The Benefits of the Knowledge-Centric Platform
Donald Thompson

As founder, president and COO, Donald Thompson leads Maana’s product vision, oversees technology and product management.

15

September 2017

The Benefits of the Knowledge-Centric Platform

This week I was excited to see two respected analyst Jason Bloomberg from Intellyx and Mike Guilfoyle from the ARC Advisory Group validate digital knowledge technology as a new breed of technology that delivers much faster business value.

There are many reasons why Maana’s customers select our platform to drive digital transformation, but without a doubt the primary reason is time to value. In almost all cases our customers have been able to complete their first use case and demonstrate business value in less than three months.  In my discussions with customers these are the benefits our customers see in using the Maana Knowledge Platform™.

Faster time to value: Reusable Models
With the Maana Knowledge Platform, models are designed to be flexible, reusable assets that can be easily and quickly repurposed to represent a related problem and its associated workflows (for example, optimizing field service or accelerating collections).

This means, that if users have already figured out when an industrial equipment is likely to fail and identified the likely potential root cause of failures (as part of a predictive and prescriptive maintenance initiative, for example), they can re-use that existing knowledge to determine which service engineers are available to maintain assets and which are best suited for each task. Using this knowledge to make staffing decisions can improve the company’s service response rate.

Another important factor contributing to faster time to value for Maana users is the use of Knowledge Assistants. As their name suggests, these assistants are artificial intelligence algorithms, machine learning, and many other techniques to augment people as they extract, explore, discover, and model new forms of knowledge describing some aspect of the business domain. This user-guided, machine-assisted approach to building models optimizes the interactions between people and machines to maximize overall productivity and effectiveness.

Integration with existing data infrastructure
Another capability of Maana’s Knowledge Platform is that data can reside anywhere and is easily accessed and indexed by Maana. You don’t need to have a data lake to get started with Maana.  But we also recognize that many IT organizations have invested in technology they want to leverage.  That’s why we designed the Maana Knowledge Platform to also support traditional, third-party systems for ETL, data wrangling, visual analytics, dashboard development, and more.

We also support any backend system, such as Google’s TensorFlow, gridMathematica, or even a proprietary system through a pluggable interface. Maana provides fluid integration and composition to enable users to use their preferred tool for each job. Whether it is a Python- or a complex, deep learning network.

Collaboration between Subject-Matter Experts and Data Scientists
Traditionally, lack of coding skills or familiarity with data science platforms has prevented subject-matter experts and business analysts from being a part of many analytics and optimization initiatives.  Maana’s platform makes it possible for them to use machine intelligence-based Knowledge Assistants to create conceptual models and contribute their expertise in a manner that Maana’s Knowledge Platform can understand.

With Maana  a petroleum engineer, for example, can easily join data from a myriad of drilling equipment and reports. An insurance underwriter can simply provide an example of a “good” insurance contract and use it to search through a global database to find other “similarly good” contracts in seconds. In the past, these kinds of tasks were considered IT or data science projects, each with high costs and long lead times that caused delays and lack of results.

With the Maana Knowledge Platform in place, data scientists also play an important role that is more suited to their sophisticated skill set. For example, they focus their expertise on building more advanced computational models that apply various solution techniques to the problem that has already been defined by the subject-matter experts that operate the asset and understand the business process.

Centralized, incremental knowledge and enterprise-wide connectivity
At a major industrial company the field service organization of that company was using the Maana platform to identify optimization initiatives to make that division more profitable. As part of that optimization initiative they discovered that certain field engineers were ordering many unnecessary parts that tied up inventory in the field.  That led to another project that recommends the right parts to field engineers based on the service use case.

The models that were created for field service optimization were leveraged and built upon by the inventory team, making the second project significantly faster in delivering results.

This is a perfect example of how the knowledge platform allows knowledge building in an agile, value-driven way. Unless people first define specific problem-questions, it can become very challenging to determine what knowledge needs to be encoded in the platform. Once models are ready and available for a particular problem, they are easily searchable and can be quickly repurposed for a related problem.

These models easily reveal the interdependencies between various processes and assets across the organization, which is an essential aspect of digital transformation and revealing these interdependencies and the re-usability of models created with the Maana Knowledge Platform is the key to accelerating digital transformation from five years to two years.

Digitized decision-making
One of the keys to achieving quantifiable business gains is improving day to day decision-making. In fact, in the context of digital transformation, the main motivation behind digitizing workflows is to improve the quality and speed of decision making. Meeting decision deadlines (i.e., decisions that need to take place along the workflow of an operation and must be made in a few hours or a few minutes) can have a big impact on gross margins.

Maana’s Knowledge Applications are the end product that subject-matter experts use  to augment decisions by machine intelligence. These Knowledge Applications are powered by the models created on our platform and embed recommendations directly into existing line-of-business applications. For example, in the “Accelerating Accounts Receivable Collections case study” each collection agent tracking late payments gets a customized call list embedded directly into the AvantGard GETPAID collection system they are already familiar with.

Knowledge Applications are tailored for business users and provide prebuilt analytic functionality for predictive and prescriptive maintenance, supply chain optimization, accounts receivable optimization, and more. Users can also build new Knowledge Applications or extend existing ones to meet custom requirements specific to their needs.

Models Learn and Adapt
We are heading to a future where software platforms can give users advice as part of a real-time dialog. Models running on a platform will be able to build a human conversation like a chain, enabling users to explore their queries in a natural, question-answer format. And through every interaction, the models will learn and provide better responses going forward.

In practice, this means that the first time a drilling engineer uses the Maana Knowledge Platform, the platform won’t necessarily know to pull all the incident reports and provide a maintenance recommendation. But over time, with each additional query made by users, related models will be enriched and start providing more meaningful recommendations. The platform itself gets smarter without anyone writing software code.  You and your employees train it the more you use it.

The shift to becoming a Digital Enterprise is a shift to digitizing all key processes and decision flows that have the biggest impact on revenue, cost, safety and risk.  Digitization means building models that provide recommendations into those decision flows, so employees can make better and faster decisions.


by Donald Thompson in Maana Leadership