Question 1

What is an AI agent and how is it different from ChatGPT?

Accepted Answer

ChatGPT is a tool you have to prompt manually. An AI agent is an autonomous system embedded in your processes (CRM, messengers, databases). It makes decisions, takes actions, and works 24/7 without a human in the loop. ChatGPT doesn't know your clients, can't update your CRM, can't sync data between systems. An AI agent does all of this automatically.

Question 2

How fast does an AI agent pay for itself?

Accepted Answer

Typical: pays for itself within 1 to 6 months depending on scope. Example: a 3-person support team costs $6,000 a month. An agent costs $3,500 to set up plus $750 a month to run. After month one you are already saving.

Question 3

How long does implementation take?

Accepted Answer

7 to 25 business days depending on scope. That covers process audit, setup, integrations with your tools, team training, and a testing period before go-live.

Question 4

How do you handle our data?

Accepted Answer

We run on closed servers or isolated infrastructure, sign NDAs, encrypt data at rest and in transit, and use guardrails so your data never reaches public LLMs. We never train any public model on your information.

Question 5

What if we want to test before committing fully?

Accepted Answer

We recommend exactly that. Pick one painful process, deploy an agent on it, measure results over 1 to 2 months, then scale to the next process. Lower risk, real proof of concept on your numbers.

Question 6

Will an AI agent replace my employees?

Accepted Answer

It depends on the task. Routine work (intake, FAQs, CRM updates, follow-ups) gets handled fully. Complex work (negotiations, edge cases, judgment calls) stays with humans, the agent prepares the context. In practice the team stops doing routine and starts doing work that actually requires their expertise.

Question 7

Why such a big price range? I have seen AI agents for $20 a month and for $10,000.

Accepted Answer

Cheap solutions are "ChatGPT in a nice wrapper." A custom agent is a digital employee that actually knows your business: your clients, your data, your workflows. That is a fundamentally different product.

Question 8

What sets you apart from other AI agencies?

Accepted Answer

1. Custom only. Every agent is built around your business, your data, your workflows. No templates. 2. Transparent ROI. We calculate the economics on your real numbers before the project starts. 3. No salespeople. From call one you talk to the people who will build your agent. 4. Automation first, AI second. We clean up processes before deploying AI, that is why our deployments actually work.

Question 9

What guarantees do you offer?

Accepted Answer

We do not start a project unless we see clear economics. In the audit phase we show you exactly where you are losing money, how much you will save, and the success metrics we will track. If the agent does not hit those metrics within 30 days of go-live, we fix it for free.

Question 10

What happens to our agent and data if we stop working with you?

Accepted Answer

Everything is yours. Code, prompts, training data, integrations, vector stores, all delivered with documentation. You can move the agent to another vendor or take it in-house. We do not lock our work behind proprietary formats.

Question 11

What if the AI says something wrong to our client?

Accepted Answer

Three layers of protection. First, every agent runs in a sandbox with strict guardrails on what it can and cannot say. Second, sensitive actions (refunds, contract terms, anything that touches money or commitments) require human approval. Third, every conversation is logged and reviewed in the first weeks of deployment so we catch and fix edge cases fast.

Question 12

Why work with you instead of hiring an AI engineer ourselves?

Accepted Answer

An in-house AI engineer in the US costs $180,000 to $250,000 a year, plus 4 to 6 months to hire, plus they spend the first months figuring out your domain. We deliver a working agent in 3 weeks, on a budget that is 5 to 10x lower for the first year. If you want to take it in-house later, we hand it off cleanly.

Question 13

Where are you based? Who is behind iSendora?

Accepted Answer

iSendora is registered in Florida, USA. Founded by Den Senkevich and Artemiy Doronkin. We work remotely with clients across the US, EU, and the Middle East. Den leads engineering and AI architecture, Artemiy leads partnerships and operations.

Question 14

How long have you been building AI agents?

Accepted Answer

We have been building automation and AI systems for B2B clients for several years across SaaS, e-commerce, manufacturing, and digital marketing. iSendora was founded to focus exclusively on production-grade AI agents for businesses that need real outcomes, not demos.

Question 15

How big is your team?

Accepted Answer

Small and senior on purpose. Two co-founders plus a tight team of AI engineers and operations specialists. You will not be passed between account managers and a separate delivery team. The people on the kickoff call build your agent.

Question 16

Do you have case studies?

Accepted Answer

Yes. Our most public case is Defure, a Dubai-based furniture manufacturer for whom we built an end-to-end AI-driven operations system that handles orders, fulfillment coordination, and client communication. See the full case on the homepage. Additional cases are available on request, some are under NDA.

Question 17

What industries do you specialize in?

Accepted Answer

We work across industries where operations are complex, repetitive, and customer-facing: e-commerce, manufacturing, professional services, agencies, healthcare admin. We do not take projects in regulated areas where we cannot deliver good results (heavy financial services, primary medical care).

Question 18

What systems does the AI agent integrate with?

Accepted Answer

Anything with an API. CRM (HubSpot, Salesforce, Pipedrive). Messengers (WhatsApp, Telegram, Instagram, Facebook). Accounting (QuickBooks, Xero). Task managers (Asana, Trello, Jira, ClickUp, Monday). Email (Gmail, Outlook, SendGrid). Plus Google Sheets, Airtable, Notion, Slack. If you do not see your tool here, ask: we work with anything that exposes an API.

Question 19

Where is the agent hosted? Can we self-host?

Accepted Answer

Default: managed cloud on infrastructure we run for you (isolated environment, no shared compute with other clients). For enterprise customers we deploy on your cloud (AWS, GCP, Azure) or on-premises. Self-hosting is fully supported for projects with strict compliance needs.

Question 20

What happens when AI models get deprecated or replaced?

Accepted Answer

Models change every few months. We monitor the landscape and migrate your agent to the best available model when it improves cost or quality. You stay on a stable interface, the underlying model updates are transparent to your team. No re-implementation needed.

Question 21

Do you offer ongoing support and SLAs?

Accepted Answer

Yes. After go-live you get a maintenance plan: monitoring, bug fixes, model improvements, edge-case handling. SLAs depend on your tier. Standard response: critical issues within 4 hours business time, non-critical within 1 business day. Enterprise tier supports 24/7 paging.

Question 22

Do we need to change our processes to accommodate AI?

Accepted Answer

Sometimes yes. If your processes are chaotic (disconnected systems, no documented workflows, outdated data), AI will not help, it will only amplify the mess. We do two steps: clean up and connect processes first, then deploy the agent into a clean environment. That is how we guarantee results.

Question 23

How do you price token usage? Are there cost surprises?

Accepted Answer

No surprises. We bundle expected token costs into the monthly plan based on your usage profile. If your traffic spikes beyond the plan, we notify you before you hit the cap, and you decide whether to upgrade. We do not pass through raw token bills with markup.

Question 24

What is the cancellation policy?

Accepted Answer

Monthly contracts: cancel any time with 30 days notice. Annual contracts: cancel at the end of the term, prorated refund only if we materially fail to deliver. No long-term lock-in beyond what you sign for. You always own the agent code and data.

Question 25

Are there setup costs besides the deployment fee?

Accepted Answer

The deployment fee covers everything: audit, build, integrations, training, testing. Third-party costs (e.g., a new API subscription you need anyway) are passed through at cost, no markup, with your prior approval. No hidden line items at the end.

Question 26

Do you offer discounts for multiple agents?

Accepted Answer

Yes. After the first agent ships, additional agents on the same infrastructure cost less because the integration work is largely reused. Typical discount: 30 to 50 percent off setup fees for the second and subsequent agents.

Question 27

Where is our data stored geographically?

Accepted Answer

Default: US-based infrastructure. For EU clients we offer EU residency on request (data and processing within EU regions, no cross-border transfers). For Middle East clients we can provision regional hosting where the underlying providers offer it.

Question 28

Are you GDPR and CCPA compliant?

Accepted Answer

Yes. We support data subject rights (access, deletion, portability), data minimization, and lawful processing through DPAs with all our subprocessors. For EU clients we sign GDPR-compliant DPAs and provide records of processing on request.

Question 29

Do you have a data processing agreement (DPA)?

Accepted Answer

Yes. Standard DPA available on request, customizable for your jurisdiction. We sign DPAs before any production data is processed.

Question 30

What happens to our data after the contract ends?

Accepted Answer

On request we deliver everything to you (training data, logs, embeddings, exports of integrated systems data). After delivery and a 30-day grace period, we delete all your data from our systems. You can also request immediate deletion with no grace period if you prefer.

Question 31

What does the discovery call look like?

Accepted Answer

30 to 45 minutes. We ask about your business, your operations, the specific pain you want to solve, and what you have tried before. By the end we tell you whether we can help and what a project would look like. No sales pressure, if AI is not the right answer we will say so.

Question 32

How does the free audit work?

Accepted Answer

Yes, completely free. We audit one operational process of your choice, not your entire company. 1. Interview with a key decision-maker (founder, COO, ops lead), 1 to 2 hours. 2. Process mapping by an AI engineer plus an ops expert. 3. ROI calculation on your real numbers. 4. Custom proposal presentation. The output is yours regardless of whether you proceed with us.

Question 33

Who from our team needs to be involved?

Accepted Answer

For the audit: one decision-maker (founder, COO, ops lead) and access to the people closest to the process being audited. For implementation: one project sponsor on your side, plus 1 to 2 hours per week from the team using the process. We minimize the load on your people.

Question 34

How do you handle change requests during a project?

Accepted Answer

Small adjustments (copy tweaks, integration field changes) are part of the build. Bigger changes (new use cases, additional integrations) get scoped and priced separately so we do not blow the timeline. We tell you before doing the work, no surprise invoices.

Question 35

How do you prevent the agent from going off-script?

Accepted Answer

Multiple layers. System prompts that anchor the agent role and constraints. Tool-level permissions (the agent cannot take actions you have not authorized). Output validation against rules (no commitments above $X, no medical advice, etc). Real-time monitoring with alerts on anomalies. Most off-script behavior is caught by the validation layer before reaching the user.

Question 36

Can we see the agent reasoning and audit logs?

Accepted Answer

Yes. Every conversation is logged with the full reasoning chain, tool calls, and decisions made. You get a dashboard to inspect any interaction, search by client or topic, and replay sessions. Useful for compliance, training, and improving the agent over time.

Frequently asked questions