Answers to the questions customers have asked most during Early Release webinars and follow-ups.
Access, Permissions & Rollout Timeline
Q: Can access be limited to specific user groups, or does turning it on make it live for the whole company?
A: Access to Eva, and the ability to create Agents, is permission-based, so it can be controlled by user group rather than turned on for the whole company at once.
Q: Will everyone get Eva access but not the agent part?
A: Both Eva access and Agent-creation access are permission-based and can be controlled separately by your account admin, so it's not automatically Eva-only.
Q: When released, will new tools appear automatically, and when will they be switched on?
A: Yes, they'll appear automatically for the specific users identified for Early Release. They'll switch on September 1.
Q: When does the trial start and end, and how do we know which users have access?
A: The trial starts September 1. End date is still being finalized (likely around a month), and we'll give you advance notice before it wraps up. We can send you the full list of users who've been granted trial access.
Cost & Credits
Q: Will there be a cost per agent after Early Release, and what will it be?
A: There's no cost to agents during this Early Release period, it's free use so you can test what saves your team the most time. Once we move toward general release (within roughly 2-4 weeks), agent costs will be introduced. We'll share exact pricing before anything is charged, so you'll have the chance to pause usage first if you want to avoid incurring cost.
Q: Will agentic credit allowances be available at the same time as pricing, or sooner?
A: Still to be confirmed, we're working on getting a definitive answer on timing.
Q: Does repeating a prompt someone else already ran use another credit?
A: Yes, for EVA, every prompt will count as a credit as it could be doing the same thing for a different record.
Q: Is there a usage limit during the pre-launch testing period?
A: EVA has a prompt limit shown in the top right corner of the screen, refreshed hourly. That limit may vary by account but should be generous. Agents have no limit during this testing period, use them as much as you'd like.
Q: Does a sequence that triggers multiple agents use one agentic credit, or multiple?
A: Credit usage is based on actual agent runs, not triggers. If a sequence fires off 4 agents, that's 4 runs. If an agent runs against 10 candidates, that's 10 runs, not 1. It's charged per run, not per trigger.
What EVA Can Do
Q: Does EVA require specific terminology, or will she understand our own vocabulary?
A: Natural language is fine, Eva does not require specific Tracker terminology to be used to understand your request. EVA will know that job spec and job description means the same thing.
Q: How does Eva determine “My Clients,” and can this use our custom ownership fields?
A: EVA interprets “my clients” as client/company records where the owner ID is your user ID. For instance, if in your session your user ID is 66, EVA would search for clients with clientOwnerId = 66. If you use a custom field, EVA can't reliably pull ownership unless it's configured for your account specifically, which would have to be done by the Tracker team. What EVA can do: use the standard client owner field, search by company name or other standard client fields, or work from a list of client names or IDs you provide.
Q: Does the AI format the Agency Resume itself, or run it through Tracker's existing template process?
A: EVA uses the existing Tracker agency resume function/template process, not an AI doing the full formatting itself.
Q: What fields can Eva update on records?
A: EVA can only change data where a specific update or create action exists. If there is no authorised action for a field, she cannot change it, even for a single record.
What EVA CAN update:
- Opportunity/job (name, description, contact, start/end date, active flag)
- Opportunity Resource/placement (start/end date, pay rate, charge rate, overtime rates, status, placement type, date confirmed, timesheet entry)
- Resume text
- Screening templates and questions (create, update, remove, reorder)
- Activities (create notes, tasks, appointments, communication logs)
- Documents (create/save generated resumes, PDFs, DOCX)
- Shortlist membership (add always, remove only if not in a working status)
What EVA CANNOT update:
- Contact core fields (owner, name, email, phone, status, job title)
- Client/company core fields (owner, status, address, type, department)
- Resource/candidate core profile fields (owner, status, contact details, headline)
- Ownership changes on contacts, clients, or opportunities (unless a specific update field is exposed)
- Any field without a supported update action
Q: What parts of a candidate record does Eva search when finding top candidates for a job?
A: EVA uses the existing longlist workflow rather than a separate matching engine. It generates a boolean search based on the job requirements, runs that search across the searchable candidate database, and then ranks the resulting candidates using the current longlist ranking method. To learn more about longlist ranking, you can view this Academy lesson: https://academy.tracker-rms.com/Home/Lesson/1068
Q: Will EVA text/email?
A: EVA can create content, but requires users to push the send button so they can edit before sending if needed. At this point, agents cannot create texts and emails. This is on our roadmap, though.
Q: Will agents create reports?
A: EVA will be able to assist with report creation and updates in the future.
Q: How can we see all Eva's features and capabilities before signing up?
A: EVA surfaces suggested prompts directly in the tool to help you get started, and you can also just ask her what she can do. Also, the LMS course (about 55 minutes) walks through every available agent and the usage dashboard in detail.
Q: Does EVA work in other languages?
A: Early indications are that it should work across different languages, but this hasn't been fully tested and confirmed across all languages yet. We're exploring this further and will share more once we have a definitive answer. This would be a great one to test during Early Release.
Agents: Setup, Triggers & Customization
Q: How customisable are agents, and are they configured at the company level or per user?
A: Agent instances and triggers are created and configured at the individual user level, so customization happens per user. Which agents are available to your team at all is controlled agency-wide via the Activation Hub, and select agents (like Resume Summary) also let you add your own custom guardrails. Depending on the agent, different options are available. For example, when configuring a resume summary, you can give the agent a name, add individual guardrails, and identify where to save the summary, as an activity, PDF, or .Docx. You then have the option to trigger this based on hundreds of action combinations.
Q: Do agent triggers on new candidates require manual entry, or will they also work with auto-added/parsed candidates?
A: Automations can be used to trigger agents when candidates are created, regardless of how they're created.
Q: If CVs aren't re-extracted, will EVA or agents be unable to assess/score candidates?
A: Assessment/scoring is based on a candidate's resume, which can be parsed when the candidate is first created or when their resume is first uploaded. There shouldn't be any need to re-extract.
Q: Can agents be used for data management, GDPR cleardown, or duplicate records?
A: Not currently on the roadmap, agent development right now is focused on the candidate and job journey. It's a good idea though, and we're taking it back to the product team.
Q: Can agents monitor and flag records based on specific conditions, like margin at a certain stage?
A: Current agents don't do this today, but this is something our product team is considering.
Q: Does the alignment agent use data from notes or previous activities?
A: The alignment agent analyzes the candidate against the job using resume skills and the job description, not notes or previous activities.
Q: Can agents handle bulk resume/cover letter downloads instead of doing it manually per candidate?
A: This isn't currently on our roadmap but could be considered. Presumably this refers to applicants, as longlisted/shortlisted candidates would already have resumes in the system.
Security, Data & Audit
Q: How secure is agent/EVA data, and does she train on or send our data outside our database?
A: Eva doesn't train herself, she is a pre-trained LLM. She cannot get outside of your database, and all data stays inside our network, it doesn't get sent outside.
Q: Does the audit history log the user, or “Eva,” for actions she completes?
A: At present, no activities are logged when EVA or agents run. The only audit trail currently captured is the prompts themselves, which are tracked by Tracker for reference. Any resulting action a user takes still logs as normal, for example, if EVA drafts an email, the user still sends it, and that logs as a sent email under that user, same as usual. Logging additional activity history has been discussed for the roadmap, but it's not confirmed yet.Feedback & Support
Q: Is there a feedback mechanism during this trial?
A: Yes, feedback goes through your CSM, just email them directly with what's working or not. If you don't have a dedicated CSM, use the support chat. We're logging all feedback, and if volume is high we'll look at building a more structured way to collect it.
Other / Plug-ins
Q: Is the Google Places API plug-in new, and does it cost?
A: It's a free plug-in available for anyone to enable.