Building AI Maturity in the Enterprise – A Guide for Consultants and Enterprise Leaders

Daniel Faggella

Daniel Faggella is Head of Research at Emerj. Called upon by the United Nations, World Bank, INTERPOL, and leading enterprises, Daniel is a globally sought-after expert on the competitive strategy implications of AI for business and government leaders.

Building AI Maturity in the Enterprise

This article was a request from one of our Catalyst members. The Catalyst Advisory Program is an application-only coaching program for AI consultants and service providers. The program involves one-to-one advisory, weekly group Q-and-A with other Catalyst members, and a series of proprietary resources and frameworks to land more AI business, and deliver more value with AI projects. Learn more at


As an AI project leader, you need to focus not only on delivering measurable near-term value through your AI projects, you also need to improve the AI maturity of the enterprise you’re working with. Whether you’re an internal enterprise innovation leader or an outside AI consultant, “plug and play” AI applications might be a useful place to start experimenting with AI, but they’re not a path to genuine AI transformation and long-term advantage.

Smart project leaders determine where they can find near-term leverage, and they align those near-term projects with strategic outcomes, as well as efforts to deliberately improve the AI maturity of the organization – what we refer to as Capability ROI.

With the goal of helping AI project leaders develop AI maturity over the course of their projects – and reap the rewards of that maturity in long-term AI value, we’ve broken this article out into three sections:

  1. Defining and Assessing Enterprise AI Maturity
  2. The Benefits of Executive AI Fluency
  3. Develop Informed AI ROI Benchmarks

We’ll begin by T-ing up the topic of AI maturity itself.

Defining and Assessing Enterprise AI Maturity

Prerequisites to Successful Enterprise AI Adoption
Emerj’s Critical Capabilities model for AI maturity. Source: Emerj Plus.

When we talk about AI maturity, we’re essentially talking about that subset of factors that makes an organization more likely to (a) successfully implement AI, and (b) see a strong, measurable benefit from a given deployment.

We’ve referred to these factors as “prerequisites to successful AI deployment,” and we’ve heard it referred to as “AI readiness.”

There are many ways to think about AI maturity.

For example: From the standpoint of a Head of AI or Chief Data Officer, “AI maturity” might mostly have to do with technical nuances, technical talent, and infrastructure.

For the sake of this article, we’ll be addressing AI maturity as it pertains to the audience we serve here at Emerj Artificial Intelligence Research: Non-technical C-level business leadership – a perspective that is both holistic and strategic.

Emerj Plus members will already be familiar with our Critical Capabilities model for AI maturity – and we’ve included one of our framework images from that model in this article in order to illustrate the ideas at play. As you can see there are three major elements of AI maturity in this model – Skills, Culture, and Resources – each with its own set of sub-elements.

So, is there a sequence to assessing these Critical Capabilities, or AI maturity factors?

While there is no one-size-fits-all order for diagnosing the state of a firm’s AI maturity, the order listed below works well as a rule of thumb:

1.1 Begin with Executive AI Fluency

Beginning with AI fluency is the name of the game. What that means is you need to strive to provide clients with a contextual understanding about what AI can do and the relevant use cases while resolving any misconceptions or unrealistic expectations they might have.

The most common misconception about AI is that it is just “plug and play.” If you fail to address that, everything else is impossible. If there is no executive AI fluency, there is no education to be had from all the hurdles, annoyances, and frustrations that often go hand-in-hand with AI deployment. There is no urgency for capability building or even acknowledgment that AI capabilities are a thing from the client, and any AI project or initiative will fail before it even starts.

You want to turn AI projects and initiatives into education that makes leadership smarter because education is the key to trust and long-term value in partnerships. You want to turn the hurdles of AI deployment such as difficult data access or siloed cross-functional teams into opportunities and strategic advantages.

When you build these capabilities, you help the client to pull ahead of the competition and deliver more results in the future. In the meantime, the competitors are spending money on little popcorn projects and never learning anything.

Ultimately, the first you need to do is achieve some level of shared understanding with leadership of how AI works and manage their expectations. You are not going to steer them wrong, you are going to build real ROI, and not make fake promises. Without this establishing shot, if you will, the rest of the picture will just be wrong.

Once you have achieved some level of executive fluency, you can move on to the next step in the process.

1.2 Establish the Need for Data Science Skills and Talent

You might think that this part is a given when it comes to AI initiatives, but the truth is you may not have a need depending on the level of talent in the company. Sometimes you are going to have a project that is going to require new data, science skills and talent, but not all the time.

It depends on the company with which you are working and their existing talent pipeline. This part is going to be dependent on the project. What you may see in dealing with one or more cross-functional AI teams, even if they have talent, almost everybody is going to be struggling with how to work together. Once you establish the need, or lack of, for additional data science, skills, and talent, you then have to figure out the way forward.

1.3 Develop the Deployment Roadmap

Emerj’s Phasic Model for Enterprise AI Roadmap
Emerj’s Phasic AI Roadmap model – a process for determining long and short-term AI initiatives that support AI maturity-building. Source: Emerj Plus.

Almost all companies will not have the skill or know-how to develop an AI deployment roadmap, and the onus will lie on the AI project leader to help guide that process.

You need to have a willingness to experiment and iterate to educate clients and point the way forward in an effective and rewarding manner. While many people do understand the need for iteration, most are going to have very little idea of how this pans out when it comes to AI projects.

AI is a different kind of iteration because it is new. It is not something most people are familiar, so experimenting with it is going to be very different from traditional IT projects. Knowing that some companies have more of it and some companies have less of it will indicate how your roadmap should go.

Note: Interested readers may benefits from a deeper dive on this topic in our AI Deployment Roadmap report.

2.1 Deal with Cultural Barriers

One thing you will almost certainly encounter is the internal culture that permeates a company or organization. While this may be a good thing or a bad thing for AI deployment, what you need to keep in mind is that iteration is unique and different with AI and needs explanation.

As an AI consultant or service provider, you have to prepare for the fact that you might need to deal with a heavy culture shift. Of course, this will depend on the company culture you find such as the value placed in cross cross-functional collaboration, but culture is very much tied to success in deployment.

The fact of the matter is that changing the culture to value collaboration is the only way that successful AI deployment is going to be possible. When you complete a number of projects where you actually work together and effectively with cross-functional AI teams and retain some of the things you learned in the form of playbooks and guides, then it becomes a template for future projects going forward.

A culture shift is something that you are able to do because you developed a deployment roadmap and carried out a data audit in previous projects. Now you’re starting to get the company to understand that this is a new way of working and should make it part of the culture.

2.2 Track Progress and Retain Lessons Learned

Like with anything that you will tackle that is disruptive in nature, it is almost certainly necessary to keep a close eye on progress and learn from what works and what does not. Any AI project will involve “retained learnings” on the value of data, and you want to do this for every single project to lock in what works and what does not.

As the consultant, you want to track and monitor what’s working so that you can develop playbooks for your current clients and to use for future projects for future clients. Instead of saying, “Hi, where’s your algorithm?” you can say, “Hi, here’s the way that our teams are going to work together.”

2.3 Promote Cross-Functional Team Collaboration

Promoting cross-cultural team collaboration tightly relates to developing the deployment roadmap, but in terms of the company culture.  You will likely develop this over time and numerous AI projects, and not with initial ones because a culture shift is historically difficult to pull off in the first instance.

That said, you should certainly aim to arrive at a cultural change with our clients in the first project. That has to be the case.

5 Phases of a Data Audit - Assessing AI Opportunity in the Enterprise
Assessing the state of our most mission-critical data assets is almost always advisable as part of the project selection process. This image is from our Emerj Plus article titled 5 Phases of a Data Audit – Assessing AI Opportunity in the Enterprise

3.1 Inspect Data Access and Quality.

Every project is going to involve data access, and if you can do a full-blown data audit of a company, that is perfectly fine. However, that is not always possible for various reasons, so you will have to be able to pivot your strategy to work with what you have in terms of data access and formats across cross-functional teams. Most of the time, the data silos of a given enterprise are going to be upgraded, harmonized, overhauled, and improved based on the requirements of a project.

Normally, you do not overhaul data for its own sake and then pick projects. You think about the capabilities you need to make a project happen, and then take steps to overhaul or improve data access and quality. In other words, the extent of the need to improve collaboration among cross-functional teams and increase data skills and talent is going to depend on the value of the available data.

You may have to emphasize this. There may not be a willingness to iterate and experiment, so you may have to emphasize it repeatedly. This is going to be part of every single project you do.

3.2 Develop Playbooks and Guides

By the end of an initiative – whether the project made it into deployment or not, and whether or not the project delivered a financial ROI, some kind of run books should be created that outline and streamline processes handled during the course of this initiative. These might include:

  • How to quickly clean and harmonize X data source
  • How to establish a productive cadence of meetings between SMEs and data scientists during X phase of X project type
  • A checklist to ask about APIs and integrations so that we can efficiently make requests of IT teams without dozens of later requests and revisions to our demands
  • Etc…

You must develop playbooks and guides as part of deploying AI and testing initial projects. These will provide frameworks that will be useful for your client. If you are an outside AI consultant, these playbooks may provide a template for use with other clients as well.

Keep in mind that while this is the preferred sequence of determining the need for critical capabilities, it might not always follow. Some steps are going to be a part of every project, and some are going to depend on the client. However, it all has to start with the first step, which is beginning with executive AI fluency.

The Benefits of Executive AI Fluency

We begin with executive AI fluency for a number of reasons. When stakeholders (the leaders who direct company funds and/or ultimately sign off on checks) understand how AI works, basic AI terminology, and the strategic benefits of AI – they’re able to improve overall AI ROI by three important means:

Picking a First AI Project – A 3-Step Guide for Leaders
For a lengthier discussion of AI project selection, read: Picking a First AI Project – A 3-Step Guide for Leaders

1. Improved Project Selection

You want to educate stakeholders so they can pick projects that are more intelligent. When enterprise leaders understand what AI can do conceptually, and the viable use case range, then they can pick projects from solid facts, and not from a place of assumptions or in response to what competitors appear to be doing.

Many times that is where these initiatives come from. However, from a grounded place of understanding, there is real knowledge about the value of this stuff and where could it take them. That’s really important.

Aside from a basic understanding of AI, you also want them to understand deployment challenges and to frame those challenges as benefits. You want clients to know that this is what it actually takes to succeed and that is actually a strategic advantage to for you to build these because most companies don’t understand this and that’s why there’s such a big failure rate

2. Unlock Long-Term Value

Another reason you want to have executive AI fluency is so you can have a shared understanding of strategic value. You want to be educating the buyer and explaining the process to them in an informative way. You will then be able to articulate where AI fits into the current thrust of the company strategy and what you want to do in the near and long term and they can follow your plan.

When you are able to articulate the relationships of near and long-term company strategies with AI projects, you are anchoring it to strategic value. You can do a much better job of strategic anchors when the executive has an understanding of what AI can do in the first place.

3. Capability Requirements

When you succeed in establishing strategic anchors, the client will then be able to appreciate the capability requirements to move AI forward. You will be able to connect capability requirements with the strategy anchors that have value for the client, and come up with capability anchors – i.e. the AI maturity-building initiatives that will unlock value. Many enterprise AI initiatives fail because leaders see AI maturity-building as only a cost, not as an investment – and only AI fluency among those leaders can reliably overcome this predictable hurdle. When you explain these critical capability prerequisites for AI deployment, you have common ground with the client,

For example, when you tell the client that you will have to overhaul data infrastructure, you need more in-house data scientists or work with cross-functional teams to move these projects forward, there will be little resistance. The client has a good understanding of their context and capability priorities, which will make any conversation you open up about improving capabilities much easier and productive. That gives you a good reason to make sure that you can develop AI maturity through the course of your project.

Develop Informed AI ROI Benchmarks

During an AI project, there are a couple of things to bear in mind. As you’re talking with enterprise leadership, the expectation should be that the updates your provide over the course of a project’s deployment will involve measurable and strategic value – including the AI maturity that the project builds. Limiting expectations of the leader to just measurable returns on investment is setting you up for failure (a topic we explore in depth in our Lead with Strategy article). As the AI project leader, you need to come to the decision-maker as an expert, and you need to cover all your ROI bases to show you know what you are doing – sharing firm expectations about how “success” is defined.

This way, your clients are primed to understand what they can reasonably expect from an AI project. As you go into the AI project, make sure you have a way to measure how you get closer to capabilities and that are high priorities for the client. You are making AI maturity development (i.e. the improvement of the Critical Capability elements in the graphic at the top of this article) as a selling point, an upside.

Setting measurable ROI benchmarks is a topic we explore in greater depth in a separate article, but suffice it to say that your measurable ROI benchmarks may include:

  • A reduction of false negatives and false positives for payment fraud by 8% or higher.
  • A decrease in overall time to resolution for new support tickets by 30 seconds or more.
  • Etc…

These benchmarks help to hold the AI project accountable to a reasonable near-term improvement to the business operations, building confidence among teams and leaderships, and leading naturally to larger investments in initiatives and maturity-building.

Your capability ROI benchmarks might include:

  • A way to harmonize and unify customer purchase and promotional response data across silos, unlocking future statistical/AI insight.
  • A new set of experimentation standards for the marketing and customer support departments, actively mandating and encouraging innovation and friendly competition.
  • A higher level of AI fluency among SMEs across the customer service department, through involvement of multiple leaders and front-line support staff in AI project selection and algorithm training.
  • Etc…

These benchmark goals help to convey to leadership that AI is driving positive organizational change, and is making the company more ready to take advantage of AI-related capabilities in the future.

The idea here is that you have a kind of maturity map, or “pulse” of maturity you can use not only as a resource as a consultant to gauge your progress with stakeholders, but as a tool to show your legitimate value to the client,  to show that you are genuinely their advisor and guide. You are helping them now and in the future.

When you focus on these foundations and these capabilities, it is going to help you move forward on future projects because you’re building the underpinning infrastructure for success in other projects. Unlike most AI project leaders – leaders who actively focus on maturity-building are painting a realistic picture of what deploying an AI project looks like, and are ultimately delivering more near and long-term value in the process.

Stay Ahead of the AI Curve

Discover the critical AI trends and applications that separate winners from losers in the future of business.

Sign up for the 'AI Advantage' newsletter:

Stay Ahead of the Machine Learning Curve

Join over 20,000 AI-focused business leaders and receive our latest AI research and trends delivered weekly.

Thanks for subscribing to the Emerj "AI Advantage" newsletter, check your email inbox for confirmation.