Nearshore Americas

Case Study: Instacart’s Massive, Automation-Enabled Agent Scale Back

A lot of text has been dedicated to chronicling and analyzing Instacart’s boom in the midst of the COVID-19 pandemic, but little is told about how the company’s back office handled skyrocketing demand.

Instacart –a grocery delivery platform– was one of many digitally native businesses which hit it big during COVID. The company reported that revenue more than doubled in 2020, hitting US$1.5 billion, as well as a 500% increase in orders following the first months of lockdowns. 

Such a massive jump in orders required an equally massive increase in the company’s capacity to provide customer service to a group which included people who buy groceries through its platform, shoppers/couriers and (to a lesser extent) retailers.

At the time, Instacart sourced CX services mainly from three BPOs and had 1,200 agents providing service to the three groups of users mentioned above. Following the boom in orders spurred by the lockdowns, the company had to scale up to 20,000 agents in just three months, according to a source formerly employed by Instacart and who was directly involved in the scale up.

“I don’t think that’s normal for anyone. BPOs were losing so many clients from travel that they were happy to have some business, and they moved fast with us,” the former employee, who asked to remain unnamed, told NSAM in an interview.

NSAM reached out to Instacart to determine if these numbers were accurate. The company replied that, in fact, it “scaled its customer experience team significantly at the onset of the pandemic […] to ensure we could meet the needs of our customers, shoppers and partners.” Nevertheless, it stated that the volume of agents peaked at 14,000.

Instacart was forced to dramatically increase its partnerships too. The company jumped from 3 to almost 20 BPO vendors during that three month period, expanding the presence of its agents to 46 cities. 

A ramp-up of such scale would be a nightmare for any company, yet it seems that Instacart was able to navigate the situation with less difficulty thanks to its own nimble nature at the time.

“Instacart, even though they’ve been around for 12 years, is still a startup, so that traditional red tape was not necessarily there in terms of the training, the VPN. We had that, but we were able to move fast. It wasn’t like we had all these layers of leadership,” the former employee explained.

As impressive as the scale up was, Instacart leadership would find itself executing an equally massive move in the opposite direction but a couple months down the line. 

Scale Back #1

At the same time as Instacart was scaling up its CX capabilities to meet exploding demand, the company was automating its self-service, chat and email channels.

Turns out the automation worked like a charm, leading to a massive cut in the number of agents required for delivery: from 20,000 to 10,000 in a couple months. 

“It took about three months to fully get to 20,000. By month two, we started working on the automation”, NSAM was told by the former Instacart employee. “As soon as we hit 20,000 we had to scale down. We reached the number, but now we have the automation. It’s working, it has to stay; now let’s go down.”

“The ramp down took probably two months. From start to end, the up and down, it probably took five or six months,” the source added.

In the same vein, the company scaled back on its BPO partnerships. From nearly 20, it ramped down to only 8, with agent presence going from 46 cities to 18. 

Instacart refused to share data on the size of this scale down. The company did recognize, however, that “overtime, our support needs changed and we adjusted accordingly.”

Scale Back #2

Already well established as a platform, and with booming demand somewhat under control, Instacart tried its hand at another experiment with automation: chatbots.

The process took considerably longer than simple chat, email and self-service automation. While the latter took a couple months, the proper implementation of chatbots required about a year, mainly due to lots of beta testing, our source explained.

The results were massive, though. Instacart was able to scale back its number of global agents even further: from 10,000 to 6,000. 

Instacart’s case shows the scale at which automation can impact customer care operations in an industry as intensive and potentially troublesome as retail. At least when applied to more simple channels (email, chat, self-service), the results can be staggering.

With a new wave of automation technologies powered by generative AI taking over the CX industry, it remains to be seen if scale backs of the magnitude executed by Instacart will be achieved.

Players as big as Teleperformace, Accenture and Concentrix are already leveraging the power of those technologies to upgrade their services portfolio and, potentially, pull off their own scale backs. Teleperformance told investors that it expects its in-house GenAI solution to reduce total volumes by 20%-30% over three years.

Sign up for our Nearshore Americas newsletter:

Business leadership remains skeptical. For every industry expert that trumpets GenAI as the biggest, most relevant thing in CX at the moment, there are at least a couple who see a technology which, though promising, has a long way to go to achieve levels that are optimal enough to justify massive deployment.

AI has proven to be fast-paced in its develoment, though. Don’t be surprised if the technology is able to traverse that long road in a relatively short time.

Cesar Cantu

Cesar is the Managing Editor of Nearshore Americas. He's a journalist based in Mexico City, with experience covering foreign trade policy, agribusiness and the food industry in Mexico and Latin America.

Add comment