7 practical ways how companies use AI to work in 2026 (or at least can use it)

It is now 3 years since OpenAI launched ChatGPT. Much has happened since. Huge investments and lots of experiments. Leaders are no longer asking if they should use AI, but how they can use it. 

Surveys from McKinsey, Google, BCG show that organizations in 2025 are using AI, but still in early stages of scaling AI and capturing value across the organization.

Here are 7 practical ways how companies use AI to work in 2026.

1. Build an effective AI strategy

It can be one page long. It can be 30 pages. Or 100 pages. It doesn't matter that much. What matters is that you have an AI strategy in place, with its key components: 1) vision for AI in your company, 2) identification and prioritization process for use cases and 3) that you actually implement initiatives and iterate on them. 

This is the backbone of everything. You need to plan how your business should look with AI to avoid standalone projects, but make it part of your overall corporate strategy.

It is a bit boring perhaps. Not the most flashy practical way to use AI in your business. But if you make a good AI strategy for 2026, you will thank us later. Read a simple guide to creating an AI strategy in 2026. 

By the end of this, you should be able to point to one document that explains why you use AI, where you use it, and how you decide what comes next.

2. Categorize AI initiatives into three buckets

Let’s continue with something more concrete. Categorize your AI initiatives into three buckets: 

1. The basics

2. Intermediate (GPTs, Gems, etc.) 

3. Advanced (AI workflows, agentic AI)

By doing this you ensure that you execute the easy initiatives with a high level of impact first. It also ensures to keep track of the other initiatives. Also, the purpose is not to make a super long list, but finding the best AI use cases for your business. 

In continuation, I will show an example of each.

3. Create a prompt library for your company

This is a classic “basics” initiative. Low risk, high leverage. It is so simple that you should do it now. Let me explain.

The purpose of a prompt library is to save your prompts in a way that you can find them later, share them with colleagues and rate them of their efficiency. The last point, to ensure that you dont drown in a list of prompts, but keep a clean list of prompts that has proven effective to solve its task.

Now, there are several ways to do this. The most simple is to set up a shared document, with the prompts. Alternatively, if you are like us at BRACAI with a fondness for spreadsheets, then you could set up a shared Google Sheets. 

For larger organizations, you could set up governance structures, ie. most people can add a prompt through a Form, but access the prompt list. You could set up an AI task force with a small group of people responsible for assessing and optimizing the prompts, to ensure that your prompt library is up-to-date. 

Setting this up takes less than 10 minutes and immediately creates shared leverage across the team.

If you want a head start, we’ve created two simple templates we use ourselves at BRACAI.

Download the templates:

– Prompt library. Google Doc

– Prompt library. Google Sheets

You can copy them and adapt them to your company in minutes.

4. Build GPT workflows

This is an example of a semi AI initiative. Not basic. Not heavy engineering either.

--> It is simple enough to build in-house

--> Powerful enough to change how work gets done

--> The most realistic path toward AI workflows

You are probably familiar with GPTs by now (or Gems, Claude Projects).

A GPT is:

• A customized ChatGPT

• Optimized for one specific task

• Using a fixed prompt, structure, and documents

Think of it as: a repeatable micro-tool, not a chatbot.

One of our most successful trainings this year has been “create your own personalized GPT”, with an average 9.58 / 10 rating. Here we teach participants how to create their own GPTs in 4 hours and how they can use them to replace parts of the workflow. No code. No engineering team.

Example: Workflow redesign with GPTs

Most teams already run the same meeting workflow over and over again. The problem is not the meeting itself. It is everything around it. Done manually, this workflow is slow, inconsistent, and varies a lot from person to person.

Below is an example of the benefits of a workflow redesign with GPTs.

Notice that the meeting itself does not change. What changes is the work before and after. The result is not automation for its own sake, but a faster, more consistent way of working that saves time on every single meeting.

5. Replace your first process with an AI workflow

GPT workflows are a strong step forward. AI workflows are the next one.

We do not recommend jumping here immediately. But for most processes, this should eventually be the end goal.

Everything shown in the meeting example above can be taken one step further and run automatically.

What changes with AI workflows

So far, GPT workflows still require a human to:

1. Open ChatGPT

2. Paste inputs

3. Run each step

With AI workflows, this changes.

An AI workflow:

• Runs across multiple tools

• Follows a predefined logic

• Starts from a trigger

• Completes the process end to end

No copy-pasting. No manual handoffs.

Example: Fully automated meeting workflow

This is not limited to meetings

Once companies reach this stage, the same pattern applies across many use cases.

For example:

• automatically create documents using your templates

• automatically extract and process data (e.g. invoices, PDFs, forms)

• automatically create social media posts, including images and videos

• automatically update internal systems and dashboards

The logic is always the same:

trigger → agents → output.

A word of caution

AI workflows are powerful.

They also require more discipline.

Before automating a process, you need:

1. A clear workflow

2. Stable inputs

3. Agreed output standards

This is why creating GPT workflows first could be a good idea.

If you want to get expert help in exploring this, get in touch with us. We help companies identify and create AI workflows that fit their way of working, not the other way around.

6. Consider AI images and videos 

AI-generated images and videos have taken big leaps forward.

They are no longer just experiments.

They are already used in:

• Netflix productions (e.g. VFX in The Eternaut)

• E-commerce teams

• Real estate marketing

• Performance-driven ad creative

For many companies, this is one of the fastest ways to apply AI because the output is visual, tangible, and easy to deploy.

AI images and videos are especially powerful when:

• traditional production is slow or expensive

• many variants are needed

• content must be localized or personalized

• speed matters more than perfection

Below are a few concrete examples.

The opportunities here are endless. From ad creatives and tutorials to product explainers and marketing assets. The key is not just creating visuals, but integrating them into your workflows. If you want to explore where this makes sense for your business, get in touch.

7. Start with low-hanging fruits

Don’t be overambitious.

The other six points matter. But the fastest way to build momentum is to start with quick wins.

Pick 3–5 simple AI use cases and implement them on a short timeline (e.g. six weeks).

Then do more.

Three super simple examples: 

Start using AI formulas in Google Sheets (categorize, summarize, data cleanup)

Start using Notebook LM (work faster with documents, research, and internal knowledge)

Start using Canvas in ChatGPT (structure writing and thinking)

These are very low risk. They require almost no investment and no engineering. And they deliver value immediately. Momentum matters more than ambition.

If you want help identifying the right low-hanging fruits, or executing any of the ideas in this guide, just send us an email and we will set up a short intro meeting.