Outsourcing to AI - How will it be
Outsourcing to AI - How will it be

The amazing times to come

Over recent interviews (@Aderson.com), we have come to get acquainted with some AI based service offering that can be Outsourced by companies to leverage their operational business processes effectiveness, quality, costs and time to market.

In the case of MarianaIQ, sales prospects can be identified by AI with laser focus accuracy, having specifically shaped marketing content forward to each individual prospect’s stake holder, within the logic of promoting and enabling/ facilitating a call to action on their side. This means to grow the sales pipeline with hot, highly qualified leads with whom the company has started a targeted 1o1 interaction. (check here)

In the case of AppZen, AI based Backoffice Automation tools allow companies to reach detail auditing and analytics results within a fraction of the standard way required time frame, hence enabling faster time to market and business momentum. (check here)

AI is tightly connected to one other Technologic recent evolution (although the conceptual scope is not new), which is Robotic Process Automation (RPA).

Per definition, RPA is constituted by coding which allows with or without mechatronics integration (robotic physical machines), to partially or fully automate human activities which are manual, repetitive and rule-based. In doing so, RPA allows companies to Outsource towards a fully automated workflow, business processes that are clearly definable and repeatable.

Machines are replacing humans as we speak, and it’s not bad

Why having humans replaced by machine based fully automated process workflows does, in fact, constitute an advantage and not a problem?

In both above-mentioned examples of MarianaIQ and AppZen, having the task performed by humans would imply a cost level which would (in most cases), simply render the Business Case unprofitable and therefore the initiative would never move forward.

Now, by resorting to these technologies, companies can accurately identify more leads and potential nonconformities, improving business profits, therefore making the company a safer and more competitive workplace for the humans that work there.


Note that by robots we do not merely consider machines, but also fully automated software.


RPA by itself presents several advantages to the corporate world like:

  • Reduced cycle times – after all robots do not sleep
  • Scalability – once a process has been automated there is room to add as many robots as necessary performing that same process in parallel
  • Flexibility – the same robots can be used to different tasks as they become not needed for the previous one, come back to it if necessary at a later time
  • Human motivation improvement – robots do not get bored with repetitive tasks, so by allowing human workers to stop doing such tasks and focusing on more “creative” ones, corporate morale has room for improvement.
  • Improving regulatory compliance – robotics is less prone to errors than humans, therefore it is unlikely that any regulatory compliance issues will arise from robotic work, provided that rules and guidelines are clearly defined in the software.

Now, AI (Artificial Intelligence) is fundamentally different from RPA (representing “the next step in the evolution ladder”) since it does not merely repeat a task based on a given set of rules, yet it learns from having done repetitive tasks and is able to make subjective decisions that will improve the initially established process as base rules.

AI is no longer Sci-Fi, but reality.

While RPA is programmable, meaning developers will code instructions that define what and how to act within compliance of business or process established rules and regulations; AI can resort to several ways of surrounding environment information recognition (visual, sound, data query over the Internet or other information clusters, other) to perform cross analysis and the decide on a given course of action.

With AI, we step into a “Machine Learning” context, where (instead of coding instructions) developers “train” the AI to acknowledge certain patterns and make a decision upon that gathered knowledge.

A simple example is how to train AI to recognize dog’s pictures and not to mistake a dog for any other similar animal; so, the developer will not define a set of rules like four legged, furry, with a tail and black nose, etc… The developer will instead start by presenting the AI with literally thousands of pictures of dogs. The AI will “look” at mixed pictures of dogs and other similar animals and every time it identifies some other animal as a dog, the developer will “tell” it to be wrong. Within several cycles, the AI will become more and more accurate until either the error margin is lower than if a human was making the identification or equally accurate yet much faster.

So, while RPA is dependent on maintaining the surrounding environment as is (problems will potentially translate into stoppages), AI’s strength lays exactly in the fact that it is designed to adapt to changes in the surrounding environment (problems constitute an opportunity to evolve and become better) while also being capable of performing what RPA does, but in an adaptive manner.

New potential Outsourcing offering portfolio

Besides both examples mentioned at the beginning of this article, let’s now undergo some creative thinking and lit some potential improvement that may result from Outsourcing to AI in a near future:

  • On demand Data-Consolidation – Some relevant information for corporate or even public service purposes resides in Big Data platforms that are neither consistent nor structurally equivalent.

The solution up until now was to build up enormous data warehouses and start a long project of migrating data from those several sources to a new central information cluster. With AI companies will be able to both extract the relevant information as well as optimize the new information cluster structure in a manner that not only best suits current business requirements but adds on-demand flexibility towards potential evolution scenarios.

Instead of having a project of 6 months to one year with several consultants walking around your premises while getting people from your team to join endless meetings where the project path will over and over be fine-tuned; you define the rules, launch the AI, coach it according to several examples and once it is able to clearly identify what is what and why within your data structure, release it to “do its thing”.

At the end, the AI will present several improvements and a flexible information cluster structure.

  • BPO onshore potential – The BPO model basis has been for decades to make a profit based on offshore growing HR that represent cheap costs.

Now being able to create an AI based Outsourcing portfolio is, in fact, more relevant if done onshore since the potential for disruption due to communication lines failing or power outages is much lower if compared to some other geographies with additionally no time zone constraints.

The potential for change is such that some well-known Outsourcing providers that have their teams core based in offshore locations (e.g. India) such as WIPRO, COGNIZANT, TCS, INFOSIS, other … have either acquired AI development companies, existing platforms or event started to develop their own AI platforms not to lose their BPO edge in the marketplace.

Financial analysis points that in such scale companies’ an RPA or AI could cost as much as 1/3 of an Offshore employee cost towards the same BPO set of tasks.

  • Efficiencies Fulfilment – The fact of the matter is that under their quest to satisfy investors regarding growing profits, big client corporations are continuously looking for efficiencies that allow cost reductions without compromising core business.

Such goals are mandatorily transmitted along the value chain to their suppliers and providers who basically need to do more while receiving less.

AI bears a high potential of automating well known and established processes allowing the client company to get the efficiency that was committed under existing contract without negative core business impact (in fact most likely an improvement), while getting a lower price tag for the service and all of it with no negative impact towards the provider’s operating profit.

There is a huge Win-Win potential here.

The above-mentioned companies have gained some contracts in the last 5 years to U.S. and European based long term established competitors at a much lower price since they were using offshoring low-cost service delivery centers towards some of the services under contract, nevertheless despite being cheaper those existing contracts still have the “efficiency gain targets” which means that the solution now is to go RPA and AI.

  • Client Indirect Operational Costs reduction – Some existing Outsourcing contracts require on-site workforce from the provider to assure service delivery as per established KPIs.

Besides the labor cost of such resources, each human being needs a workplace (space) which comprehends cleaning, power, furniture, maintenance, water and other inherent indirect operational costs.

Needless to say, that AI simply discontinues such costs, even local power, since it shall be provided from a remote Data Center.

  • More assertive KPIs – Most of the current Outsourcing contracts have KPIs that aim at measuring efficiency, based on a number of dedicated FTEs.

The point is that having 5 guys dedicated to a given BPO service is fallacious in two manners: to begin with they often are in fact shared with other clients or project in order to compensate for established efficiencies and maintain profit levels, and having 5 does not necessarily mean that performed work will be twice as good as if only 2 were involved (humans do not represent linear metrics).

Now if the KPIs move to outcome based measures, where results are measured instead of involved resources while those resources are AIs, the client at the end get a more accurate view of how the service is being efficiently rendered.

  • Intellectual property and business support continuity – In most big BPO outsourcing contacts, the human factor is predominant with regards to specifically tailored expertise.

Meaning, some dedicated experts bear specific knowledge about client’s processes and IT Systems that are not documented or it is very hard to document in detail which makes that handful of people highly valuable for the client (as well as the provider).

The problem is that humans have both feelings as they tend to eventually retire or die.

This may seem to be a very cold direct approach, but having those types of resources is not a blessing but a potential disaster waiting to happen.

Now AI has the potential of dually tackling this problem for if in one hand it can outlive even the company lifespan on the other it directly assures that the “intellectual” property belongs to the client corporation since it is directly connected to such corporate reality and context.

Near future facts

  • AI and RPA will have a tremendous impact in 2018 onwards regarding the way companies operate and buy BPO services.
  • Outsourcing will undergo a dramatic change both in terms of processes, the Offshore/ nearshore/ onshore balance and scope
  • Contracts will need to be fully re-written according to these two disruptive technologies.

Soon, RFPs will have a strong AI component

Previous Article What is sensible to Outsource
Next Article Outsourcing Sales