Most of the rules of the corporate world were written in the 1960s, when employees were separated into the “haves” and “have nots.” The “haves” got Madmen-style perks and benefits, including dedicated support from executive and administrative assistants, to ensure that they could do their jobs without being unduly bogged down. The “have nots” didn’t rate this support and had to managing their own schedule, priorities, and line up for shared services to get anything done.

In the working world today, this pattern still holds true with executives reaping the benefits of dedicated Chiefs of Staffs and Executive Assistants while ordinary workers are left to fend for themselves. At the same time, the pace of work has increased, with ambitious OKRs, unannounced meetings, and distracting one-off tasks.

Xembly was founded to bridge the productivity gap by supporting the everyday needs of knowledge workers. Tasks like scheduling meetings, taking notes, and managing one’s own working hours may seem trivial, but they collectively drain a worker’s capacity to focus on the important conversations and workstreams needed to excel in their jobs.

What is an AI Chief of Staff? This is a tough question to answer as this emerging class of AI-enabled software is rapidly evolving. Think trying to envision what a smartphone would become in June 2007 when the iPhone was first released.

What is an AI chief of staff (ACS)?

As a starting point, it’s important to understand that an ACS cannot and will not replace the need for a truly talented human Chief of Staff operating at the highest levels. Humans are capable of complex reasoning and abstraction that no AI model can currently (or likely ever) match.

The Harvard Business School presents a three level model for how a Chief of Staff might function which serves as a helpful contextual backdrop for this discussion on what an ACS is and is not. See table below for an abbreviated version of their 3 Levels of Chief of Staff model.

Level 1 Level 2 Level 3
Chief of Staff Role Help the leader become better organized Manage important projects well Help executives navigate through uncertainty and risk
Chief of Staff Capabilities -Understands the business

-Can do project management

-Can manage relationships

-Communicates well

-Organizes the executive’s office

In addition to level 1 capabilities…

-Can simplify complexity

-Does strategic thinking & problem analysis

-Can manage the process of idea to execution

In addition to level 2 capabilities…

-Can anticipate and avert problems

-Can grasp and add value to the leader’s vision

-Has organizational and political intelligence

Source: https://hbr.org/2020/05/the-case-for-a-chief-of-staff

At a starting point, an effective AI chief of staff should be able to handle the Level 1 responsibilities outlined in the chart above. The ACS may not be capable of managing direct reports or providing long-form analysis, but should be able to communicate, help the leader become more efficient, and maximize efficiency with minimal change – all through regular, transactional, brief interactions with the leader.

How does an AI chief of staff (ACS) work?

Here are four principal guidelines for how an ACS should function:

An ACS functions proactively – not just reactively.

What separates an ACS is its capacity to function proactively – repeatedly offering suggestions or even completing tasks without user prompting. This level of automation requires both a generalized AI understanding of how work is accomplished as well as a narrower understanding of the user and how they engage their everyday work.

This proactive working model contrasts with most project management tools which require users to input most if not all changes.

 

An ACS learns over time.

For an ACS to be able to function proactively, it will have to learn the user’s preferences. Some of this could be completed in an initial intake experience, but much of this learning will happen over time.

This is both an explicit and implicit process where the software can routinely offer the user opportunities to provide direct feedback while also simply picking up on user cues (e.g. a feature’s engagement rate).

An ACS must work primarily within the context of existing workflows.

Similarly, an AI chief of staff would be expected to operate primarily in the framework of a user’s existing tech and communication stack in order to minimize the learning curve and maximize a user’s potential engagement rate.

In the modern working world, the most common systems one might expect an ACS to operate through would include the user’s email, calendar, and chat-based collaboration tool (e.g. Slack).

An ACS must simplify work – not create new work.

An ACS must be ruthless in its elimination of time-draining, focus-expending task bloat.

Person writing notes next to computer.

Source: Unsplash

For instance, prompted suggestions an ACS might offer (e.g. would you like me to move your meeting to Th 3pm?) should operate out of a basic yes/no or poll-based framework that minimizes the cognitive load induced upon the user.

Who needs an AI chief of staff?

Looking through a long-term time horizon, it’s possible that every knowledge worker could be a potential user of ACS software in the same way that every worker currently uses email. In the short-term, constraints (e.g. IT budget for software, willingness to test new tools, ACS’ current shortcomings) will limit the use of ACS to a wave of early adopters who are likely to see and realize the immediate benefits of the ACS value proposition.

Characteristics of individuals, teams, or companies which may be great fits for ACS adoption include:

    1. High meeting load – The function of scheduling and rescheduling meetings is such an ongoing tedious workload that it’s often become its own job in larger companies. In addition, the process of note taking can drain a person’s ability to focus in a meeting. Minimal sharing of notes can also limit the organizational impact of a meeting’s strategic thinking or decisions. Individuals or groups that suffer from meeting mayhem are ideal candidates to adopt ACS.
    2. Multiple workstreams – For individuals with multiple and changing priorities, an ACS can help them stay focused and balance competing priorities. This is readily apparent with roles like product management that naturally intersect across a range of business functions.
    3. Tech savvy – As AI chief of staff software is still in its infancy, tech savvy companies willing and accustomed to testing new tools are best positioned to quickly integrate an ACS into their existing systems.

Person on computer.

How is an AI chief of staff different from existing productivity solutions?

AI Chief of Staff Legacy Software Solution
Initiative Proactive Reactive
Primary Interaction Model Conversational Manual input
Scope Generalized (touching many common tasks) Narrow (touching only a few or one tasks)
Platform Primarily dependent (leveraging existing systems and processes) Primarily independent (creating new systems and processes)

What is the best ACS solution out there?

The current market for ACS solutions is rapidly evolving.

Person writing notes next to computer.

Source: Unsplash

As a starting point, many larger incumbent players have found product market fit in distinct functional areas such as project management, scheduling, calendar management, meeting note taking, but very few tools have a vision for a broader cohesive solution that ties together these various activities under one assistant who can support the leader.

Xembly is at the forefront of innovating in this space alongside other providers who hit on parts of the entire package. In the coming months and years, we expect and hope to see many competitors providing differentiated offerings in this space, validating the market opportunity for everyone and creating daily wins for the end consumer.


Key Sources