It’s already clear that business automation using AI is a big and lucrative topic. But here’s an interesting twist! There’s no need to create “universal platforms for automating business processes” — instead, the focus should be on creating digital AI employees “with professional experience”. All the necessary technologies for this already exist. What remains is to properly package them according to this principle.
Project Essence
Today’s startup, created in September this year, managed to get selected in the new Y Combinator batch starting this winter and raised its first $2.3 million in investments.
Artisan AI is set to create digital employees who can work as full-fledged team members alongside living humans.
The startup first plans to create a digital salesperson, designer, and marketer.
The main difference between a digital employee and a regular AI bot is that the digital employee is “out of the box” trained to perform entire task chains related to their area of expertise, interacting via API with necessary cloud services like CRM, email marketing platforms, social networks, email clients, messengers, and everything else required for their work.
You can configure the digital employee and monitor their performance statistics through a special dashboard.
Assigning tasks to the digital employee and asking them questions can be done via a regular messenger. In this sense, communicating with a digital employee is no different from communicating with a remote human employee. Ideally, you wouldn’t even notice the difference between the two 😉
A digital employee can even write to you on their own initiative—if they encounter a complex case they can’t handle, or when the information they have obtained or found requires your immediate action.
In the process of communicating within the team and performing assigned tasks, the digital employee constantly learns by tracking the effectiveness of completed tasks and mastering new actions.
In early December, the startup plans to release a digital salesperson who can replace 10 people working on outbound client communications but will cost 10 times less than one human employee.
The digital salesperson will have access to a constantly updated database of 265 million company employee contacts, who could become potential clients of the hiring company. They will be able to select the most suitable contacts for communication using 65 filters, applying them as needed depending on the task at hand.
The digital salesperson will compose and send letters (or messages) to selected contacts, personalizing the texts and refining their writing techniques based on the responses or lack thereof to already sent letters.
In addition to sending letters, they will be able to process responses—sending a new letter based on the received response and gradually intriguing the interlocutor in the product being sold, leading them to an online meeting with a live salesperson.
Once a response from a potential client indicates they are sufficiently “warmed up,” the digital salesperson will schedule a meeting with a live sales department employee.
In doing so, all letters will be sent in the name of a real person. When scheduling a meeting, the digital salesperson will inform the live employee of all important details of the previous correspondence, so they are aware of what they have been discussing with this client 😉
There’s no need to specifically instruct or ask the digital salesperson to take the next step in communicating with clients. It’s enough to set a general task like “find so many potential clients according to these criteria, offer them this thing, and schedule as many meetings as possible”—after which they will start working on autopilot, adding meetings to the calendar and contacting the live sales department employee only in case of any problems or questions.
You can set up a new digital salesperson in 10 minutes, and within an hour they will send the first letters to potential clients.
What’s Interesting
I wrote about a very similar startup, 11x, which raised $2 million in investments in its first round.
It also creates digital employees and started with a digital salesperson, but next in priority for them are a recruiter (already released) and a customer service employee.
Artisan AI, conceived after 11x, does practically the same thing—but despite this, got into Y Combinator and raised $2.3 million in investments at the idea stage. In my opinion, this is another great confirmation that good and timely ideas can and should be copied 😉
In essence, all the individual actions performed by a digital employee can be done using various AI tools—for example, asking ChatGPT to compose a personalized sales letter, and then manually sending it to the client.
However, any business is a set of business processes. Therefore, companies need not “a set of different tools,” but “a platform for automating business processes.” A digital employee is essentially a constructor of business processes, in which a set of specific tasks for this category of employees can be arranged in a sequence suitable for a given company. The picture above shows the constructor window for setting up the work of a digital recruiter on the 11x platform.
Such narrow specialization fundamentally distinguishes digital employees from universal business process automation platforms, which can also use modern AI tools as building blocks. Although such universal platforms also turn out to be in demand—just recall Relay, which raised $8.1 million in investments, as I wrote about in October.
From an engineer-programmer’s point of view, a universal platform is much better. But from the point of view of a client company, this may not be the case. Because a universal platform is more difficult to configure to solve a specific task, and it may not have some important specific bricks for this type of task.
And hiring a digital salesperson or digital recruiter for a company may turn out to be much more understandable than paying for “a platform for automating work that can do more than others because it uses AI assistants, allows organizing interaction between AI and humans, and also involves several participants in the process” — as Relay describes its platform on its website 😉
In general, this is very similar to the fact that companies do not like to hire even smart beginners who still need to be trained in the profession—but prefer to hire people who already have relevant work experience, even if they are not as smart in the general human sense 😉
Where to Run
The general direction is to wrap a set of specific skills for a given category into the packaging of digital AI employees.
All such tools already exist separately. Therefore, the whole trick is in creating “digital employees with corresponding work experience,” who are ready out of the box to perform all the necessary actions—which only need to be instructed to perform them in the right style, in the right sequence, and interact with the platforms used in a given company.
Let’s take a very simple example.
This month I wrote about the Abyssale platform, which helps automatically create different versions and formats of advertising graphic and video banners based on a once-created designer template—allowing, according to the startup, to create banners 10 times faster, without hiring additional designers.
But why not offer companies to hire a “digital designer” instead—who can do 10 times more for a tenth of the salary of a live designer, just like in today’s Artisan AI formulation? After all, it’s enough to take the same technological basis, slightly spice it up with magical AI sauce, and repackage it. At the same time, it’s not excluded that such a digital employee may even sell better 😉
There are actually a ton of such packaging and repackaging options. So the question is—what kind of digital employee would you create? What should they be able to do? What existing technologies and tools would need to be inserted into them for this?
About the Company
Artisan AI
Website: artisan.co
Last round: $2.3M, 11.15.2023
Total investments: $2.3M, rounds: 1