I entered the digital marketing industry about 15 years ago.
Back then, we used to print stacks of online advertising reports, put them in paper bags, and bring them to clients saying, “Here are this week’s numbers!” (Yes… that was normal. haha)

Things have changed a lot since COVID, but in Japan, the culture of
“creating a report, putting the numbers into PowerPoint or Excel, and presenting them to the client”
still remains surprisingly strong. And my impression is that many clients simply listened and nodded along—without much discussion.

■ The depth of data handling can (and should) go much further

When we work with data, there are actually multiple layers we can dive into:

• Monitoring

• Visualization

• Analysis

• Interpretation (insight extraction)

• Hypothesis building

• Action planning

• Estimation / Forecasting

• Validation

Until recently, this was as far as we could go.
But the rise of generative AI has shifted things a little.

■ What generative AI truly adds is “dialogue”

We can now load data, ask questions, and have a back-and-forth conversation with AI to dig deeper.
This style is certainly new.

But personally, I feel this:

If a company lacks real dialogue experience between humans
(agency ↔ client),
there’s no reason to expect the quality to suddenly improve
just because the “other person” is now AI.

Generative AI is still subject to “Garbage in, Garbage out.”
Shallow data and shallow questions will only produce shallow answers.
And unlike a human, AI won’t say:

“Hmm… this data looks messy. Are you sure about this?”

■ That’s why we must first review our own “depth”

Thanks to generative AI, we can now move through analysis, insight extraction, hypothesis creation, action planning, forecasting, and even validation—together with AI.

But before that, the more important questions are:

• Where are we right now in our data-handling depth?

• At which layer does our process get stuck?

Companies that have clarity here will accelerate the moment they adopt AI.
Those that don’t… won’t see much change, even with AI.

■ What I ultimately want to say

Using generative AI is not a technology discussion.
It’s a discussion about revisiting our questions, dialogue, and thinking.

Compared to the days when I carried paper bags full of printed reports to clients,
we’re now in a time when the quality of a single question can instantly change the depth of the entire discussion.

That’s why I want to reevaluate, including for myself,
how we engaged with data before AI existed.
In the end, I believe this is the fastest way to improve the quality of our dialogue with AI.

Masaki “Mark” Iino
Founder & CEO
SOPHOLA, Inc

P.S.

I recently had lunch at Obusedo for the first time, and it was far more delicious and enjoyable than I expected.
If you visit during the autumn foliage season, I highly recommend it!