Frequently Asked Questions

An AI employee is a specialized autonomous agent that performs real business tasks – finding leads, sending emails, answering customer questions, monitoring competitors – 24/7 without human supervision. Unlike chatbots, AI employees don't just answer questions. They take actions: search the web, send emails, write reports, and deliver results directly to your messenger.

ChatGPT, Claude, Gemini, and other LLMs are conversation tools – you ask, they answer. Our AI employees are autonomous workers. You set them up once, and they execute tasks on their own: find leads, write personalized emails, monitor competitors, generate reports. They work in the background 24/7 and deliver results to your Telegram, WhatsApp, or Slack – no prompting required. Think of it this way: ChatGPT, Claude, or Gemini is a consultant you have to call every time. An ewpire agent is a full-time employee who just does the work.

About 5 minutes total. After checkout, your agent environment is provisioned in roughly 30 seconds. You receive a welcome email, click one button to connect Telegram, answer 8 quick onboarding questions (your business, target audience, tone of voice, preferred language), and the agent starts working. No technical skills required.

No. Everything is managed through your messenger (Telegram, WhatsApp, Slack). You communicate with your AI employee the same way you'd message a human colleague.

Telegram is live today. Slack, Discord, iMessage, Microsoft Teams, Signal, and Matrix are rolling out via our channel integrations. WhatsApp is planned for a later phase. All agents share a single memory across every connected channel – start in Telegram, continue in Slack, context carries over automatically.

Yes! Starter plan includes a 3-day free trial. Card required at signup, cancel anytime in 1 click before day 3 and you won't be charged. Pro and Business plans don't have a trial – they're for teams ready to commit.

Yes! All plans include voice message support. Simply send a voice note in your messenger (Telegram, WhatsApp, Slack, Discord) and your AI employee will understand it. Voice messages are automatically transcribed and processed as text – no extra cost.

All AI employees are multilingual. During onboarding, you choose your communication language – the agent communicates with you in that language (reports, alerts, approvals). For content-producing agents (Sales, Growth), you can also choose a separate content language – e.g., communicate in German but produce marketing content in English. Language settings can be changed anytime via /settings.

Each AI employee is powered by the latest frontier AI models and runs on our infrastructure 24/7. They're not chatbots – they're autonomous specialists. Your Sales Agent uses 4 proven email frameworks (PAS, BAB, QVC, AIDA) and detects trigger events for 3x reply rates. Your Research Agent analyzes every article across 9 dimensions including sentiment and weak signal detection. Your Support Agent aims for first-contact resolution with confidence scoring. Every agent operates with expert-level prompts, built-in quality checks, and delivers results directly to your messenger.

Yes, literally. AI employees don't sleep, take breaks, or call in sick. They work every hour of every day, including weekends and holidays. Your Sales Agent sends emails during business hours in the recipient's timezone, but the system itself never stops.

Through your preferred messenger. Send a message to adjust settings, ask for a report, or change priorities – just like messaging a human colleague. The AI understands natural language, so you don't need special commands.

Absolutely. During onboarding, you define the agent's parameters: target audience, tone of voice, priorities, working hours for outbound, and more. You can adjust these settings anytime through your messenger.

AI employees have built-in confidence thresholds. When unsure about a decision (e.g., a customer question it can't answer), it escalates to you via messenger with full context. All plans include approval mode – you can review critical actions like emails before they're sent.

All plans have the same limits: 50 personalized emails per day for Sales Agent (with mandatory 4-week warmup: 5→10→20→30→50), and 50 AI-generated images per day for Growth Agent ad creatives. Core tasks like support (unlimited queries), research (unlimited scans), lead qualification (unlimited conversations), tender monitoring (unlimited searches), and document processing (50 documents per batch) have no daily caps.

The Sales Agent automatically selects the best framework per prospect: PAS (Problem-Agitate-Solve) for pain-aware prospects, BAB (Before-After-Bridge) for visionary buyers, QVC (Question-Value-CTA) for C-level (shortest, most direct), and AIDA (Attention-Interest-Desire-Action) for cold-cold prospects. Selection is based on the prospect's seniority, industry, and company stage – CEOs get 60-word peer-to-peer emails, managers get 100-word practical ones.

Trigger events are recent company changes – hiring, funding rounds, product launches, leadership changes. Emails referencing trigger events get 3x higher reply rates because they feel timely and relevant. The agent detects these automatically from public data and leads with them when found.

All plans: up to 50 emails per day. Mandatory warmup: Week 1: 5-10, Week 2: 10-20, Week 3: 20-30, Week 4+: up to 50. Each email is unique, research-backed at 75-100 words optimal length, and passes anti-AI detection filters.

Yes. The agent uses anti-AI techniques: varied sentence lengths, natural contractions, imperfect phrasing, odd-numbered lists. Combined with SPF/DKIM/DMARC from your own domain, emails bypass 2026 Gmail and Outlook AI filters.

4 unique strategies over 14 days: Day 2 – different angle, Day 4 – social proof (results from similar companies), Day 7 – value-add (useful insight), Day 14 – breakup email. Each adds NEW value – never 'just checking in'. The breakup email paradox: giving permission to leave often triggers responses.

Every tender is analyzed for: relevance, complexity, bid/no-bid decision (GO/NO-GO signals), win probability (0-100%), evaluation method (MEAT/QCBS/lowest price), scoring weights, positioning strategy, compliance flags, deadline urgency, competitive advantages, estimated effort, strategic value, consortium advice, and lot recommendation.

When your company can't deliver alone, the agent doesn't just say 'skip'. It recommends a partnering strategy: what capability gaps exist, whether you should be prime contractor or sub, and what type of partner to look for. This expands your pool of viable tenders.

40+ countries via official government APIs: all 27 EU states (TED), USA (SAM.gov), UK, Australia, Canada, UN (UNGM), Ukraine (ProZorro), Mexico, Colombia, and more. Free, legal, no scraping.

The agent aims to completely solve the customer's issue in a single response. It reads between the lines – 'it's not working' becomes a targeted diagnostic. Every answer cites sources from your knowledge base. Frustrated customers get empathy first, then solutions.

Every answer gets a score (0-100). Below 60%, or after 2 low-confidence responses in a row, the agent auto-escalates to a human with full context. You can also set keywords that trigger instant escalation.

Upload documentation, FAQ, help articles (PDF, DOCX, Markdown, URLs). The agent chunks, embeds with vector search, and retrieves the most relevant context for each question. It only answers from your knowledge base – never makes up information.

Exactly 3 days (72 hours) from activation. You'll receive a reminder email 24 hours before the trial ends.

Only the Starter plan – both monthly ($199/€199/£199 per month) and annual ($1,990/€1,990/£1,990 per year). Pro and Business are for teams ready to commit – contact us if you need to evaluate them.

Stripe places a small authorization hold to verify your card is valid. The hold is released within 5–7 days. No actual charge happens during the 3-day trial. After the trial, your card is automatically charged the Starter plan rate (monthly or annual, depending on what you selected) unless you cancel.

Send /cancel to the Telegram bot – the easiest way. Alternatively, go to the Stripe Customer Portal via /billing. You can also email [email protected] (best-effort processing). If you cancel before day 3, you won't be charged.

One free trial per email address and per payment card. We use both checks to prevent abuse. If you've previously had a trial, please contact us.

When you start the trial, you explicitly waive your 14-day right of withdrawal under Article 16(m) of EU Directive 2011/83/EU. This is required because the digital service starts immediately upon trial activation. You consent to this via a checkbox at signup. However, our voluntary 14-day money-back guarantee applies after your first paid month – see the next question.

After your first paid charge (after day 3), you have a 14-day money-back guarantee on your first month's subscription. This is a voluntary commercial offer separate from EU statutory withdrawal rights. Email [email protected] within 14 days of first charge for a full refund.

If you cancel before day 3 and don't convert to paid, your account data is deleted within 30 days per our privacy policy. You can request immediate deletion at any time under GDPR Article 17 by emailing [email protected].

Monthly: Starter $199/mo (1 AI agent), Pro $499/mo (3 AI agents), Business $1,499/mo (all 9). Annual: save ~17% (2 months free) – Starter $166/mo, Pro $416/mo, Business $1,249/mo, billed yearly. Add extra agents for $199/mo each. All plans include the same features. No hidden fees.

No. Monthly plans are month-to-month. Annual plans are billed once per year at ~17% discount (2 months free). Cancel anytime – no cancellation fees, no questions asked. Your agent stops working at the end of the billing period. Every new subscription comes with a 14-day money-back guarantee.

All plans include: approval mode, custom knowledge base, all messenger integrations, voice messages, daily reports, 50 emails/day, 50 AI images/day. The only difference is the number of agents: Starter (1), Pro (3), Business (all 9). You can add extra agents for $199/month each.

Yes. Send /upgrade in the Telegram bot to switch from Starter to Pro, or Pro to Business. The upgrade takes effect immediately with prorated billing – you pay only the difference for the remaining days in your current period. Downgrades happen via the Stripe Customer Portal (run /billing for the link) and take effect at the next renewal.

Yes. Send /cancel in the Telegram bot. For trials: cancels immediately, no charge. For active subscriptions: your agents keep running until the end of the current billing period, then stop. No cancellation fees, no questions asked. You can also cancel in the Stripe Customer Portal via /billing.

Go to ewpire.com/recover and enter the email you used at checkout. You'll receive a secure one-time link to reconnect your agents to a new messenger account. The entire process is automatic – no waiting, no support tickets. Your agents and data remain safe on our servers.

We accept all major credit and debit cards (Visa, Mastercard, Amex) through Stripe. All payments are processed securely with PCI-DSS compliance.

All prices on the website are displayed excluding VAT. Applicable taxes are calculated automatically at checkout based on your billing country. EU businesses with a valid VAT ID are exempt from VAT (reverse charge mechanism).

No. If you provide a valid EU VAT ID during checkout, the reverse charge mechanism applies and you pay 0% VAT. Your VAT ID is validated in real-time against the EU VIES database. The invoice will state 'Reverse charge – VAT to be accounted for by the recipient'.

We detect your location automatically and show prices in € for Eurozone countries, £ for the UK, or $ for the rest of the world. The nominal amounts are the same across currencies ($199 = €199 = £199). Applicable taxes are calculated at checkout based on your billing country.

Yes. Add extra agents for $199/month each. All agents are available in all plans – the only difference is how many are included. If the math works out, upgrading to the next plan is usually cheaper than adding agents individually. Business includes all 9 agents.

Use /swap in the Telegram bot to replace an agent slot with a different specialist from the 9 available. The bot shows a confirmation prompt before acting. Warning: this permanently deletes all conversation history, settings, and the knowledge base for the replaced agent. The new agent starts fresh. This is different from /use, which just switches which agent in your existing roster you are talking to – non-destructive.

Yes. Two ways: (1) During onboarding, upload files (PDF, DOCX, Markdown, URLs) – the agent chunks and indexes them for retrieval on every response. (2) Any time after onboarding, run /knowledge add <url> in the bot to feed a specific URL into the active agent's knowledge base. Uploaded knowledge persists permanently and is used on every relevant response. Each agent has its own independent knowledge base – uploading to Sales does not affect Support Agent.

Yes on Pro and Business plans. Run /model byok-add in the bot and provide your API key from your preferred AI provider. Your key is used for your agent's inference – you pay the provider directly at their rates. The default configuration handles everything automatically if you prefer not to manage this. BYOK is optional, not required.

Yes! Every new subscription includes a 14-day money-back guarantee. If you're not satisfied for any reason, contact us within 14 days and we'll issue a full refund – no questions asked.

Yes on Pro and Business plans. The admin runs /invite <agent_slug> <email> – the colleague gets an email with a Telegram link, picks their own communication language, and starts chatting. Only one person can use a given agent at a time (admin or operator, not both). Admin can revoke access any time via /operators revoke <id>. Starter plan does not include team delegation.

Just type a message in your messenger (Telegram today, WhatsApp + Slack + Discord rolling out Day 25+). Your agent replies in 30-60 seconds. You can use plain language – no special syntax required. Need more control? See the commands below.

15 standard commands grouped into 4 sections: Account: /start · /onboard · /settings · /billing · /invite Multi-agent: /agents · /use · /current · /reset · /memory · /transfer Monitoring: /status · /report · /stop Help: /help Plus admin commands: /swap (replace an agent with a different specialist), /upgrade (switch plan), /cancel (cancel subscription), /operators list/revoke, /channels add-bot (connect your own branded bot, Pro/Business), /model byok-add (bring your own AI provider key, Pro/Business), /human (request human review), /knowledge add <url> (feed a URL into the active agent's knowledge base). Tap the command menu button (☰) in Telegram to see them all with descriptions.

Use /use <slug> – example: /use sales. The agent becomes sticky until you switch again. All new messages go to the active agent. Check who's active with /current. Each agent has its own separate memory – switching doesn't mix conversations.

Type @<slug> at the start of your message – example: @concierge book Monday 2pm with John. This is a one-shot route – the mentioned agent replies, but your active agent doesn't change. Perfect for quick questions to a different specialist without losing your main conversation.

/memory shows the last 10 turns of conversation with the currently active agent (you + agent messages). Useful for reviewing what was discussed. /reset wipes the memory of the active agent – next message starts fresh as if you'd never talked. Other agents' memories are untouched.

/transfer <slug> hands off the conversation from the current agent to another with a 2-3 sentence summary auto-generated by AI. Example: you've been discussing prospects with Sales → /transfer concierge → Concierge receives a handoff brief and continues naturally. Useful for multi-step workflows that span specialties.

/billing shows your plan, status, running agents, and a link to the account portal. /invite (Pro and Business) auto-emails a colleague a secure deep-link – they click, pick their own language, and start using the agent. Admin keeps full control: /operators list to see who has access, /operators revoke <id> to remove.

Yes – your agent has a single shared memory across all connected channels. Start a conversation with Sales in Telegram, pick it up in Slack an hour later, and the agent knows exactly where you left off. Memory is per agent, not per channel – so your Sales agent has one conversation history that spans every messenger you've connected. Use /memory to review the last 10 turns. Use /reset to start fresh across all channels at once.

Nothing breaks. Your agent treats every channel as a window into the same conversation. Switch any time – the context, the task you described, the decisions already made – all carry over automatically. You don't need to explain yourself again. The agent picks up where it left off.

Each agent has its own independent memory – Sales doesn't share context with HR, and HR doesn't share context with Concierge. Within each agent, memory spans all channels and the last 10 conversation turns stay active as rolling context. For deeper long-term learning (your internal SOPs, product docs, company-specific knowledge), upload a knowledge base via /onboard. That knowledge persists permanently and is used on every response.

Multi-agent commands aren't needed – everything goes to your one agent. /agents shows it, /use / /current / /transfer are unnecessary. /memory + /reset still work. You'll still see the commands in the menu; they'll reply with a friendly note if not applicable to your plan.

Axi is the Support Agent product live on our own website – you're looking at us dogfooding. The name Axi is short for Axiom, because support answers shouldn't be guesses. Click the chat bubble bottom-right on any ewpire.com page to talk to Axi, ask pricing/product questions, book trials, or try features like voice input. Axi answers from our knowledge base with source citations and confidence scores.

Yes – that's Product #2 Support Agent. You subscribe, upload your knowledge base (PDFs, URLs, FAQs), configure tone, and embed a single-line <script> snippet on your site. Your customers get the same experience Axi gives our visitors: first-contact resolution, confidence scoring, voice support, multilingual, automatic escalation when stuck. Price is included in Pro and Business; add to Starter as +$199/mo add-on.

Axi uses a speed-optimized AI configuration tuned for sub-second response time, because customer-facing support needs instant answers. The other agents use a more powerful reasoning configuration optimized for longer analytical tasks (typically 30–60 seconds). Your own Support Agent deployment uses the same speed tier by default. Both are production-grade and tested 24/7.

Axi conversations are retained 30 days for quality monitoring (anonymous session IDs, no PII beyond what you voluntarily share). Your Support Agent deployment for your customers follows the same pattern – you own the data, we provide the infrastructure. GDPR Article 17 delete requests honored within 24 hours.

Before subscribing, you can see one of our 9 agents – Support – live on our own site. If Axi answers your questions well, the other agents (Sales, Research, Growth, HR, Tender, Concierge, etc.) work the same way for your business. It's proof-of-work, not a pitch.

Run /invite <agent-slug> <email> in the bot. Example: /invite sales [email protected]. Jane receives an email from ewpire with a Telegram deep-link. She clicks it, answers 2 onboarding questions (her communication language, then a context summary), and she is ready to use the agent. No technical setup on her end. /invite is available on Pro and Business plans.

No. Each agent supports one active conversation at a time. If your colleague is chatting with the Sales Agent, that session is theirs until it finishes. As the admin, you can view conversation history and stats even when an operator is active – but you would not send messages into the same session simultaneously. This ensures the agent maintains coherent, uninterrupted context for whoever is working with it.

This is a known Safari deep-link edge case. Safari on macOS sometimes does not pass the activation token through to the Telegram app. Fix: open the Telegram app directly, search for @ewpirecom_bot, and type /start and send it manually. The bot recognizes your account from your Stripe email and completes the binding identically. Chrome and Firefox on macOS handle the deep-link automatically.

Yes. Run /settings in the bot at any time. An inline keyboard lets you update your communication language, brand voice, content language(s), and ICP. Changes take effect on the next message. You can also re-run the full guided onboarding flow via /onboard if you prefer a step-by-step experience.

Yes on Pro and Business plans. Create a bot via BotFather in Telegram, copy the token, then run /channels add-bot <token> in your ewpire bot. Your custom bot becomes the entry point for your agents – your customers see your brand name, not ewpire. Your agents, memory, and all commands work identically through your own bot.

Communication language (set in Q1) controls how your agent talks to you – reports, alerts, status messages, confirmations inside the chat. Content language (set in Q7) controls what your agent produces – cold emails, articles, tender responses, proposals. They are independent. Example: you communicate in German, but your Sales Agent sends cold emails in English to US prospects. Each operator you invite sets their own communication language, but content language is set business-wide by the admin and applies to everyone.

Run /operators list. It shows all operators across your agents with their assigned agent, last activity timestamp, and operator ID. The /status command shows per-agent health including current session state. A more detailed analytics view is planned for a future release.

Run /operators list to get the operator ID, then run /operators revoke <id>. Revocation is immediate – their Telegram binding is dropped and they cannot send messages to the agent. Their conversation history remains in your account for continuity. You can re-invite them at any time.

AI Presence shows how your brand appears when people ask AI engines questions in your market – across ChatGPT, Google AI Overviews, and others. If a buyer asks 'what's the best logistics software for mid-size manufacturers?' and your brand is absent from the answer, you are missing a new class of buyer who never clicks search results at all. The free scan gives you a snapshot. The $39 full report runs 50 targeted prompts and delivers 5 actionable fixes.

The full report runs 50 prompts specifically targeted to your market and competitive set, covers 5+ AI engines (vs 2 in the free scan), provides a full share-of-voice breakdown comparing you to competitors, and delivers 5 specific improvements you can implement this week. Delivery within 24 hours. One-time purchase, not a subscription.

The $39 report is a single snapshot. The $49/month subscription reruns the same 50-prompt scan weekly and alerts you when your share-of-voice drops or a competitor gains ground. It is for businesses that need ongoing visibility into how AI engines describe their brand over time.

ewpire's Growth Agent ($199/month) acts on the findings from your AI Presence report. If AI engines consistently describe you as a 'regional player' when you operate nationally, the Growth Agent writes the long-form content and landing copy that corrects that framing over time. AI Presence identifies the gap. Growth Agent closes it.

Yes! Our Referral Programme pays 20% of the first month's subscription for every customer you refer. Sign up at ewpire.com/refer to get your unique referral link. When someone subscribes using your link, you earn a commission paid after their first month completes (approximately 30 days after payment). Payouts are processed via Stripe Connect to your bank account. Important: all tax obligations on referral income are your responsibility under the laws of your country of residence. ewpire does not withhold taxes or provide tax advice. Full terms at ewpire.com/referral-terms.

ewpire agents run on a frontier AI model selected for speed, reasoning quality, and cost at the scale we operate. The default configuration handles the full range of production workflows – qualifying leads, drafting emails, screening candidates, analyzing tenders, processing documents – without any tuning on your side. You focus on outcomes, not on which model to pick.

No. We use models under commercial agreements that exclude use of your inputs and outputs for training foundation models. Conversations are retained only as long as your agent needs them for context (last 10 turns as rolling memory, knowledge base you upload for persistent grounding) and are deleted under GDPR-compliant retention windows. Full details are in our Privacy Policy.

Yes. We use AES-256 encryption at rest, TLS 1.3 in transit, and complete data isolation between clients. Your data is never shared with other customers or used to train AI models. We are GDPR-compliant (registered in Estonia, EU).

No. Your business data is used solely to power your AI employee's tasks. We never use client data for model training, benchmarking, or any other purpose. Data is deleted when you cancel your subscription.

All data is stored on secure European servers (EU jurisdiction). Our infrastructure provider maintains SOC 2 Type II compliance. We perform regular security audits and penetration testing.

Only your AI agent accesses your data during task execution. Our engineering team has no access to customer data in production. All access is logged and audited.