re:solved, AI automation agency
← Back to blog
Cătălin Andrei Bora, founder of re:solved

By · Founder, re:solved

·8 min read

AI Agent Examples: What They Actually Do for a Small Business

Built examples and scenarios for companies with more than 25 employees. Reactivation and intake are built. Bookings and invoice follow-up are offers.

A small business team at desks with translucent AI agent figures working alongside them on chats, scheduling and content.

GreenScope, our own carbon reporting company, used an agent we built for research, writing, image generation, LinkedIn publishing and a report after every post. It has not been running since September. This is use in our own company, not a client result.

These examples distinguish what we have built from what we could build. Large-company results are not promises for your business.

We distinguish the built Ornella Design and Multidiet examples from GreenScope, our own company, and scenarios not yet built.


Customer-facing work: catching what you currently miss

At Ornella Design, a print shop in Timișoara, the agent reactivates past customers on WhatsApp, answers questions and prepares quick estimates from the shop’s price rules. When a customer wants to order, it hands over to a salesperson. Stop requests are honoured automatically.

A smartphone with glowing ribbons of light flowing out of it, representing messages, calendar events and tasks handled automatically.
An agent quietly handles the work that used to sit unanswered in your inbox or message queue.

At Multidiet, Kara responds on WhatsApp, 24/7. It explains the programme, asks screening questions and checks required lab tests are present without interpreting values. It routes to the app or practice and hands practice cases to staff. We report counts only, never individual patient details.

What we could build: an agent for booking requests, confirmations, reminders and reschedules, with clear human handover rules. We have not built this agent for a client yet.


Sales work: stop losing leads to slow replies

What we could build: an agent could follow up on unanswered enquiries and quotes under rules agreed with you. Negotiations go to a person. We have no measured commercial result for this scenario.

What we could build: an agent that prepares a conversation summary and CRM information. A person checks what is saved. We have no client result or measured time saving for this scenario.


Marketing work: keeping the lights on when you're busy

At GreenScope, our own company, the agent researched topics, wrote, generated images, published on LinkedIn on a schedule and sent a report after each post. It has not been running since September. This is use in our own company, not a client result.

The weekly industry roundup. An agent monitors a set of news sources through the week, then on a fixed day pulls the most relevant developments together, summarises them in plain language, and publishes or emails the digest. It's the kind of consistent, useful presence that's almost impossible to maintain by hand. And trivial for an agent to keep up forever.


Internal work: clearing the admin pile

What we could build: an agent for message triage, with written limits and human handover. Without measurement we do not promise time saved.

Logging the things that always slip. Quotes, orders, enquiries buried in email threads. The agent reads them, extracts the important details, and logs them in one place so nothing gets lost and you can actually find it later.

Hypothetical scenario: an agent could prepare answers about the business from approved data, showing its sources and requiring human checks. It is not a client result.


Use in our own company: GreenScope

GreenScope, our own carbon reporting company, used an agent we built for research, writing, image generation, LinkedIn publishing and a report after every post. It has not been running since September. This is use in our own company, not a client result.

The best automation doesn't feel like automation. It feels like someone reliable finally joined the team.

That's the standard we hold an agent to. If we're going to tell you one can run part of your operation, we should be running ours on one first.


What an AI agent won't do

It's worth being honest, because the examples above can make it sound like agents do everything. They don't.

An agent is brilliant at work that's repetitive, predictable, and happens over and over. The follow-ups, the scheduling, the reminders, the reporting. It's the wrong tool for work that needs real judgment, a genuine human relationship, creative strategy, or physical presence. A good provider tells you which is which, instead of selling you an agent for a problem an agent can't solve. We'd rather lose the sale than oversell.


What it costs

An agent takes over a defined process. A person remains responsible for boundaries, exceptions and checking the result. We do not use salary comparisons to justify the price.

The 30 minute conversation is free. The specification costs 400 EUR, credited against the build if you sign within 60 days and free with an existing DIGIVEST strategy or feasibility study. A build starts at 1,500 EUR, fixed from the specification. Monthly operation for agents connected to business systems starts at 250 EUR.


Who this is for (and who it isn't)

A good fit:

  • Businesses with more than 25 people losing real hours to repetitive, manual work.
  • Owners who hired someone just to handle follow-ups, bookings, or outreach. And it's still inconsistent.
  • Industries where client data is sensitive (healthcare, legal, finance), where running the agent on your own infrastructure is a genuine advantage.

Not a fit:

  • Solo operators with no client base to re-engage and no operational load to lift.
  • Anyone after a website FAQ widget. This is operational work, not a chatbot in the corner of a page.
  • Companies that need enterprise procurement and large projects before defining one process.

Frequently asked questions

Which of these examples should I start with?

Start with whatever is costing you the most time or the most lost revenue right now. Usually that's lead follow-up, bookings, or content. One role, done well, beats trying to automate everything at once.

Can one agent do several of these jobs?

We start with one process and add others only after the first works.

Will it work with the tools I already use?

That's the goal. Agents connect to the tools you already work in. WhatsApp, email, calendars, CRMs, Drive. So nothing changes in your team's daily routine.

How fast can one be running?

The start date is agreed in the specification after checking the process and connected systems.


Want to know which of these fits your business?

Book a 30 minute conversation. In 30 minutes we look at one process that costs you time. We tell you honestly whether an agent can take it over and what it would cost. No sales presentation.