The AI Journey: How Do You Take Insights to Decisions?

Let’s Talk AI


It’s been 67 years since Stanford professor John McCarthy defined artificial intelligence as “the science and engineering of making intelligent machines.”

Share on facebook
Share on linkedin
Share on twitter
Share on email

About the author

Rigvi Chevala

CTO, Evalueserve

Rigvi Chevala is Evalueserve’s chief technology officer (CTO). He has more than 16 years of experience leading high-performing product engineering teams in building enterprise-scale products and applications.

In-house AI experts

Cathy Feng

Swapnil Srivastava

When using AI to solve business challenges, we need to draw out the distinctions between how AI can aid decision-making vs. AI that makes decisions for us.

As an illustration, when an autonomous vehicle senses a pedestrian crossing, it could either:

  • send the driver an audio alert (i.e., aid in decision-making), or
  • brake automatically (i.e., decide on your behalf).

If you consider a B2B setting, you could have an algorithm that can rationalize inventory levels by adjusting purchase orders or providing procurement teams with the intel and data to make those decisions themselves.

In this eBook, Rigvi Chevala helps teams understand the way domain-specific AI solves real-world problems that organizations face, with tips from him and other industry experts!

AI is kind of dumb by itself because it only does one thing, and it does one thing very well. The trick is to identify things we can’t make effective decisions on and see if it’s mathematically possible to identify patterns and then use them for your decision-making.

Rigvi Chevala, Evalueserve’s CTO

Understanding the Challenge

0 %

of organizations had implemented AI

0 %

AI employment increase between 2015 and 2019.

Source: 2019 Gartner survey 

Did You Know?

99% of firms reported they were actively investing in AI and Big Data.
However, according to a 2021 NewVange Partners survey, only 24.2% have forged a solid data culture, and only 29.2% are experiencing transformational business outcomes.

Analytics Maturity

A mature AI-powered organization with a strong focus on data governance leads you to where you want to be going: upwards on the analytics maturity curve. A 2019 Forrester report found that about 57% of organizations were classified as beginners on the maturity curve, 35% were at an intermediate level, and only 8% were at an advanced level.

0 %

of organizations were classified as beginners on the maturity curve

0 %

were at an intermediate level

0 %

were at an advanced level

Briana Brownell

Pure Strategy

Brownell mentioned some things to keep in mind when organizations are kick-starting their journey up this maturity curve:

Successful and mature organizations often have an internal innovations team in place. The internal innovations team is centralized and helps educate different departments, building their skillsets with available insights.

Ask questions every step of the way:

  • Is your data set uniform before you build an analytics model?
  • Are you talking to a variety of teams across the company and getting their input?
  • How can you improve your products?
  • What markets can you enter?

Macro Levels of Data

Every day there are about 2.5 quintillion bytes of data produced, a Forbes article reported. That is a lot of data with which to keep up. For optimal business results, it’s key to narrow down on the right type of data for your business using AI and a team of experts.

You must understand the space and take a methodological approach, which begins with data collection and understanding, even managing, that big data. Then, before you can provide the capability for decision-making, you have to combine natural language processing with human curation.

This was a core topic in the Decisions Now episode featuring Cam Mackey, in which we discussed using competitive intelligence to drive better business outcomes.

Cam Mackey


Have a growth hypothesis in place to poke holes in your data sets that help teams understand the why and the context of the challenge and data, without which data is meaningless.

Data Governance

Achieving results depends on your company’s data governance and data management strategies.

When sorting through the large pools of data that can be processed by your AI algorithms to deliver decision-ready insights, it becomes important to have quality data. 

In the fifth episode of the Decisions Now podcast, we asked Bin Mu, Adobe’s Chief Data Officer, for his take on the need for data governance and creating that framework.

My advice will be to have a broader view of the company of how data is being used. I start with a high-level strategy, scope it, and do that one step at a time showing the achievement, implementing the data governance strategy and the approach. So, start to show the value, and train the users to get used to it. Follow it, build into their nature, and use that quadrant to roll out to the entire company.

In our conversation, Bin gave these pointers on how to build a quality data governance strategy:

Make data governance the front and center of your business model. Have a team of experts in place as well as automated processes that prioritize this strategy.​


Have a centralized data hub, a system that streamlines all operations across different departments and processes data in real-time.


Ensure that you are making your goals scalable is essential. Know what KPIs (key performance indicators) and parameters you want to monitor, so the tool you use can notify the team when something in your system spikes or is flagged as out of the ordinary.


Data Delivery

After you’ve sorted through the data, analyzed it, and connected the dots to links and trends – bam! You have insights.

An Accenture study reported that 74% of participants said they felt overwhelmed and unhappy when working with data yet 87% percent of them believe in the value of data.

That is why data delivery is important, and analytics teams need to learn that different people understand data differently. Evalueserve’s co-founder and director, Marc Vollenweider, joined us on the Decisions Now podcast to discuss AI, analytics, and insights and how to go about turning them into business decisions.

Marc Vollenweider


Some people want a quick statement, the quick message, maybe on WhatsApp or whatever, maybe with a link to touch if they’re interested in doing some more, other people want this to be spelled out to them in text, and some other people just want the hard facts, the data, and they’re happy with it. So, understanding the end user is critical in this.

Data analytics teams need to know their end-user well and deliver data to them in the manner that benefits them the most.

Marc emphasizes the importance of ensuring your decision-maker understands the insights and data you present to them for optimal results. We asked him what some key things are to look out for when delivering data.

Here is his four-step process to ensure accurate data-delivery: 


Getting the pure raw data


Converting that data into information


Turning that into insights


A bunch of insights together convert into knowledge at the fourth level.

Drawbacks of AI

An integral part of being an AI and data-driven mature organization is knowing that AI is constantly evolving.

Companies should not become too dependent on their decision-making tools. When machines don’t understand the context they are built for, they spit out skewed and error-filled insights.

Algorithms are only as good as the data you feed them, and many sets of data, especially historical ones, can include biases.

Humanyze co-founder Ben Waber talked about AI biases on the seventh episode of the Decisions Now podcast. Specifically, he suggested what to keep in mind when thinking of AI biases:

Ben Waber


Know your tools, be cautious of the vast data pool, and really narrow in on the questions and predictions you’re trying to figure out.

With context becoming so important, we need to focus on retraining our models with time. Something that was acceptable five years ago may not be today, or a fact we once knew may have been modified or debunked. To ensure we have real-time, up-to-date insights, we must work with quality data and AI.


The AI and analytics journey can be a long one with many twists and turns, but it is one worth taking to find success in today’s business world. Fortunately, many have experienced this journey, and sharing those lessons learned can make adaptation easier and faster, allowing you to make better decisions swiftly.

In this eBook, we talked about what it means for leaders to incorporate AI and strengthen their data and analytics culture. Now that we have gone over all the ups and downs of AI, you might be ready to take the next steps and kick-start your AI-powered journey.

Make sure to listen to the Decisions Now podcast wherever you get your podcasts. You can find it online at In the interim, we put together some questions you may want to ask yourself before incorporating AI into your processes:

Ask questions every step of the way:

  • Do you know where you want to apply AI?
  • Are you starting with the end-user in mind?
  • Is the outcome measurable?
  • Have you considered what data is needed?
  • Do you have the required skillset within your organization, or would you have to outsource those skills?
  • How critical is time to value?
  • Have you taken into consideration the costs and budgets before you start your project?
  • Do you have a dedicated person or team that drives this goal?
  • Are you aware of the routine maintenance and upkeep to account for model drifts?