AI implementation and operations

AI for real work.

We integrate AI into the tools your business already uses, implement a concrete process, train your team and support daily operations.

An implementation example

Every request has an owner.

An assisted inbox for document requests.

  1. A request arrives

    It is logged and assigned to an owner.

  2. AI prepares a draft

    It uses approved information and identifies its sources.

  3. Your team reviews and replies

    It checks the content and keeps pending requests visible.

Service example. We assess suitability before implementation.

From a recurring task to a process you can track.

01 / One concrete process

AI implementation and integration

We configure agents and automations in your tools to log requests, prepare information and track work. We agree which checks remain with your team.

02 / Agreed ongoing service

Monthly operations with an agreed scope

Periodic reviews, incidents, sources and usage costs. We define in writing what we handle, how often and during which hours.

Explore operations

Discovery and scope

We review how the work is done today, its exceptions and what your tools already handle. We define scope, price and acceptance before starting.

Team training

Practical sessions on the agreed work, with examples, a user guide and time for questions. Your team learns to check results and handle exceptions.

Usage rules and oversight

We define owners, access, approved sources and exception handling. We document how the process is reviewed and when a person needs to intervene.

A clear process from the first step.

01Scope

Understand and agree

We review a work sample, the tools and the exceptions. We agree owners, price and acceptance tests.

02Getting started

Implement and train

We configure the connections, test the full process with authorised data and train the team that will use it.

03Agreed ongoing service

Review and operate

Under a monthly service agreement, we review results, incidents and usage costs. Extensions are agreed before they are delivered.

«AI proposes. Your team decides.»

Human review

We define what a person validates and what happens when a result needs review.

Acceptance criteria

The sample, valid outcome and exceptions are agreed before implementation.

Environment and access

We review permissions, providers and data handling for each client’s environment.

An ERCEL reference

García Baquero, an ERCEL client.

García Baquero is part of our client portfolio.

View reference
Assessment tools

Review is part of implementation.

Testing helps identify failures and define usage limits. Alongside implementation work, ERCEL provides tools for assessing AI systems.

ERCEL Audit

Tools to explore technical risks and organise assessment results. Review their methodology and scope to understand what they check.

EU AI ActOWASP LLM Top 10ISO 42001NIST AI 600-1

Show us how it works today.

Tell us which task repeats, who handles it and which tools they use. In an initial conversation, we’ll assess suitability and the next step.

Scope, price and acceptance are agreed before work begins.

664 24 13 28About usAn initial conversation to understand your process.