Is Accenture’s SynOps automation solution a killer app or a job killer?

Profile picture for user jbowles By Jerry Bowles February 17, 2019
Accenture is now selling SynOps, a combination of AI software and services that it has already used internally to automate tasks and redeploy about 40,000 workers.

One of the surefire ways to tell if a new enterprise solution is a potential game-changer or not is to look at the marketing rollout and observe the linguistic contortions the company goes through to persuade potential customers that “this is going to save you millions of dollars” without alerting policymakers that “this is going to mean thousands of workers are going to lose their jobs.”

It is an awkward dance that is likely to get more and more frequent as automation continues to spread like a tsunami washing away thousands of jobs (the marketers prefer the word “roles”) in its wake. Consider this impenetrable paragraph from the official Accenture press release:

SynOps is an assembly of talent, capabilities and technologies, including artificial intelligence (AI) and advanced analytics, that gives organizations a fit-for-purpose and flexible roadmap for achieving higher levels of efficiencies across the enterprise. SynOps enables companies to make their organizations more flexible, agile and responsive, helping them generate additional value by driving growth and scaling operations quickly.

What the press release doesn’t say is that SynOps is a system that Accenture has developed and used over the past five years to automate processes in areas such as finance and accounting, marketing and procurement and one that has allowed it to eliminate 40,000 “roles.”

The SynOps platform—which Accenture helpfully calls “a human-machine operating ‘engine' " sits on top of a company’s existing databases and record-keeping systems and uses artificial intelligence, advanced analytics, and human insight to optimize business processes and improve returns. It can do things like predict which invoices are likely to be rejected because their terms don’t match the contract, and in some cases automatically correct those invoices before sending them on to a customer.

When used in procurement, the system can take an order, generate an invoice, check that invoice against a contract, correct any errors and then email it to the customer without human intervention.

Clients can set custom rules for whether the system will process something automatically based on how much confidence the AI software has in its own predictions. For instance, if a system were handling routine insurance claims, a company might allow it to automatically pay those claims as long as it had at least a 95 percent confidence that the decision was correct. If its confidence falls below that threshold, the claim would be routed to a human.

The system also allows Accenture’s clients to benchmark how efficient their processes are against other companies in the same industry or across industries.

So far, about 100 Accenture clients have used the system for procurement, with around 20 having tried the finance product and five clients currently piloting the marketing offering. Said Nirav Sampat, group technology officer at Accenture:

If you are a chief procurement officer of a company, you are really looking at how to get the best out of your money and spend. For a large beverage company, one of the things we did was look at all the indirect spend and really look at spend analysis in order to optimize what they needed to have. Then, we got all the information to the buyers, so they have all the right information to make the decision versus really having to look at a few things and make a decision for themselves. And, if you put it all together, it helped them to save millions of dollars.

The human element

Accenture spokespeople have gone to great lengths to stress that those workers whose “roles” have been displaced were retrained and redeployed and that Accenture has more employees now than it had five years ago. Debbie Polishook, group chief executive officer of Accenture Operations, the company’s outsourcing unit, which once used human workers in mostly low-wage countries such as India, to handle routine data entry and customer service tasks for clients, offered the standard bromide:

This is not trying to get rid of the human but to make them as productive as possible and get them to focus on the work that a human really needs to do.

That is rapidly becoming the corporate equivalent of “No animals were injured in the making of this movie.”

Clients who buy SynOps will get to make their own decisions about whether to retrain and redeploy workers or take the extra money and run with it. Human nature being what it is…yada yada yada.

My take

As Phil Fersht recently opined on the topic of RPA:

...while some of his (Blue Prism CEO) competitors have been in stealth mode, raising all sorts of private investment and offering licensing models that appear (on the surface) a lot cheaper, while selling the “This is easy, this is no/low code, we can train you in weeks and get you a nice certificate to share with your friends on LinkedIn”.  This is what I personally detest about the software business… anyone can sell dreams, confuse executives too scared to ask critical questions like “how exactly does this work again?” especially when you have the lovely term “robotics” to excite greedy CFOs and CEOs eager to find new ways to increase margins...

...Now it’s all about stitching the wonderful skills of building scripts, macros, document processing and screen scrapes with the emerging excitement of Machine Learning, Natural Language Processing, Augmented Reality and Computer Vision.  Yes, folks, you thought the hype-train of the past 30 years was bad, the one we’re venturing into is going to drive many of us completely nuts.

Back to SynOps.

Despite the innocuous name and low-key introduction, SynOps is a very big deal. By adding AI and machine learning capabilities to standard robotic process automation this newly evolved category of automation tools analyze prior decisions and actions, learn over time and then get smarter and more intelligent at making decisions.

SynOps is just the latest example of the way Robotic Process Automation (RPA) is transforming the enterprise. McKinsey research estimates that RPA, in which software is used to handle repetitive data entry and other routine tasks, will have an economic impact of more than $6.7 trillion by 2025.

For large companies with lots of costly “repetitive cognitive” processes that can be automated, AI-enabled RPA is the next stop on the road to the holy grail of “intelligent operations.” For low-skilled workers, it is a one-way ticket to Palookaville.