People who search for Amelia conversational AI are usually weighing an enterprise virtual agent platform against something lighter. This article does not review that product or describe its features. It answers a different question: what does a firm of 10 to 250 people actually need from conversational AI, what should it check when comparing enterprise platforms with an agent built into its CRM, and how do you set up the CRM option on the levelpartner.ai platform?
The question behind the search
Enterprise conversational AI platforms are built for large organisations: many departments, deep integrations with internal systems, dedicated teams to design and maintain virtual agents. A mid-sized firm has a different problem. It needs to answer leads and clients quickly on a handful of channels, book meetings and keep records in one place, without hiring a team to run the AI.
So the first question is not which platform is most capable. It is which platform matches your scale, your channels and the people you have to maintain it.
What to check when comparing options
Use these questions with any vendor, including us:
- Who builds and maintains the agent day to day, and how many hours a week does that take?
- Does the agent write to your CRM directly, or does it need an integration project?
- Which channels are covered out of the box: web chat, SMS, WhatsApp, social, phone?
- Can the agent book into your calendars and update your pipeline?
- How are hand-offs to people handled, and where does your team reply?
- How is usage billed, and what does a typical month cost at your volume?
- What does it take to change an answer: a knowledge edit, or a developer?
If an option needs a project team to answer those questions, it is built for a larger organisation than yours.
Before you start on the platform
- A clear first job for the agent, such as answering new enquiries or client logistics.
- Your channels connected to your account.
- A knowledge base drawn from your current FAQs, service pages and process notes.
- A calendar and pipeline the agent can use.
- A named owner with an hour a week for upkeep.
How to set it up
- Go to AI Agents > Conversation AI and click + Create Bot.
- Select Guided Form Setup.
- Click Create New Bot.
- Choose General Q&A or Appointment Booking for its first job.
- Select or create a Brand Voice.
- Configure Bot Settings.
- Configure Bot Goals.
- Save the bot.
- Choose Suggestive mode for the first week, then Auto-Pilot.
The trade-off between enterprise and built-in
An enterprise platform offers depth: complex integrations, fine-grained conversation design and support for many business units. You pay for that depth in set-up time, specialist staff and ongoing cost. A CRM-native agent offers less depth and much faster time to value, with the agent's work landing directly in your contacts, pipeline and calendars.
For most firms in the four industries we serve, agencies, consultants, mortgage brokers and real estate, the CRM-native route covers the actual need. The enterprise route makes sense when the AI must work across many internal systems that are not your CRM.
Say route one is an enterprise virtual agent with a three-month build and an integration to the CRM, and route two is an agent inside the CRM platform, live on web chat in a week and on the phone in a month. If the firm's need is enquiry handling and booking on four channels, route two meets it; route one offers capability the firm would not use.
Time to run it
A CRM-native agent takes a day or two to build and about an hour a week to maintain. Treat any vendor's estimate of upkeep as a serious selection criterion; an agent nobody maintains loses accuracy within weeks.
Buyers compare feature lists and ignore who will run the thing. The most capable platform with no owner performs worse than a simpler agent someone reviews every week.
When the built-in route is not enough
If you need the AI to act in several internal systems, such as a case management system and a document platform, or you serve many departments with separate processes, an enterprise platform may justify its cost. For context on the wider market, see enterprise conversational AI and the conversational AI magic quadrant article. For how agents work on the platform, read conversational AI.
Troubleshooting
The agent cannot answer account-specific questions. That data lives outside the CRM. Bring key fields into contact records or keep those questions with people.
Stakeholders expect enterprise-style analytics. Set up the Reporting dashboard with the four or five measures that matter and agree them upfront.
The pilot stalls. The first job was too broad. Narrow it to one channel and one outcome.
Answers drift over time. No one owns upkeep. Name an owner and book their weekly hour.
Frequently asked questions
Does this article compare features with Amelia?
No. We do not describe or rate another vendor's product. Use the checklist above with each vendor you are considering and ask for answers in writing.
Is a CRM-native agent less secure than an enterprise platform?
Security depends on configuration and data handling, not on size alone. Ask every vendor how conversation data is stored, who can access it and how long it is kept.
Can we start small and move to an enterprise platform later?
Yes. Your knowledge base, hand-off rules and test scripts carry over to any platform. Starting small also shows which capabilities you actually need.
What does it cost on the levelpartner.ai platform?
The licence covers AI agents with every other feature. Messaging, calls and AI usage are billed as used.