An AI Professional Services Roadmap
Service company executives recognize artificial intelligence offers significant opportunities for revenue growth, productivity improvements, cost savings and competitive differentiation. However, planning the design and build of this new technology can quickly become overwhelming.
It was Mark Twain that said, "The secret of getting ahead is getting started. The secret of getting started is breaking your complex, overwhelming tasks into small manageable tasks, then starting on the first one." That's the approach we suggest, using our Johnny Grow AI Professional Services Roadmap.
It's a build for scale model. And it's versatile. It can be used for initial deployments; to get from pilot to production; or to scale. A summary illustration is shown below.
I'll highlight a few of the most influential steps in the journey.
PREPARE
The first step is about ensuring the prerequisites are in place.
History and experience show it's the first step that is often glossed over or minimized. And that of course impacts the effectiveness of all downstream steps.
Every Artificial Intelligence journey starts with data.
A through data assessment is an upfront investment that may be the single biggest determinant of success or failure.
Recognize that poor data quality is like a disease in two ways. First, it often goes undetected until it's a big problem. Second, left untreated it only gets worse.
A data strategy, data quality standards and data management practices are the cure.
A data strategy identifies what data should be captured and how it will be used.
Data quality is the ability of data to serve its intended purpose.
And data management is the process to ensure data quality. Data management practices analyze, validate, categorize, clean, dedupe, append and archive data.
Know that most dirty data comes from human error and most human error comes from missing data quality standards. It's a virtuous circle.
To keep your systems clean and efficient you need to define data standards and show what good data looks like. Create a data policy to define how data should be entered and standardized. Then create system validation checks with automation to enforce your data policy rules.
And one more thing. Refrain from the popular but ill-advised more-data-is-better thinking.
Unused data creates a distraction that must be separated when creating AI models. It's not about how much data you have but how much gets used.
You will also want to assess IT readiness. That includes assessing skills, your talent mix, resource availability and your professional services technology strategy.
In fact, you would be well advised to update your IT strategy to include an AI strategy. This aligns technology capability and investment with business benefits and outcomes.
For example, a common tech strategy is called PACE. It aligns the methods business leaders use to create value and competitive advantage, described in terms of common ideas, different ideas and new ideas, with different types of AI technology, such as predictive AI, generative AI and autonomous AI.
Many people think Gartner created PACE. They didn't. But they made it very popular.
DESIGN
The second step is Design.
The most successful AI doesn't attempt to retread the cow path with new technology. It reimagines how work gets done. And the best way to reimagine this is with a design thinking workshop.
The workshop is typically a one-day exercise. It facilitates a cross functional team to surface the highest impact and most important success criteria. It measures success in user, customer and business outcomes, and according to the people that will most use or benefit from the technology.
It also cuts through complexity and clutter. It simplifies, humanizes and focuses on what's most important in helping people achieve their objectives.
Design thinking shifts goals from being measured in software utility, and expressed as technology features and functions, to being measured in business outcomes.
BUILD
The Build step often begins with professional services use cases. It's a good idea to segment use cases according to business processes and the three types of AI technologies (shown below).
Client acquisition processes may include the following use cases:
- Gen AI can help professional services salespeople generate sales correspondence with prospects and customers or create proposals and SOWs.
- Use Predictive AI to forecast which leads will become opportunities and which opportunities will become customers.
- Autonomous AI can use agents, such as SDR agents to converse and qualify online prospects. It can then schedule appointments for qualified prospects with sellers or partners.
For services delivery processes, consider the following use cases:
- Use Gen AI to create weekly status reports, risk registers or governance updates. The technology can highlight or dive deep into the measures or items most important to each recipient.
- You can use Predictive AI to forecast troubled projects, identify services staff in need of help, or forecast financial profitability.
- Autonomous AI can distribute knowledgebase artifacts and lessons learned from prior projects or clients to people who can benefit from them.
Client relationship management processes and use cases may include the following:
- Gen AI to create correspondence for drip or nurture marketing campaigns.
- Predictive AI to identify clients which are not in your top tier segment but should be (something often called look-alike modeling). Or it can forecast clients at risk of churn – with estimated timeframes and confidence levels. It can also recommend retention levers.
- Autonomous AI can engage with your clients. For example, it can deliver surveys as part of a Voice of the Customer program.
IMPROVE
Delivering AI innovation doesn't happen in a single step. An iterative process is needed.
Agile methods work particularly well with these deployments because they bring together the following 4 critical success factors:
- Collaboration among self-organized and cross-functional teams
- Close user involvement
- Iterative and adaptive deployment processes, and
- The frequent delivery and inspection of incremental releases
There are many types of agile methods. Use the one you have, or if you don't have one, I'd suggest starting with Agile Scrum.
SCALE
Scale is the final step of the AI professional services roadmap.
That means shifting from pilots to production in order to get from validation to value.
And value is not measured in technology terms. It's measured in business outcomes.
Value must show progress of the company's priorities and be stated in financial or monetary terms.
One More Thing
There's one more insight you may find helpful.
I'm sometimes asked if small or midmarket companies have a disadvantage is adopting AI professional services technology.
The answer is that they have both advantages and disadvantages. But more advantages.
Smaller companies are more nimble and agile. They are able to move more quickly. They often have much tighter executive sponsorship.
They also benefit from what they don't have. They are less likely to have lethargic hierarchies that require advance approvals to begin. They have fewer turf wars or political fiefdoms to pacify. There are fewer bureaucratic approval processes to slow down the journey. There are fewer managers who will view new technology as a threat to their influence, domains or career path.
Large companies have more skilled resources, and they tend to better at creating strategies. But they have far more organizational complexity. They also have more sacred cows and layers of resistance. They have more people that want to preserve the status quo.
Any company regardless of size can be successful. And smaller companies are not at a disadvantage.