Why write about limits?
Most AI marketing describes what agents can do. Knowing what they can't do is what lets you deploy them safely. Every limit below is a design input for a stop rule.
Can AI agents guarantee correct answers?
No. Language models generate plausible text. Even with good source content they can misread a question, mix up details or state something confidently that isn't true. Grounding answers in your approved content, limiting the topics an agent answers and logging conversations all reduce the risk, but don't remove it. That's why high-stakes answers go through a person.
Can AI agents exercise judgement or take responsibility?
They can follow rules and weigh information you give them. They can't be accountable. When a customer disputes a charge or a patient describes symptoms, someone has to own the decision. Agents prepare; people decide.
Can AI agents fix a broken process?
No. If your team disagrees on how refunds work, an agent will apply that confusion faster. Automation amplifies whatever process it's given. Fix and agree the process first. Our consulting often spends most of its time here.
Can AI agents work with messy or missing data?
Only partly. If stock levels are out of date or customer records are duplicated, the agent will give wrong answers with confidence. Data quality is often the hidden first project.
Can AI agents handle every edge case?
No. There's always a message nobody anticipated: a forwarded voice note, a photo of a handwritten order, a customer who's grieving. Good design recognises when it doesn't know and hands over gracefully, instead of guessing.
Can AI agents build relationships and sell?
They can qualify, inform and follow up. Trust, negotiation and closing still depend on people, especially in high-value or relationship-driven sales like real estate or B2B.
Should AI agents take irreversible actions alone?
Generally not. Cancelling orders, issuing refunds, sending legal or financial statements, or deleting records should require approval. The cost of a wrong irreversible action outweighs the time saved.
So what are AI agents good at?
- Reading and categorising large volumes of messages
- Drafting replies from approved content
- Looking up and summarising information across systems
- Routing work to the right person with context
- Doing the same repetitive step correctly, every time, at any hour
Use the stop rule template to decide where each task belongs.
How do we apply this in client work?
Everything on this page feeds into how we run pilots. We start with a consultation to understand the process and the people involved, then write a one-page scope: the trigger, inputs, steps, outputs, a success measure and a stop rule that says what the agent may do alone, what needs approval and when it must hand over. We build inside the tools you already use, test on real or realistic cases including awkward ones, and review the results with your team before anyone talks about expanding.
We keep a few habits on every project: we never publish invented results, we label examples as illustrations, we list third-party costs separately, and we say so plainly when something isn't worth automating yet. If that sounds like the kind of partner you want, read the pilot process or message us on WhatsApp.
What should you prepare before talking to us?
You don't need a technical brief. A useful first conversation needs only a few things, and gathering them is valuable even if you never work with us:
- The process in plain words: what triggers it, who does what, and where it usually goes wrong.
- Rough volumes: how many messages, orders, tickets or invoices per week. A one-week tally is enough.
- Your tools: WhatsApp app or API, CRM, sheets, accounting or helpdesk software, including versions or plans where you know them.
- Three real examples: typical messages or cases, plus one awkward one, with personal details removed.
- What must never happen: the mistakes that would hurt most. These become the first lines of your stop rule.
- An owner: the person who will review the workflow once it runs.
Send these on WhatsApp or by email, or run the pilot scorecard first.
How does a pilot with us work?
Every engagement follows the same four steps. Consult: you walk us through the process, the tools and where it breaks. Pick the pilot: we choose one workflow and write down its inputs, outputs, tools, stop rule and success measure. Ship: we build it into the tools you already use and test it on real cases. Review: your team checks what the agent did and what it escalated, then decides whether to tune it, expand it or stop. Read the full pilot process.