a free trial of the Radar tier. If you need to evaluate the REST API or custom scoring for a large account portfolio, contact us and we can arrange a structured pilot program.
There is no setup fee on any tier. You pay the published monthly or annual rate and nothing else. Onboarding is self-serve for Pulse and Signal tiers, with documentation and a guided import wizard to get your first accounts scored in under 30 minutes.
Radar tier accounts receive a 60-minute onboarding call with the Avara team, included in the plan. If you are migrating from a CRM or custom analytics setup and need additional configuration support, contact us and we can quote a structured onboarding engagement separately.
When you cancel, your account enters a 30-day grace period during which you retain read-only access to your data and can export it in CSV format. After the grace period, your account data is queued for deletion. Deletion is completed within 60 days of cancellation.
If you cancel and re-subscribe within the 30-day grace period, your account and data are restored in full. Reactivation after the grace period starts fresh with no historical data retained.
Still have questions?
Talk to the team directly
We are a small team and Henrik answers most inbound himself. If you have a question that is not covered here, reach out via the contact form or email directly.