From AI tools to better processes: five changes your business should act on now

The core idea

For years, adopting artificial intelligence often meant opening a separate tool and asking it to help with a task. That is changing. AI is moving into WhatsApp and Instagram, Google Workspace, Google Ads and Analytics. At the same time, new rules are raising transparency expectations, while digital channels continue to gain weight in Mexico.

For a growing company, the right question is no longer which AI tool to buy. It is which process to improve, what information the system may use, who reviews the outcome, and how the business will know it works. AI adoption is becoming common; operational integration is becoming the differentiator.

1. Customer conversations are becoming an AI-assisted process

Meta introduced Business Agent on June 3, 2026 and announced its expansion to businesses of different sizes across WhatsApp, Messenger and Instagram. According to Meta, it can answer business-specific questions, recommend products, qualify leads, book appointments, hand conversations to a person and support sales conversion.

That can help businesses receiving repetitive questions after hours. But a fast answer creates little value if it uses an outdated price, makes a promise the company cannot fulfill, or fails to bring in a person at the right moment.

• Choose one pilot use case, such as common questions or initial qualification.
• Create an approved source for prices, availability, hours, service areas, policies and delivery times.
• Define handoff rules for custom quotes, complaints, sensitive data, negotiation and uncertainty.
• Track response time, qualified leads, errors, handoffs, appointments and sales.

2. Automation can now follow a team’s internal standards

On August 26, Google announced new Workspace Studio features. Teams can create reusable skills that use their own rules, templates and reference files across Gmail, Docs, Slides, Drive and Chat. Google also expanded controls for agent access management and data loss prevention.

The practical opportunity is not simply to have an agent. It is to turn a good way of working into a repeatable flow: meeting notes, opportunity summary, next steps, follow-up draft and owner reminder.

• Document the current flow and its exceptions before automating it.
• Keep human approval before sending messages or changing customer records.
• Measure time to follow-up, missing information, correction rate and reply rate.
• Use least-privilege access and retain an action log.

3. Google Ads will automate more campaign decisions starting in September

Google says automatic upgrades to AI Max will begin in September 2026 for campaigns using automatically created assets and campaign-level broad match. AI Max adds controls for brands, locations, text, URL exclusions and reporting, while expanding automation across queries, ad text and landing-page behavior.

As the system makes more decisions, signal quality matters more. A campaign cannot reliably distinguish a strong prospect from an irrelevant one when conversion goals are wrong or the sales team never feeds lead quality back into the process.

• Identify campaigns using automatically created assets or campaign-level broad match.
• Review conversion goals, locations, brands, search terms, URLs and eligible landing pages.
• Save a baseline for spend, qualified leads, acquisition cost and revenue.
• Use a controlled test or short monitoring window before increasing budget.

4. AI transparency is now a concrete obligation in the European Union

Article 50 of the EU AI Act has applied since August 2, 2026. European Commission guidance includes duties to inform people when they interact directly with certain AI systems, along with marking or labeling requirements for specific types of generated or manipulated content.

A small company in Mexico or the United States should still examine the scope if it serves European customers, publishes content for that market, or uses AI in an experience that may fall within the rules. Applicability depends on the business’s role and use case, so real exposure calls for legal review.

• Inventory chatbots, generated content, assisted calls and customer-facing automation.
• Record provider, purpose, data used, owner, human review and disclosure method.
• Do not design an automated interaction to impersonate a human.

5. Digital selling in Mexico is growing faster than retail overall

INEGI data published on August 21 show that real retail revenue grew 2.0% year over year in June 2026. Within the report, the category for sales exclusively through the internet, catalogs, television and similar channels increased 23.2% year over year. The category is broader than pure e-commerce, but the gap confirms the rising weight of remote channels.

More digital demand can also create more operational disorder: ownerless messages, inconsistent quotes, abandoned carts, unreconciled payments and duplicate follow-ups.

• Map Instagram, WhatsApp, web and referrals through quote, payment, fulfillment and follow-up.
• Assign an owner and response-time target for each entry point.
• Track conversion, abandonment and repeat purchase by channel—not only reach or message volume.

A practical 30-day agenda

Choose one process with visible commercial impact and work in a controlled way before scaling.

• Document the current process, owners, exceptions and recurring failures.
• Create an approved source of truth for pricing, offer, policies and responses.
• Define permissions, human handoffs, prohibited actions and rollback.
• Record a baseline and run a small two-week pilot.
• Review errors, time saved, customer quality and commercial outcomes before scaling.

Merluca’s role

Merluca is not a CRM or SaaS product. It is a professional service that helps growing businesses design, document, connect and improve the sales and operating processes behind the tools they already use.