TeckelApplied AI firm

A Teckel campaign

Campaign piece: a mall totem with the headline Less busywork. More business. and a long receipt spilling out of it.
Campaign piece. Generated composition; not a real installation.

How much work does your team do before it can do its work?

Copying the same data into two systems. Chasing a signature. Rebuilding Monday’s report. Your team knows this work by heart.

Start with one process that needs to move. See what changes

Illustrative tasks. Your operation’s list is something we map together.

The work around the workAny given week
  1. Copy the order into the spreadsheetx 14
  2. Ask on WhatsApp if it arrivedx 23
  3. Chase a signaturex 6
  4. Find the email with the attachmentx 9
  5. Type the same thing into the systemx 14
  6. Tell the client it shippedx 17
  7. Rebuild Monday’s reportx 1
  8. Check whether someone already checkedx 11
  9. Copy the order into the spreadsheetx 14
  10. Ask on WhatsApp if it arrivedx 23
  11. Chase a signaturex 6
  12. Find the email with the attachmentx 9
TotalHours nobody bills

Which task keeps coming back?

A request arrives. Someone copies the details, asks for the missing information and follows up again. Before automating anything, we map what repeats and where someone has to decide.

The first step is to identify one bounded process, its owner and the result worth measuring.

This is what changes when the system keeps the record.

A request comes in through the channel your team already uses. The system gathers the context, does only what it is allowed to do, and writes it down. When something falls outside its rules, it asks a person.

TimeInputDecisionWhoResult
09:14“Has 2291 cleared yet?”, from the client on WhatsApp.Look up operation 2291 in the record and reply with its status.SystemAt customs since 08:50. Reply sent to the client.
09:16Note from the account executive: a document is missing on 2310.Set the status to “awaiting document” and notify the owner.SystemStatus updated. Karla M. notified.
09:31“Change its status to released”, from a number that is not in the catalog.Not in the catalog. Not classified. A person is asked.Karla M.Karla confirms it is the client. Status: released. Author and time go in the log.
09:40The client asks for the receipt.Attach the receipt already on file.SystemSent. Nothing is rewritten; everything is appended.

Example with synthetic data, modeled on Grupo ILT’s system. Not a client screenshot.

What is not in the catalog does not get classified: the system asks. That is a design decision in Grupo ILT’s system, not a universal promise. The decisions behind the system

One operation. A shared status.

Before, the status of each operation lived in lists spread across several boards, and progress depended on where a card had been placed. Today each operation has one of eight statuses, and most updates are recorded by the system, not typed by a person.

Client
Grupo ILT
Business
Customs brokerage, Mexico
Work
Operations tracking over WhatsApp with one shared record
In operation since
July 2026
Evidence reviewed
September 21, 2026
Build
8 active days
Before

29

lists across 4 boards

Today

8

statuses in one record

4 of every 5 updates are written by the system.

Diagram of the structural change. Each thin rule is a list; each stretch of the record, a status.

Records H-080 · H-087 · H-089 · H-090 · H-114

No operation counts, by written condition of Grupo ILT. The comparison describes; it does not estimate a causal effect.

Read the full case file

What needs to work better?

Bring a process, a pilot or a question. You talk directly with the people who will build it.

Book a conversation

Opens Calendly in a new tab.

george@teckel-ai.com
Prepare your call with AVA

AVA asks a few questions so you arrive with your case in order.