Which work must be AI Automated?
![Javi Sanchez](/_next/image?url=%2Favatars%2Fjavi.jpg&w=64&q=75)
Javi Sanchez
@jvrsanch
![Which work must be AI Automated?](/_next/image?url=%2Fblog-assets%2Fwhich-work-ai.png&w=1920&q=75)
Work Before AI Existed
Before modern AI existed, for any task requiring the slightest amount of reasoning you had 4 options:
1. Do it manually
When a workflow is critical, waste working hours performing the task.
2. Outsource the work
Large corporations outsource the work to a firm that do it at scale for cheaper. This is a viable option but it's hard to govern the quality of the outputs and there is significant overhead in getting an operation like this up and running. Besides, communicating with the outsourced firm is a time consuming process and often requires going back and forth because they lack your business' domain knowledge.
3. Build custom software
When work desperately requires automation, many organizations will opt for creating custom software in-house.
This has been the most scalable approach historically, but comes with major issues. It's extremely expensive, drawing engineering power away from the core business. It is time consuming as it requires significant communication among teams to specify requirements and plan the development.
4. Don't do it
Unfortunately the most frequent path. There's tons of work you wish you could do at scale but haven't yet because automating didn't seem feasible.
What should you automate?
We like to break this down in two questions:
Is this worth automating?
Value = Repetition x Pain x Potential for Growth
-
Repetition: If a task is performed frequently enough to impact productivity, it might be worth automating.
-
Pain: Consider how much time the task takes and how painful it is. Sometimes a task is performed 100 times a week but only takes seconds each time. This work is a mild annoyance and probably not worth the effort to automate.
-
Potential to Grow: Does this use case have the potential to become more common as you grow (e.g., support tickets), or do you want to encourage it to become more common (e.g., personalized sales outreach)?
Can AI handle the task?
The golden question. As a rule of thumb:
"If you can describe it as a list of steps, like you would for a new employee, it can be automated"
AI has revolutionized software. Previously, automating data extraction from a PDF required an exact format; a slight change broke the automation. Now, AI is capable of human-level understanding, unlocking a huge amount of skills for computers, previously accessible to only humans:
- Read, understand and extract data from PDFs, Word or PowerPoint, without rigid formatting rules.
- Search on Google and extract information from the results.
- Understand and extract information from audio and video.
- Data Analysis: Derive insights from diverse data sources.
- Content Generation: Produce written content, images, audio and video.
- Language Processing: Translate, summarize, transcribe and interpret text.
Some example use cases:
- Automate data entry from diverse document types into spreadsheets.
- Extract and categorize information from receipts, resumes, and customer feedback.
- Write proposals, contracts or emails based on company data.
- Populate and update various management systems (CRMs, ERPs, applicant tracking, patient records, etc.).
If you feel you can leverage AI to automate your company, talk to us!
Or explore real-life case studies and how companies are using AI to automate and augment their work.