Adding AI to Workflows? Here’s What We’ve Learned at Indeed

Key Takeaways

• Indeed is deploying Salesforce’s Agentforce to help employees focus on more meaningful, strategic work — boosting engagement, productivity and overall job satisfaction.
• Implementing AI comes with challenges. Companies must embrace the learning curve and refine processes to maximize AI’s benefits.
• Not every task needs automation. Focusing AI on high-impact areas ensures efficiency without compromising quality.

Every day, I am inspired by Indeed’s mission to help people get jobs. With AI-streamlined workflows, we can have an even greater impact on the world of work.

Like many of you, I’m finding ways to leverage AI tools to “supercharge” our workforce at Indeed, helping them become more effective in their roles. This includes enhanced productivity and time savings, improved work quality and greater employee satisfaction by eliminating low-value tasks. Ultimately, if successful, this should also have a positive impact on Indeed’s bottom line — and we’re only just starting to scratch the surface of what AI is making possible to revolutionize hiring and the job search.

For talent leaders exploring AI options to boost productivity, business outcomes and employee satisfaction, I want to share what we’re learning from our AI-based pilots. Here’s how we’re using Salesforce’s AI platform, Agentforce — one of the many tools in our arsenal of AI resources — to support our mission.

The Objective: Integrate AI for Faster Hiring

I hold two titles at Indeed: Chief Information Officer (CIO) and Chief Security Officer (CSO). As CIO, I’m passionate about making the job search on Indeed more efficient by optimizing our employee efforts. As CSO, I'm dedicated to doing so in a way that minimizes risk.

With 595 million job seeker Profiles¹ and 24 million jobs, Indeed already uses AI to ease the job-search process by recommending jobs to job seekers and matching candidates to roles. Now, we’re aiming to shorten the job-search duration by 50% by 2030, in part through adopting AI technology that will help our support teams work faster and smarter.

Enter Agentforce: Salesforce’s digital labor platform for bringing trusted, autonomous AI agents into the flow of work. These AI “agents” take on repetitive tasks from our client-facing teams, essentially forming a 24/7 digital support team so employers can move through the hiring process quicker and easier.

Say, for example, an employer notices Indeed has flagged one of their job postings. They can exchange messages with an Agentforce-powered AI agent to learn the reason, like having an overly long job title. Agentforce can suggest how to resolve the issue, allowing the employer to quickly move forward in the hiring process with an improved, high-performance job posting.

But it took some work getting here, and we picked up some strategies along the way.

Key Learnings: Adding AI to Workflows

1. Slow and steady wins the race

AI is exciting so, naturally, there’s been overwhelming interest across Indeed teams to use Agentforce. However, being intentional about where you integrate new AI tools helps you learn the best ways to train teams and implement systems — especially before scaling.

AI has the potential to reshape roles, creating new opportunities for efficiency and innovation. We chose to prioritize the productivity of Indeed’s sales and support functions, enabling these teams to better serve our customers who are under pressure to quickly make good hires. We’re investing the time to reshape our technical teams, product delivery processes and risk mitigation strategies to suit the new world of agentic products. The skills and processes required to build and manage agentic solutions can look different from traditional products, so it’s worth taking the time to get these things right before scaling.

2. Don’t leave room for error

When implementing any new AI applications, we have to manage potential risks. This means establishing standard design patterns, governance and testing practices. Since 2020, we’ve remained committed to our responsible AI principles, updating policies as needed with technological advancements. 

When our clients interact with AI systems, they'll push boundaries: that’s just the reality and risk of using AI-powered technology. So we must monitor how we securely provide data to these systems, ensuring AI agents don’t act in ways that contradict our core mission and values when they engage with the customer. For example, Indeed sits on more than 320 terabytes of data on employer and job seeker behavior and connections. Of that data, we integrate what’s relevant to our sales and marketing goals, providing context to Agentforce’s AI agents. This creates a unified view of our 3.5M employers, which we continuously enrich with fresh data.

3. Show your teams what’s in it for them

Perhaps most importantly during this process of integrating Agentforce, we’ve been transparent about our goals and vision for transforming our go-to-market (GTM) operations with AI. 

It's critical for employees to see how this transformation will benefit them personally, like reducing friction in their daily work life and building skills that will help progress their careers. Clearly communicating and demonstrating how it adds value to their roles also helps reassure and retain workers concerned about AI taking their roles. Ultimately, AI is about augmentation, not replacement, and the work should remain human at its core.

4. “Embrace the suck” with experimentation

Even if you do everything right, things might still go wrong. Effective organizational change management anticipates and plans for an initial dip in productivity and satisfaction as people resist change. When you hit roadblocks, it’s time to “embrace the suck” — a blunt adage I learned while serving in the military that rings true today. New risks and pitfalls come with building AI agents, so we have to adapt. This means adjusting your approach until you find what works and then applying what you've learned to achieve a better solution.

We encourage innovation and experimentation within our teams. Practice makes perfect, and providing space to explore allows our employees to adapt to new, more efficient workflows.

Moving Forward: Measure the Success of AI Implementation

When establishing new workflows, it’s important not to pursue AI for the sake of AI. Instead, look past the hype and into how the technology can actually drive important metrics for your business. For example, here are some ways we’re gauging success with Agentforce:

• Revenue per employee: We’ll measure productivity improvements with AI by our GTM team’s ability to carry more revenue per employee. 
• Revenue opportunity: We can directly uncover more net-new revenue opportunities with tools (like Agentforce for Sales) that automate early phases of the sales process, such as identifying leads.
• Return on investment: With careful and thoughtful investments in AI, we believe we can expect positive ROI. 

Through the power of AI, we can make the future of hiring and getting a job dramatically simpler, faster and more effective.

Read more about how Indeed uses Agentforce to speed up hiring under Salesforce’s customer stories.

Sources

¹Indeed data (worldwide), job seeker accounts that have a unique, verified email address.

Learn more from leaders in technology and AI:

AI, Talent Shortages and the Future of Work: Insights From Indeed’s CEO
How Indeed Uses AI to Provide Better Matching Context for Job Seekers
In Conversation: Indeed CEO Chris Hyams and OpenAI CEO Sam Altman
AI in Hiring: 5 Ways Talent Leaders Can Overcome Fear and Embrace Change