The Status: Is Your Agency Ready to Reap the Rewards of AI?

Artificial intelligence may still need to mature, but it’s rapidly improving. Follow these basic data steps to get your business AI-ready.

Overhead shot of a lake with train tracks

A recent research paper shows that popular AI assistants from Apple, Google, and Microsoft have an IQ that is less than a six-year-old. The researchers claim that the average IQ of six-year-olds is around 55, whereas the IQ of popular artificial intelligence technologies ranged between 25 and 50 when the research was conducted throughout 2016.

We’re all wondering how artificial intelligence will impact everything around us, from self-driving cars, to how we order groceries, to how we run our businesses. But would you let a six-year-old make decisions for your company?

Today’s AI technology may still need to mature, but it’s rapidly improving. And the anticipated positive outcomes on businesses are very promising: from better products and services, to improved efficiency and faster decision-making. So, even if you aren’t ready to entrust your business to the immature six-year old, don’t wait to start preparing to take advantage of AI.

The key ingredient to AI readiness and effectiveness is data. Yet, in a recent survey 10,000ft conducted in collaboration with SoDA, agencies reported that they have a long way to go to achieve the best practices of collecting, integrating, and interpreting data.

Here are the steps you need to take to make your agency AI-ready:

Start Collecting Data Today

We teach our young children many things early on, hoping it pays off later in their lives. We may instill values that aren’t relevant to their current situation, like saving money or giving to charity. Knowing these values will be instrumental as our children become financially responsible adults.

The same concept applies to AI in business. You can begin gathering information today, so your agency can fully reap the benefits of AI when you’re ready to bring it into your workplace.

One of the early ways AI will start providing value is making forecasts. For example, an AI can interpret data about work productivity, examining how changes in team dynamics have an impact on this productivity. When a project manager goes on vacation, team utilization goes down. When certain kinds of projects enter the pipeline, utilization goes up. By considering these factors against historical data, AI will be able to forecast the future productivity of your team.

According to our survey of over 100 creative agencies, only 13% of respondents consider themselves experts on making forecasts based on data.

AI can help, but it needs accurate data to be valuable.

Make Sure Your Data is Accurate

A forecasting system needs a lot of historical data. Even if you don’t yet know how to interpret all of the data, collecting this information will pay off in the long-term.

The data you collect must meet two criteria:

  • Your data must be consistently tracked over-time.
  • Your data must be accurate.

Almost half of our respondents reported that getting accurate information is a challenge in their industry.

It will still be quite difficult (or impossible) for AI to have all the data needed to be completely accurate. Unexpected events can influence forecasting and utilization, and these anomalies are likely still outside the realm of AI’s understanding. If there’s a week-long snow storm that slows productivity, the system likely can’t predict it based on your historical data. We still need people to interpret this information.

Because the forecast still requires human interpretation, these systems will primarily support—not replace—peoples’ decision-making. AI will help us proactively call attention to data trends, so people can decide whether to act on it, and how to move forward.

Translate Data into Strategy

46% of our respondents said it’s a challenge to act on data. If you’re already tracking accurate data, how can you use it to support your decision-making abilities today?

When you make a strategic decision, look at the data you’re tracking to understand whether or not this decision will have the intended impact.

For example; imagine your team keeps telling you they need to hire another designer. How do you know whether or not you need to hire someone full-time?

If your team has been tracking their time, you could simply look at your design team’s hours over the last three months and see whether they’re consistently overworked. Are all of your designers so overworked that it justifies a new full-time hire? Or is the bulk of the workload resting on only one or two of your designers, while the rest of your team’s workload remains light?

This data helps you identify the root causes of issues like these, so you can accurately determine whether you truly need to hire another designer, or if some other, deeper issue needs to be addressed within your design team first.

You can also use data to verify (or refute) the hypotheses of your decisions. If you’re collecting data over time, visualize it to easily identify trends. Look at these trends over the next few months to see whether the impact of your decision matches your original hypothesis.

Don’t Forget the Human Factor

If you focus exclusively on collecting measurable accurate data, your insights may still be fairly shallow, unless you augment your data with what HBR calls “thick data” — insights and observations about human behavior and its underlying motivations.

Think of thick data as the intuitive knowledge you develop as a result of seeing how a particular account director interacts with a team of developers, for example. Or anticipating a client blow-up because you see how the personalities of the team and the client are creating a combustible friction. Finding a way to record such insights today will enrich your decision-making algorithms, evolving your AI system from a primary stimulus to one with the beginnings of emotional intelligence.

If the average IQ of artificial intelligence is equivalent to a six-year-old, there’s still a lot of growing and learning left to do. But that doesn’t mean we can simply wait for AI to mature. To stay competitive, start collecting accurate, relevant data today to maximize the positive impact of these systems in the future.

Perfectly reasonable business advice, delivered to your inbox.