The jury’s out in how many white-collar office or professional jobs will be displaced by artificial intelligence. But one thing is certain: many of these jobs are being supplemented or augmented by AI copilots and agents.

That’s the word from Angela Strange and James da Costa, partners with Andreessen Horowitz, in a recent note on emerging opportunities in the AI economy. “Every white-collar role will have an AI copilot,” they state. “Then an AI agent.”

What does this mean? Their implication is that while AI won’t necessarily replace a lot of office and professional jobs, it will certainly handle a lot of these workers’ tasks. In many cases, at least half of their tasks are open to automation. Left unsaid is the freeing up of workers’ time means those that can move forward are those that take on a more activist role within their businesses — advising, teaming up, and even identifying other areas open for AI assistance.

Strange and da Costa point to last year’s study out of OpenAI and the University of Pennsylvania to illustrate their point how many professionals are ripe for AI copilots and agents. The study showed that with access to a large language model, “about 15% of all worker tasks in the U.S. could be completed significantly faster at the same level of quality. When incorporating software and tooling built on top of LLMs, this share increases to between 47% and 56% of all tasks.”

Inspired by the OpenAI-University of Pennsylvania research, Strange and da Costa pulled employment data from the U.S. Bureau of Labor Statistics for 2023 and identified the top 50 roles in which 50% or more of tasks could be performed by AI. These roles include architects, financial analysts, meeting planners, insurance underwriters, HR assistants, and customer service representatives.

While Strange and da Costa are targeting their observations at startups willing to market AI-based solutions for these purposes, these are directions managers, professionals, and companies at all stages need to consider.

“The most natural place for these copilots and agents to live is the incumbent workflow or system of record — such as sales agents launched from Salesforce,” they point out. “The system of record is where the data agents need to complete specialized tasks lives, and it’s also a natural launchpad for any new user interface to reside — prompting the agent.”

Here’s where copilots and agents seem destined to take on more of the heavy lifting:

Data collection. “Data for loans or insurance policies is still often collected via email and PDFs,” Strange and da Costa stated. The key is to “AI-ify this workflow and own the data before it gets to the incumbent SOR.”

For example, they continued, “a virtual loan officer or insurance agent could own the initial back-and-forth customer document collection and appointment scheduling. Similarly, virtual sales development representatives (SDRs) can gather all the information about a potential customer and own the initial correspondence before a record is even created in the incumbent SOR.”

“Painful workflows” being slowed by requirements for outside information. “There are few things more tedious than the Know Your Business (KYB) onboarding process in banking, which involves document checking, internet searching, and back-and-forth correspondence between businesses and financial institutions,” Strange and da Costa observed. AI providers or platforms “will auto-parse every document that’s uploaded, extract the needed information, and follow up with the customer for missing information.”

Healthcare, another industry with painful workflows, is also ripe for AI-ifying, they continued. AI systems could “take in every medical document hitting a fax machine, extract patient and diagnosis details, and even run insurance pre-qualification to streamline patient visits to medical practices.”

Integrating the right data. Disparate data sources need to be converged “to create a new multimodal system of record,” the VC partners urged. “Significantly more data exists and is relevant to the job to be done than what is currently held within incumbent systems of record. For example, sales data doesn’t just exist in Salesforce or Hubspot: there are also emails and Slack messages, sales enablement materials, product usage data, customer support records, news and financial reports.”

Integrating these unstructured data sources can lead to new systems of record that will fuel AI copilots and assistants, Strange and da Costa concluded. This data — text, image, voice, and video data — forms the foundation of today’s white-collar work.