Get Started

AI Cultural Change From First Principles

Scaling AI Doug Bryan

So, you want to change your company culture to be more data driven and use AI everyday. You want to get advanced analytics beyond the Center of Excellence (or what one of my customers call the RSPD, Really Smart People Department, and another calls the bottleneck) and into all your business units and business functions. This article discusses how executives can use price and demand to increase AI development.

Price: SaaS and cloud computing have reduced AI platform prices to only about 12 work days per year per analyst or 5% of compensation. Prices are at historic lows. (Assumes 235 workdays, $100K average data analyst total compensation, and $5,000 software costs per analyst per year.)

Demand: Many factors drive demand in companies including personal factors like user experience, social ones like peer pressure and network effects, and business ones like return on investment (ROI).

demand formula

User experience: If a platform is easier or more fun to use, then demand generally increases.  For example, when the iPhone was released 14 years ago it didn’t make better phone calls than its competitors but its user interface, iTunes, and a myriad of third-party apps made it fun and it quickly gained market share.

Social pressure drives demand via:

  • Safety in numbers: If a lot of people on your team are using a platform then it probably works well or, at the least, won’t put you at a competitive disadvantage relative to your coworkers.
  • Network effects: The value of many platforms increases with the number of people on it by facilitating reuse, finding problems quickly, and fixing them quickly.  For example, in 2004, IBM found that Wikipedia was more accurate than Encyclopedia Britannica because Wikipedia had many more readers and editors fixing problems.
  • Management directives: When managers give directives to use a platform, then not doing so may have an adverse effect on your promotions and compensation.

ROI: We touched on price, which leaves value. What value does your AI platform generate? What’s the benchmark for your current AI approach, if any? How does overall business value increase or decrease with cost, revenue, profit, time to market, quality, agility, risk, and innovation? Businesses should consider them all (and their tradeoffs) since it’s often difficult to improve one without worsening another. 

Recall Mark Zuckerberg’s famous quote, “Move fast and break things,” which suggests that reducing time to market may also reduce quality. An example is Amazon.com’s checkout pipeline reliability was less than 90% (less than one 9) during the 2004 Christmas season yet revenue grew 31%. Revenue went up while quality went down.

What Can Executives Do to Accelerate AI Adoption?

Many drivers are available.

Reduce prices: The U.S. government subsidizes electric vehicle development and purchases.  Managers can do the same for AI. Many companies have a central IT department that handles AI platform provisioning and charges expenses back to business units. We’ve seen big differences in the markup of these chargebacks from -50% to +400%. Managers can accelerate AI adoption by subsidizing business units cost and reducing IT markups.

User experience: Provision platforms that are fun and easy to use by a wide variety of data analysts, not just expert data scientists. (Renee Boerefijn, Ph.D., Director of Innovation at Bunge Loders Croklaan said Dataiku’s platform “enables fun,” high praise indeed.)

Social pressure: Evangelize the benefits and internal successes of AI at every level of the organization. It’s insufficient to just have some lunch-and-learn sessions for developers. You need executive-level, C-suite evangelists too. Also, fund the management of a sustainable learning community so that network effects are obvious via vehicles such as internal conferences, awards, and laptop stickers. (Yeah, never underestimate how much developers like laptop stickers and how well they signal community membership.) Lastly, managers can mandate or recommend AI adoption headcount goals.

ROI: A key here is that managers use common cost and value metrics so that different AI approaches may be compared, apples to apples. Managers should also set priorities between completing goals such as cost reduction, revenue increase, time to market decreases, quality increases, agility increase, innovation, and risk reduction, based on their strategic goals. I’ve even seen managers go so far as to quantify priorities, such as $1 of incremental revenue is worth $4 of cost reduction.

In summary, to drive AI adoption across their business and change to a more data-driven culture, executives can:

  1. Subsidize AI platform costs.
  2. Provide fun and easy-to-use platforms.
  3. Evangelize results at all levels of the organization.
  4. Fund a sustainable learning community.
  5. Set AI adoption headcount guidelines.
  6. Align on clear, common cost, revenue, profit, time to market, quality, agility, risk, and innovation metrics.

You May Also Like

Simplify Anti-Money Laundering (AML) Alerts With Data-Backed Triage

Read More

Key Steps Involved in the Machine Learning Process: A Primer

Read More

Get Rid of Friction & Trigger an AI Chain Reaction With MLOps

Read More