(Shared) Business Services in the Age of Data & AI

In the age of AI business & service models will change

“AI is becoming the new operational foundation of business – the core of a company’s operating model, defining how the company drives the execution of tasks”*. This truth is fueling an unprecedented technology and expectation hype where fancy buzzwords (“hyperautomation”) and superficial expertise labeled as research create a toxic environment for sound application. At the same time unrealistic and thus mostly unkept promises build a solid argumentation basis for AI skeptics, at least in the short term. IMHO in most cases adoption and not technology is to be blamed. According to our nature we overestimate the short term impact and underestimate the mid-term paradigm change. And we blame ”intelligent” technology where our failure to use it intelligently should be blamed. Yet the paradigm change to come is inevitable and will disrupt businesses and the (Shared) Services landscape.

Technology: from hype to hubris

Let’s stick with technology first. In the process automation domain “the hype” started with simple, highly unintelligent low code task automation (“if this than that”) aka RPA. Process & task mining joined the hype curve later but will outperform most growth expectations as process insights on transaction and click level are the foundation to thrive towards a “fluid enterprise” (though to really get there far more than process re-engineering is required. Mining is a dirty business and few have the stamina and endurance to be successful). BPM tools (aka workflow & task management) have always been there, but only are joining the hype now by embracing the same low code and UX essentials that made RPA a category (outside IT test mgmt./services monitoring). All these technologies bring “new skills” to the “low code automation toolbox” (any programmer would add “for dummies”). And to put the cherry on the hype all these tools are now packed with AI/ML to a level that makes science fiction on PowerPoint a commodity. 

The value is in the asset, not in the enabler

But why are we discussing technology when we should be discussing data? In essence technology (RPA, Process & Task Mining, BPM, ML, -> just add any here) is highly irrelevant for value creation. Technology evolves and adoption needs constant enablement (and no, absolutely no “leapfrogging” here). It’s available on a highly competitive market and accessible for everyone (praise to XaaS from the Cloud and pay-per-use license models). But produce value a real asset is needed and this asset is not technology for (process) automation or analytics. It’s data. You will never be able to outperform competition by buying or using technology at scale, but only by monetizing your data at scale.

Simply put: technology is an enabler. In essence it helps you with:

• Creation of data via digitization of information (ICR/OCR, NLP, VCTR, etc.)
• Virtualization of data flows (BPM, APIs)
• Virtualization of tasks (bots = virtual labor arbitrage)
• Automation of decisions and rules (ML models, algorithms fueled by data)
• Data enrichment (labeling, knowledge graphs, etc.)
• Visualization of flows & data (Process & Task Analytics, etc.)

So when talking about deploying technology to “automate” what we really should mean is to “make data available & flow in real time”. There is no automation without data. That’s why any manual input hurts. And that’s where ML is challenging automation frontiers because it adds complex non-linear decisions to the list of automatable tasks (“if you can draw it on paper you can automate it” is no longer true”). If a Bank wants to enable an account opening within 3 minutes it has to build upon a “human out of the loop process” that includes decision taking. Don’t try to get there via the continuous improvement of legacy processes (and systems).

Value beats efficiency

Why do I think all this is vital to Shared (Business) Services? Because it’s a service business that is (still) highly transactional on the surface but also highly data intensive at the core. Virtually no other unit puts large scale transaction processing “as a Service” at the very core of their value delivery and business model. Build on the promise of synergies scaled Shared Services work in horizontal end-to-end process setups which require process execution and data handling across functional / BU silos. So in every case there is a huge bundle of domain and service expertise to start with. This expertise has been used – in many cases – to realize efficiencies via technology deployment at scale (so far). With the emergence of “low code” automation tools there was an unprecedented jump in automaton initiatives as deployments were no longer bound to change requests in core applications (= IT resources). Simply put: there was more liberty for DIY and structural dependencies (e.g. ERP value flow, customizing) were reduced. If theses initiatives reached their automation goal (many did not -> human intelligence failure) the side effect was always a structured “datafication” of information (inputs, throughputs, outputs). My favorite example are RPA bots: they log every activity. These logs are a goldmine of information. Or Chatbots: NLP datafies a complete conversation incl. emotions / tone.

Still not many Shared Services impressed the C-level with what they did. Why? Because most focused on pure bottom-line impact. Technology was used to drive processing (transactional) cost down but not with a clear focus on value-add from data. From a customer’s perspective the lack of enthusiasm is understandable. Their willingness to pay is bound to service levels ("money for value"). Providing the same service level at lower cost is in line with expectations when prices are reduced accordingly (if you advertise productivity you need to deliver on the price). Getting a “wow” requires service innovation i.e. either providing a new service (Bot as a Service, Analytics as a Service, etc.) or reshaping services in the portfolio (HR chatbot for employee queries). I firmly believe that Shared Services will move from “high volume at low cost” to “high volume at high value” offerings. To get there they need to add Data to their DNA of Service, Process and Technology expertise.

So let’s talk about the HOW…

Open platform beats closed shop

In a very first step it’s important to recognize the limitations of the ecosystem. Even if Shared Service organizations invest massively and take resource deployment seriously this will not do the trick in today’s vibrant economy. The secret sauce is collaboration. If there is one key lessons learned from successful FinTech startups (take N26 / digital banking) then it’s the power of open platforms and ecosystems. This requires serious partnerships with other Shared Services, ScaleUps, Software Vendors, Implementation Partner and even the competition (which in a captive Shared Services world is called BPO). The rationale behind is very simple:

• You can’t transform your assets as fast as an open ecosystem can transform you. Getting the required inputs and driving adoption to scale requires massive investments. Sharing is caring has never been more true.
• Customers value easy, seamless integration. Think of your car interacting with your mailbox calendar, your smart watch and phone to manage your day-to-day journey. From traffic alerts (I know where you want to go and when) to stress reminders (your next meeting is in 15min, have a walk to lower you blood pressure first) to task management (this sales order is really important to our cashflow, please review now). There will never be the one platform that can host it all. Thank god there are APIs.

In essence the decision “make vs. buy” has become “make vs. buy vs. co-create”.

Interaction engine beats transaction engine

I already touched this topic: scaled Shared Services in multi-tower setups are in the sweetspot of corporate value chains. They are not only knee-deep in Financial Reporting and HR but also in cash-relevant P2P and O2C processes (amongst others). Just imagine how many transactions they process, how much data they manipulate and how many business partner interfaces (vendor, supplier, employees) they manage. They host the people, the technology and the knowledge to influence sales, engage business partner, generate actionable insights and boost operational efficiency. Unfortunately they are mostly known for bringing transactions to benchmark levels … and not for generating value from interactions based on an integrated view on formerly siloed activities. Especially with the latest push to commoditize ML technologies (= package highly complex models into low code software) it’s time to monetize the value from interaction and prioritize vs. the never ending reduction of transaction costs (mostly the lemon has been squeezed already).

AI is not a service offering. Data is in all your offerings.

The first reflex to the AI hype has mostly been “great, let’s make AI offerings”. Let our reporting factories sell predictive analytics. Let’s hire data scientists to build the models. Only the second thought was: …wait. The data scientists need data. Let’s hire data architects as well. The obvious result was: now we have data scientists and architects but we need domain expertise to ask the right questions, interpret results, produce actionable insights and package those into a compelling service offering. This event chain was caused by a focus on technology (AI for Analytics). With a focus on data the first question would have been: where do we have data assets that are worth mining? And the second: which additional skills (technology & people) do we need to add value? My point here is: to get a real competitive advantage one needs to leverage existing data capability. Anybody can hire people an deploy technology. Not everybody is able to put domain expertise on top. In the case of Shared Services operations are increasingly digital and the virtual delivery share (bots) is growing. This means that data has never been so abundant. Combined with the existing expertise on processes and technology there must be opportunities to drive value for customers when they’re only expecting efficiency.

To summarize: in the age of Data & AI Shared (Business) Services are uniquely positioned to drive value (and delivery efficiency on the way). The key is a data and value focused approach to technology and AI.

* Quote from: Competing in the Age of AI, Marco Iansiti & Karim R. Lakhani