As head of AI at work management platform monday.com, Or Fridman stands at the forefront of an ever-evolving field. As artificial intelligence gradually transforms the workplace and becomes integral to project management tools, Fridman said driving AI innovation can often feel like a race to stay ahead.
“It is a race, but it’s a fun race,” he told TechRepublic.
For Fridman and monday.com, however, it is important to get the deployment of AI technology right. Their platform is used by 225,000 customers in 200 industries worldwide, often by non-technical teams who leverage monday.com to carry out daily tasks. As he explained, the way AI is rolled out is critical.
“What is important for us, and for me personally, is to do it the right way,” Fridman said on the sidelines of the Elevate 2024 conference in Sydney. “But what’s important for us is to actually see people using [AI features] and to remove the barriers from using them.”
Making AI user-friendly and seamlessly integrated into workflows
After the launch of monday’s AI assistant, which was more of a rapid reaction to ChatGPT’s release, Fridman said monday.com began thinking more deeply about how to integrate AI. The company began seeking more customer insights into perceptions of AI and what customers were using it for.
SEE: Our 2024 review of monday.com, including pricing, ease of use, pros and cons
“We really found out about the adoption challenge, and the fear that existed of being replaced, and the hesitation to trust this thing,” Fridman explained. This led to monday.com’s current approach to AI, which is really to follow much of the same recipe with products it has been using to date.
He added: “We took an approach similar to what monday.com has done for software in general. monday’s approach is to democratize software, to reach out to different types of people, tech-savvy and non-tech-savvy, and allow them to use tools in work to do their job better.”
This approach allowed the company to follow its previous strategy of creating low- and no-code interfaces — previously used for automations in monday.com — that carry monday into a new AI era. monday also now talks about “no prompt or low prompt” actions, Fridman said, which make it easy for non-technical users to utilise AI.
Another direction was to move away from AI being used in a separate interface, such as a chatbot.
“For some use cases, it totally makes sense, and I’m using it that way as well,” Fridman said. “But we really believe people want to work [with AI] in their workflow, where the work is. If I’m managing my work in a [monday.com] board, I want to work in the board. I don’t want to work in a separate interface.”
Making AI easier with pre-packaged task actions
monday.com’s strategy is now taking shape in the platform. One plank of the strategy will be adding product capabilities to “supercharge” existing products. For example, the firm plans to introduce a risk analyser to its portfolio manager product, using AI to help identify risks within a portfolio of projects.
The other main way AI is being rolled out is via pre-packaged, AI-supported automations or tasks. Known as “AI blocks,” these elements are designed to resemble the platform’s current brightly coloured blocks for easy integration and visibility within workflows. Fridman calls them “easy-to-use AI actions that can be integrated into your workflow.”
SEE: Top 10 project management software systems in 2024
At Elevate 2024 in Sydney, Fridman demonstrated how monday’s AI can help recruiters. He said AI could be used to extract specific information, such as email addresses, from a resume document. Recruiters could also use AI to provide a resume summary or extract a candidate’s skillset.
Users will also be able to create their own custom actions using natural language prompts. For example, Fridman said a recruiter could create an action comparing candidates’ resumes with the job description, highlighting any potential matches or gaps for the candidate pool they are reviewing.
“You can see that we actually kind of automated the review process of job candidates,” he told the audience. “AI really helped us to make the review process of candidates much easier and help the reviewer and the manager really focus on their work. And this is what is important for us to do.”
Building user trust over time in monday.com’s AI outputs
While monday.com emphasises simplicity for users, it won’t limit them to only basic, pre-packaged automations. For example, Fridman said more complex and robust workflows will be possible, using different triggers and conditions in a similar way to its automation centre.
However, as Fridman pointed out, building trust in the technology is perhaps the most critical factor in user adoption. If users do not trust the actual outputs — such as getting an accurate reflection of a job candidate via skills or resume summaries — they will not use them with confidence in their daily workflows.
SEE: Our review of monday.com’s CRM
Fridman said monday.com is trying to overcome this issue by giving users control and visibility. In terms of visibility, Fridman said users will be able to see what is generated by the AI in monday.com. Control will be achieved by allowing users to correct AI results in a non-technical way, helping the AI learn over time.
“AI is constantly learning,” Fridman said. “It will learn from all of your work, from your boards, from your data. The combination of visibility and control — those are the keys, in my opinion, to meet the trust challenge.”
AI will supercharge rather than replace teams
Fridman said that, although there is a level of fear and mistrust around AI, it is here to stay. He believes that, rather than replacing large numbers of people, it is more likely to augment their capacities.
“We think that it will supercharge teams and will help them not to be replaced, but to maybe shift their work and focus on things that AI will not be able to do,” he said.
Fridman gave the example of a project manager who, in the past, would spend significant time nudging team members and collecting status updates on projects.
“It is a lot of manual work,” he acknowledged. “We believe, with AI, those types of things will be redundant, and then the project manager will be able to focus on how to orchestrate the project. They will get tools to predict what’s ahead.
Fridman concluded: “Those types of things will lead us also to new developments in the product. So for each one of our products, we will imagine how it will work moving forward. And we’ll just use AI to supercharge it.”