Service
AI Agents and Intake
Answering, qualifying and booking, on your own documents and your own rules, at the hours you are not there.
An agent is worth building when the same forty questions arrive every week, when calls go unanswered after 5pm, or when qualifying an enquiry takes ten minutes of somebody's time and half of them do not qualify. It is not worth building because the website looks dated without one.
Statistics Canada put virtual agents and chatbots at 28.2% of Canadian AI use in the second quarter of 2026, third behind data analytics at 36.6% and text analytics at 34.5%. It is a mature use, which also means the failures are well understood.
Chatbot, agent, or neither
| You want | What it is | Rough cost |
|---|---|---|
| Answers to known questions | A retrieval bot grounded in your own documents. It quotes your pages, it does not invent. | $4,000 to $7,000 |
| Qualify and book | An agent that asks, scores against your rules, and writes into your calendar or CRM. | $7,000 to $12,000 |
| Handle an existing process end to end | An agent with tools, permissions and an approval step where money or commitments are involved. | $12,000 to $15,000+ |
| A box that says hello | Nothing. Use the money on [workflow automation](/services/workflow-automation/). | $0 |
The distinction matters more than the price does, and it is set out at length in the note on the difference between an AI chatbot and an AI agent.
Grounding, and why it is the whole job
An agent that answers from a general model will answer confidently about your return policy, your service area and your prices, and it will be wrong at a rate you cannot predict. An agent grounded in your own documents answers from the text you gave it and says it does not know when the text does not cover it.
Building that grounding is most of the cost. It means finding out what your policies actually are, which is frequently the first time anybody has written them down in one place, and it means a rule about what happens when the answer is not there. Handing off to a person with the conversation attached is a feature, not a failure.
The privacy question, before the features question
An intake agent collects personal information by design. The Office of the Privacy Commissioner of Canada found in May 2026 that OpenAI's initial training of ChatGPT did not comply with Canadian privacy law, and its guidance treats prompts, retrieval corpora and outputs as personal information handling like any other. What that rules in and out for an Ontario business is covered in what PIPEDA means for your AI tools.
In practice: a public-facing agent that collects a name, a phone number and a description of a problem needs a stated purpose, a retention rule, and a processor that will sign an agreement. That is achievable and it is not free, and it belongs in the quote rather than in a surprise later.
What handover includes
- The agent running on accounts in your name, with the model provider billed to you.
- The document set it is grounded in, in a form your staff can edit without calling me.
- A transcript log you own, with a retention period you chose.
- An escalation path to a human, and a rule for when it fires.
- A monthly running cost you saw before you agreed to build, usually $80 to $600 depending on volume.
Before you spend anything
Tell me how many people work there, what the busiest hour of the week looks like, and which task everybody complains about. That is usually enough to say on a first call whether an assessment is worth your $999 or whether you have one obvious problem that needs one obvious fix.
Reply within one business day ยท Mon to Fri, 9am to 5pm Eastern