Most "AI support" is a search bar in disguise — type a question, get back a paragraph someone wrote six months ago. It doesn't know your account. It doesn't know what actually happened. It just matches keywords and hopes.

Clara doesn't work that way. She's built directly into the Talkpush CRM, not bolted on as a chat widget — which means when you ask her something, she can actually go look.

Clara responds first when you open a conversation, and a human Product Support agent is always ready to step in when needed.

🧠 Why Clara is different

Most support chatbots can only match your question to a list of pre-written answers. If your issue doesn't fit one of them, you're stuck waiting for a human anyway.

Clara isn't layered on top of the product as a separate chat tool — she's built directly into the Talkpush CRM. That means she can look at how a feature actually behaves in your account and reason through your specific situation, instead of just recognizing keywords and reciting a canned response. In practice, that makes her:

More accurate — answers are grounded in how the product actually works, not a static script.
Available everywhere — the Clara widget lives on every screen in the CRM, so you never have to hunt for a "help" button.
Context-aware — she keeps the thread of your conversation, including any ticket she's opened for you, so you never have to re-explain yourself.
Knows who she's talking to — Clara recognizes you as the person in the conversation, not just a logged-in account. Your own past actions get labeled "you," not just your name.
Scoped to your company — Clara only ever looks at your own account's data. She can't see or check other companies' records, even if you ask her to.

💬 Clara at a Glance

Answer product questions — how a feature works, where to find a setting, or what a status means.
Check on open requests — get an update on an existing ticket or escalation.
Walk you through common fixes — step-by-step guidance for frequent issues, like sync errors or campaign setup.
Point you to the right help article — Clara can surface relevant help center content.
Open and track a support ticket — no forms — just describe the issue in your own words, and Clara creates a tracked ticket for you automatically. Mention the same issue again in that conversation and it stays as one ticket instead of opening a duplicate.
Accept screenshots, files, and voice dictation — attach a screenshot or screen recording of the bug, or tap the mic and speak instead of typing — Clara transcribes it straight into the text box for you to review before sending.
Loop in a human when needed — if your question is complex, sensitive, or needs account-specific investigation, Clara hands off to a PS agent.

⚙️ How it works

Open the Clara Messenger widget and start a conversation, as you normally would.
Clara responds first and works through your question — for simple queries, that usually means an answer within a minute.
For more complex or technical questions, Clara takes extra time to work out a full explanation, so a response may take a bit longer.
If the issue needs a human, Clara flags it and a Product Support agent picks up the conversation. You won't need to repeat your question — the context carries over.

💡 Tip: The more specific your question, the easier it is for Clara to help. Include the campaign name, error message, or candidate ID when relevant — or just attach a screenshot.

🔍 Real problems Clara can dig into

Clara doesn't stop at "here's how that feature works." She can actually investigate what happened in your account — the same kind of digging that used to mean opening a ticket and waiting. Here's the range of things she handles:

Category	What you're running into	What Clara checks
Audit & attribution	You need to know exactly who moved a candidate, edited a campaign, or made a change — and when.	Pulls the specific action, the person or automation behind it, and the exact timestamp — whether it was a teammate, an Autoflow rule, or another automation.
Assessments & scoring	A candidate's assessment results or score never showed up on their profile — or you just want to understand how a score was calculated.	Whether the assessment actually triggered and where a result got stuck — or breaks down exactly how a score was calculated, component by component.
Candidate readiness checks	You want a quick read on whether a candidate's file has enough data behind it to support a decision.	Flags missing scores, placeholder answers, or incomplete assessments — and says plainly when it's a "needs human review" situation instead of guessing.
Autoflow & pipeline behavior	A candidate seems stuck in a stage, or wasn't moved or rejected the way you expected.	The exact Autoflow rule involved, whether it fired, and which folder or label condition did or didn't match.
Messaging delivery	Candidates say they never received an invite, SMS, or WhatsApp message.	Whether the message actually sent, which channel it used, and why it may have failed — opt-in status, a carrier rejection, a missing link.
Integrations	A candidate's status doesn't match your HR system, or you're not sure how a connected integration — HRIS, job boards, lead ads — is actually set up.	Compares both sides of an HRIS sync to find a delay, mapping gap, or dead webhook — or explains exactly how something like a job board feed or Facebook Lead Ads connection is configured on your account.
Ad hoc reporting	You need a breakdown of applications by source, time period, or outcome — but no report is set up for it.	Builds the summary straight from your account data, instead of pointing you to a dashboard you'd have to configure yourself.
Duplicate & data checks	A candidate looks duplicated, or something on their record looks off.	Cross-checks the records involved and flags exactly where and how the mismatch happened.
Configuration questions	"Can we set up X?" or "where do I find Y setting?"	Answers directly from how your account is actually configured — and says so plainly if the setting doesn't exist yet, instead of guessing.

🗨️ See Clara in action

A few examples of what a conversation with Clara looks like:

Notice the detail: Clara doesn't just say "this application was moved" — she says "Crismin Lagamayo (you) manually moved this application." She knows exactly who's asking, not just which account is logged in. No more guessing who moved what, or digging through group chats to reconstruct what happened — and when you already have a ticket open, she gives you the real status straight from the source, not a generic "we're on it."

Drop in a screenshot of a failed message and Clara reads it, cross-checks what actually happened on her end, and tells you plainly whether it's worth escalating — or already resolved. Need a breakdown that doesn't exist as a report yet? She'll build it from your account data on the spot.

This is the difference between a chatbot that explains how Autoflow generally works, and an assistant that can tell you exactly which rule fired on this candidate, on this date, and why. Same goes for campaign changes — Clara knows exactly who created what, and when.

Here's the part most AI vendors won't show you: what happens when there isn't enough data for a confident answer. Asked to judge whether a candidate was hireable, Clara didn't guess — she flagged the missing assessment and placeholder answers, and pointed straight to human review. That's not a limitation. That's the design.

Questions like "how does our VONQ integration actually work?" normally mean waiting on a solutions engineer. Clara can answer it immediately — because she's reading your account's real configuration, not reciting generic documentation.

Clara is scoped strictly to your own company's data — she can't see or check another client's account, even if asked to. And she can just as easily hand you a straight number, like total emails sent this week, without redirecting you to a dashboard.

🗂️ Your support history, always at hand

Every conversation and ticket you've had with Clara stays organized inside the widget:

Tickets — see every open and closed ticket at a glance, with status pills so you know where things stand without asking.
Messages — your past conversations, most recent first. Click any of them to pick up right where you left off.
One click back in — reopening a ticket or past chat brings you straight back into that conversation, with full context, instead of starting over.

Q: Is Clara available 24/7?	Yes. Clara is available around the clock, while human agents follow regional support hours.
Q: Why is Clara taking a while to respond?	For complex or technical questions, Clara works through a fuller explanation before replying, which can take longer than a simple lookup. If you need a faster answer, you can ask to speak to a human at any time.
Q: Can I ask to speak to a human right away?	Yes, just ask, and Clara will route you to the next available Product Support agent.
Q: Does Clara have access to my account data?	Clara can reference general account and ticket information to help answer your question, but she's scoped to your own company's data only — she can't see or check other companies' accounts, even if asked to.
Q: Can I attach files or dictate instead of typing?	Yes — use the paperclip icon (or just paste or drag a file in) to attach screenshots or documents. To dictate, tap the mic and speak — Clara transcribes your speech into the text box so you can review or edit it before sending.
Q: What happens once my ticket is closed?	That conversation is closed for new messages. If the issue comes back or you have something new, just start a fresh conversation with Clara.
Q: Will I see Clara on every account?	Clara is rolling out company by company during this beta. If you don't see the widget yet, it just means your account hasn't been enabled for it yet.
Q: Clara is in beta — what does that mean for me?	Clara is still learning and improving. Most questions get a solid answer, but if something seems off, a Product Support agent is always just a message away.

Have feedback on an answer Clara gave you? Let your PS agent know — it helps us keep improving.

Meet Clara: Your Talkpush Support Assistant