Vendasta
AI Employees for Agencies: What They Are and How to Sell Them
An AI employee takes a whole job, like answering calls or replying to reviews. Brendan King, CEO of Vendasta, on how agencies sell AI employees to SMBs.
The short answer. An AI employee is software that takes a whole job, such as answering the phone or replying to reviews, and does it using a business's own data, with a person approving the results. AI employees for agencies are those workers packaged so an agency can sell them to small business clients and manage them for those clients. The agency brings the expertise and the relationship. The AI does the work.
I am the CEO and co-founder of Vendasta. I sold computers starting in 1989, moved into SaaS in 1999 and started Vendasta in 2008. Vendasta builds software for local businesses and the partners who serve them. Today that means an AI workforce for small business. Here is how I define AI employees, and where I think agencies fit.
What is an AI employee?
An AI employee holds a job. A receptionist answers the phone, books the appointment, updates the record and sends the confirmation. A review specialist reads each new review and replies. A content creator writes and posts. A support agent works the tickets. A sales coach helps the team close.
Every one of these is a package. The package is what a business owner understands, because it matches a job they already know how to hire for. Underneath it sit three parts that I wrote about last October:
- Knowledge. The business's own data. Reviews, contacts, products, tickets. It has to stay current without a person maintaining it.
- Capabilities. The skills. "I can respond to a review." "I can work out what a customer wants." "I can book an appointment."
- Tools. The actions that carry a skill out. Look up the review. Post the reply. Check the calendar. Create the booking.
Build each skill once and every employee that needs it gets it. The employee is what the owner hires. The skills are what the platform owns. The context is the moat: anyone can generate code now, and nobody else can generate your context.
How is an AI employee different from a chatbot or an AI tool?
A chatbot answers when asked. An AI employee is assigned a job and owns the outcome.
I put it this way in Ask Is a Habit. Do Is a Decision.:
| AI tool or chatbot | AI employee | |
|---|---|---|
| AI is | An oracle you consult | A worker you assign |
| The unit of work | The answer | The outcome |
| Who starts it | A person, by typing | An event, such as a missed call |
| Where it lives | A chat tab beside the workflow | Inside the workflow, with the data and permission to act |
| How you measure it | Time saved | Work completed without a human touch |
| The ceiling | The number of humans | Compute |
There is a quick test. If the default button on an AI employee is Chat, it is a persona on a chatbot. A real AI employee takes a job and runs with it.
Take an owner who asks a chatbot for a marketing plan. They get a nice document and a to-do list they will never finish. Take an owner who assigns an AI receptionist to the inbound calls. They get booked jobs while they are on a roof. Same technology. Only one of them changes the business.
In my Evolution of AI framework there are five rungs of autonomy: Macro Recorder, Copilot, Coworker, Agent and Autopilot Employee. A chatbot helps a person do the work. An AI employee does the work and a person checks it.
The job also needs a trigger. If a person has to start every job, AI is waiting to be asked. I wrote the schema for this in AI Should Start the Job.
Why agencies?
Most small businesses will never build software. A small share of owners like technology and will happily do it themselves. Most run a business and want someone they trust to do it for them. That majority is the market, and selling tools to people who do not want tools is a losing game. Sell the outcome and let a trusted partner deliver it.
Agencies are that trusted partner. They already hold the relationship, and they know the local plumber, the dental office and the franchise network. A technology company brings horizontal technology. The agency brings vertical expertise and a human in the loop. Neither wins alone.
Agencies also share the job list of the businesses they serve. Capture leads. Respond fast. Book meetings. Run campaigns. Manage reputation. Post on social. Keep the website current. Support customers. At Vendasta we have the same list, and we run our own company on what we sell. An agency can do the same: put AI employees on its own work first, then offer them to clients.
What jobs can an AI employee do for an SMB?
For a local business, there is no support department, no sales team and no marketing team. There is one conversation, and whoever answers it owns the customer. I wrote about it in The One Conversation. A phone call, a chat, a form or an email, and most of them go unanswered.
AI employees can take the jobs around that conversation:
- Answer and book. Pick up the phone or the chat, qualify the caller, book the appointment, update the record and send the confirmation.
- Capture leads. When a visitor shows buying intent, take a name and one way to reach them, create the contact and tell them a person will follow up.
- Handle reviews. Ask happy customers for a review. Reply to the ones that come in.
- Support customers. Answer the common questions and hand the hard ones to a person.
- Follow up. Send the reminder, the next campaign message or the thank-you.
- Create content. Draft and publish the social post or the page update.
These jobs need hands and feet as well as intelligence. An AI that cannot message, publish, book or update the record is a consultant.
How should an agency price and sell AI employees?
I will leave specific prices out of this. They change, and they depend on the agency's market. Here is how I think about the selling.
Sell the job. A business owner does not want to buy "review response capability." They want someone to handle their reviews. Name the employee after the job, and show the owner the finished work.
Let the work show up done. Small business owners want proactive work, ready to approve. The job arrives finished and the owner says yes.
Give them certainty. Small business buyers want low cost and no surprise bills. They do not want to watch a meter.
Report in the owner's terms. We measure success in leads, bookings and revenue for our customers. An agency should report those, and leave prompts and hours out of it.
Configure once, then repeat. Set up an account one time and push it to a group of similar businesses, such as a vertical, a brand or a franchise network.
Partners also matter here. An AI employee will sometimes hit something it cannot handle. A trusted agency gives the owner a human to turn to. Most AI vendors cannot offer that.
Where do humans fit?
Humans orchestrate. AI does the work.
That does not mean humans disappear. Their job changes from doing to directing: setting strategy, training the AI employees and checking their work. At an agency that means the account manager stops being the person who posts, replies and books. They become the person who owns the outcome for the client and runs the AI employees that deliver it.
Three rules govern the handoff.
- Approval is a dial, not a gate. Each job gets a level. Do it. Do it and tell me. Ask me first. A job earns more freedom over time.
- External messages start as drafts. A sent message cannot be recalled. Keep a person on that door until the AI employee has earned it. Reads and drafts are cheap to undo. Sends, deletes and bulk changes are not.
- Memory is the job. A correction is remembered once and applied forever. Every time a person steps in, that fix becomes a rule the employee keeps.
The person checks at the end of the loop, where judgment is worth the most. They approve, edit or reject the finished work.
I have not tested this in every kind of business. A regulated business will put more steps behind approval, and it should. The checkpoint belongs at the end of the loop, and how hard the end is to reverse decides how heavy it needs to be.
The full list is on my AI Principles page, with the longer set in Our Guiding Principles. The first principle covers everything above: AI does the work. Humans orchestrate. You can read more about me and Vendasta on the About page.
FAQ
What are AI employees for agencies?
AI employees for agencies are role-based AI workers that each own a defined job, such as answering calls, replying to reviews or following up on leads. The agency packages them, sells them to small business clients and oversees the work. The agency brings expertise and the client relationship. The AI does the job.
What is the difference between an AI employee and a chatbot?
A chatbot answers questions when a person asks. An AI employee is assigned a job and owns the outcome. It works from the business's own data, takes action in connected systems, and a person approves the finished work. If the default button is Chat, it is a chatbot with a persona.
How do agencies sell AI employees to small businesses?
Sell the job. Owners want someone to handle their calls or their reviews. They do not want a list of capabilities. Show finished work, keep costs predictable and report in the owner's terms: leads, bookings and revenue.
Will AI employees replace the people at an agency?
AI employees take over the doing. The people move to orchestrating: setting direction, training the AI, checking results and handling the cases the AI hands back. The agency's relationship and expertise stay human.
Who stays in control of an AI employee?
A person does. Approval is a dial with three settings: do it, do it and tell me, or ask me first. Anything that goes outside the business starts as a draft until the AI employee earns more freedom. The person checks the finished work at the end of the job.
Continue the conversation
Have a perspective to share?