Skip to content
Maor KertsmanMaor Kertsman

AI Assistants & Integrations

Assistants on your data, with a person in the loop.

Illustration: asking an assistant about today's stuck leads, a structured answer by owner, and a person who checks it.

Where AI actually helps

AI is useful where the input is messy: a voice message, a free-text inquiry, a long thread someone needs summarised, a task hidden in a conversation. It's not useful for steps that have a clear rule, where deterministic automation is cheaper and more predictable, and it shouldn't take decisions that carry commercial or personal weight. I build assistants connected to your CRM and business data through tools, with explicit rules for what they do on their own, what they propose, and what waits for a person. AI lead warming and personalised landing pages are presented here as concepts, until I can show delivered work.

Example workflows

Structured intake through conversation

  1. A customer writes or sends a voice message
  2. The assistant understands the intent and asks for what's missing
  3. Structured fields are written to the CRM
  4. A person reviews anything ambiguous before it moves on

Operational summaries and task extraction

  1. Calls, threads and notes from the day are collected
  2. The assistant drafts a short summary and lists proposed tasks
  3. The owner accepts, edits or discards each proposal
  4. Accepted tasks are created in the CRM with the source attached

AI lead warming over WhatsApp (concept)

  1. A new lead receives a guided conversation in the business's tone
  2. Qualifying answers and voice messages are understood and recorded
  3. Follow-up continues under clear stop rules
  4. A person takes over with the summary at the agreed moment

What an engagement can include

  • Assistants connected to your CRM and business data through tools
  • Intent handling and structured task extraction
  • Summaries of calls, conversations and daily activity
  • Internal questions answered from your business information
  • Explicit rules for what is automated, proposed, or decided by a person
  • Logging and review of every assistant action
  • Deterministic automation wherever a rule does the job better

Connected tools

Fireberry, WhatsApp Business Platform, Google Sheets, Make, Webhooks and custom APIs

Tool names identify compatible tools and do not imply endorsement.

Bring the messy input

Tell me where information arrives unstructured and someone has to make sense of it. We'll see whether an assistant, a rule or a person is the right answer.