Staffing Industry Spotlight is an interview series featuring leaders shaping the staffing industry, sponsored by Ascen, a leading back-office and employer-of-record for staffing agencies. This edition highlights Karen Somal, founder of Athena CRM, who brings extensive technical expertise from her background in computer science and specialized software integration. 

In this episode, Karen shares her insights on transitioning from rigid, one-size-fits-all legacy software to custom, niche-specific workflows tailored to an agency's unique business processes. She also discusses the strategic integration of automation and AI, correcting common industry misconceptions around database readiness, and explaining why maintaining a human touch remains essential to successful recruitment.

‍

Staffing Industry Spotlight: Karen Somal, CEO at Athena CRM

‍

Francis Larson:

Karen, thanks for being on Staffing Industry Spotlight. First thing we'd like to know is a little bit about you and, most importantly, how you got into the world of staffing?

‍

Karen Somal:

I think, like everybody else I know, nobody chooses staffing. You fall into it. Especially in the UK, I don’t think anybody goes to college or university and says, “My ideal role is going to be in staffing.”

I was headhunted for a role at a company called Microdec. My background is computer science, so I’ve been technical throughout. I was headhunted for one of the first global ATS and CRM systems in the UK for recruitment and staffing agencies. They were the top player back then. They had a system called Profile, and they had the larger enterprise clients: Robert Walters, Michael Page, and all the big names. That’s how I fell into it.

‍

Francis Larson:

You started out in software for staffing and recruitment agencies. I’m assuming the experience there is what led you to found Athena.

‍

Karen Somal:

Yes. I’ve always been on the tech side, and I’ve always remained in recruitment tech.

One of the big problems I had was that when I was headhunted for this role, I was told I needed to have recruiting or staffing experience, which I did not have. They loved my technical abilities, so they put me on a really intensive program where I had to sit in the recruiter’s seat for a couple of weeks and learn the recruitment processes.

At that point, for me, recruitment was, “You’re looking for a job, you give your CV,” the typical permanent recruitment flow. I had the shock of my life. I learned about temp, retainers, executive search, and contract, which back then I thought was the same as temp. It was a shock to the system when I realized what the staffing world was like.

I always remained on the technology side. From there, I was then headhunted for Bullhorn Systems Integrations, a company they were working with. At that point, I think Bullhorn wanted to focus on licenses and take out the professional services side, so I was headhunted to work for a company there. I helped that business grow and put in the implementation processes.

Years later, I decided to step out on my own and launch my own business. It was because of that and all my experience working with small agencies, larger agencies, and big enterprise agencies like Manpower and Kelly Services. It sat with me that there was a massive gap. People were spending millions, not thousands, on implementing new tech, CRM, and ATS platforms. They were going live, and they still weren’t working. There was something fundamentally missing.

I remember saying years ago, “I’m going to build my own recruitment ATS and CRM,” and everybody laughed it off as a throwaway comment. Then one night, I was thinking, “After working with agencies of different sizes and different industries, why is it just one-size-fits-all? Why is there not a solution where people can build a CRM based on their recruitment process or business process?”

You have to shoehorn everything you do as a business into your CRM, and it’s half-baked. That’s when I started thinking, “Let’s do some research and think about how I see a CRM and ATS working.”

I designed it all out and spent six months researching what was out there. The only thing that aligned with what we did was Salesforce. Originally, I was thinking about building it on Salesforce, but Salesforce comes with a really heavy price tag, and you’re stuck with the Salesforce ecosystem. So, I decided to build my own from the ground up.

Originally, it started as a hobby, but it evolved and evolved. Everyone I spoke to for advice said, “Don’t make it the way you’re making it. It’s going to be really hard to support.” I understand what they mean, but we’re very consultative in our approach. We have to be because of what we have built for clients.

‍

Francis Larson:

I’d like to double-click on that. You mentioned that one-size-fits-all is how previous ATS and CRM systems would do things. How is Athena doing it in a way that is more customized to the workflows of agencies today?

‍

Karen Somal: 

The technology we’ve built gives you the tools to build what you need.

When it comes to workflows, one of the things I realized is that bigger organizations, and even smaller organizations, might start off doing permanent recruitment with one or two people. Then they decide they want to start working on retainers, RPO, or move into contract. They’re stuck with that rigid workflow because that’s what their CRM provides.

What we’ve done is allow people to build as many workflows and pipelines as they need from the ground up, and then customize them. That can even be client-specific. You may be working with a particular client that has an assessment stage or a seven-stage interview process, whereas that’s not your standard norm.

We allow customers to build what they need from the ground up, whether it’s for contract business, permanent business, or both in one system. They can customize and build those workflows out. That includes all of their sub-statuses. What does the next stage in the process mean, from “submitted” to “CV sent”? What are your terminologies? What statuses would you use as a business? What actions follow that?

We completely customize that and allow clients to build it. It’s really difficult doing demos, actually, because a client will come to us and say they want a demo of the product, but it’s very hard to show the system. We’re so consultative in what they do and their workflows that it’s hard to show until we’ve had that conversation.

‍

Francis Larson:

What is the craziest workflow you’ve seen from a customer? 

‍

Karen Somal: 

We worked with a modeling agency, and their requirements were very unique. We configured the whole CRM and ATS around facial attributes, weight attributes, and height. Some of their assessments were very specific. Their workflow was huge. The first stage would be, “Has the candidate submitted a video showing that they can walk in a straight line?” Then it would assess their angles. They had to fit the criteria their clients were looking for, including hair color, eye color, and frame. Their workflow was fun to build, actually. It was very different.

‍

Francis Larson:

That’s so interesting about staffing and recruitment. The best agencies usually pick a very specific niche. As you were talking about, they have a workflow they’ve optimized for that industry and their space within that industry. It makes sense why the CRM and ATS would have to be customized. Otherwise, they’re going to lose some of their insight into what works in their industry.

‍

Karen Somal: 

Exactly. In some countries, there are certain things you can’t do, but depending on what they do, we utilize AI and natural language searching.

They need to capture things like a candidate’s eye color. Especially when they work in a temp model, they might have a model who needs to come in tomorrow. Or they get a call at night saying, “One of our models has fallen ill. We need somebody tomorrow.” So, it’s panic. Do a quick search, see who’s available, and send out a message. If they respond, great, and they fill that position. It’s all very varied from client to client. The ones that do something very niche can work really well.

‍

Francis Larson:

What about AI? Everyone right now is trying to build an AI-native staffing agency, or at least see how they can implement AI into their agency. Some are having no luck, and some are having amazing success. What can you talk about with AI that you see working with Athena today? Where could it go? What doesn’t work? What’s working and what doesn’t work yet?

‍

Karen Somal:

AI is my favorite subject right now. I feel like there’s a massive misconception of what AI is.

A lot of agencies come to me panicked. They say, “We’re falling behind. We’ve got no AI implemented,” or, “We need this and this.” I always say to them, “Think about your business process. What do you do today, and what are you missing that you need help with?”

Some companies are using AI quite strategically. It depends on what you need. Where AI is helping people is on the admin side. I’m a big fan of driving AI with automation, so automating a lot of what you do and then triggering AI to do certain aspects or deep learning as well. We have a whole automation program that allows you to automate tasks that then trigger AI to do certain things.

The things I find people are adapting to are changing as we grow. If I look at my children, there are certain things they’re doing where I scratch my head and think, “How did you learn that?” If you think about the next generation and the traditional ways we do things, it’s very different.

The way I see the world of AI, and even the world of CRMs and ATS systems, is that it’s going to be very conversational. WhatsApps, telephone conversations, compliance documents, SMS messages, emails, and historic migrated notes will all sit in one section.

Imagine you had something like ChatGPT sitting on top of your CRM and ATS. You do the clever stuff. For example, we have a client that sends out their client terms. We’ll use AI to verify those terms, make sure they’re signed correctly by the correct person, and upload them into the CRM. Then we use automation to drive notifications to that client’s owner to say, “These terms are about to expire. Warning.”

There’s a lot you can do with AI. I’ve seen CRMs and ATS systems doing it. We do things like recommended candidates. I’m a big believer in utilizing your database. People have built their databases over the years, so use that first. Your CRM should be your first point of contact. It shouldn’t be a filing cabinet.

When you add a job requirement, we’ve built an AI resume parser that truly understands a candidate’s profile based on their experience, how long they’ve been there, and the context behind it. For example, if I’ve been an IT project manager for six weeks and then went into catering, it understands that I’m not truly an IT project manager.

We use that against jobs currently open in the system and match based on the role. Every role requirement is different. In some cases, location is really important. We allow clients to change the weighting of what’s more important, whether that’s qualifications, skills, experience, or location. Some jobs don’t require any experience, but location is critical. You can change the weightings per role and use AI to smart match candidates, while understanding why it believes that person is a good match.

‍

Francis Larson:

It sounds like there are a lot of different areas: matching, document work, and automations. Do you think we’ll get to the point where AI agents handle the CRM as a database, and recruiters aren’t really interacting with the interface anymore? Will recruiters just be WhatsApping with an AI agent, saying, “Can you find me six people to do this thing?” How far off are we from that, or maybe it will never happen?

‍

Karen Somal:

Recruitment is all about personal touch. Whether we like to admit it or not, recruitment is about being personable with somebody. If I knew an agent had called me, I would never entertain it. Recruitment always has to have that human touch and human element.

People may want to try it, but AI sometimes misses things. We tested it on a massive RPO project with a client of ours. They said, “We want AI to screen our applications, and if candidates say no to this, reject them.” These were not low-level entry jobs. Candidates were applying for these roles and didn’t understand some of the questions. They were answering things like, “Do you have the right to work?” with “No,” because they weren’t sure. Or, “Do you have the correct visa requirements?” with “No,” because they weren’t sure.

When they reviewed a lot of the applications, they were auto-rejecting a lot of candidates who could have been suitable. I’ve always said to people, when you’re implementing these things, you have to be smarter. It needs to think like a human.

If that happens, send follow-up questions to ensure they understand. We put that layer in where AI would say, “These are the candidates who said no. Let’s send them an email or text message to say, ‘You’ve answered no. Are you sure?’” Candidates would come back saying, “I didn’t really understand the question, so I put no.”

Then what do you do at that point? You need someone to explain that to the candidate, because every question is different. I would rather somebody call me to explain that to me as well. So, in some aspects, yes, and in some aspects, no. I agree with using AI to get rid of admin, but in terms of actually doing the recruiting, I think there are missed areas.

‍

Francis Larson:

Changing gears a little bit, are you selling to direct end clients, meaning non-staffing agencies, or is it all recruitment agencies?

‍

Karen Somal:

At the moment, it’s all recruitment agencies. But I recently had inbound requests from a few talent acquisition managers, and we’ve spoken to them. We did a massive deep dive with one client, Tesco. Tesco is a massive shopping supplier. You must have had a meal deal in the UK. They were talking to me, and I was asking questions because they had just implemented a new ATS system and were really struggling with it. One of the questions I asked was, “You have these roles every single time. You’re posting them again and again.”

I think they had a UX/UI designer role, and they’d only filled a very similar role a couple of weeks earlier, but they had another opening for that requirement. They never utilized their current database pool. They have thousands of candidates sitting there.

My question to them was, “You had Joe, Francis, and Karen who were right at the end. They chose Francis because he was a little more qualified, but Karen and Joe would have been perfect for this role. At that time, they only had one role to fill. Would you ever reach out to those candidates?” They said, “How do we see that on our ATS? It doesn’t work.”

So, it has opened up a few possibilities. We could potentially do that. All we would do is take off the CRM system element, which goes against our name a bit, considering we’re called Athena CRM, but we could potentially supply talent acquisition companies as well.

‍

Francis Larson:

I’m somewhat convinced that because of automation and AI with ATSes and CRMs, a lot of external recruitment agencies were really being paid to click buttons and move people through funnels. In the near future, a lot of talent acquisition end-client folks may say, “Why am I even using a recruitment agency when I can plug in Athena or something similar, let it run on our big talent pool, and hire from there?” I wonder if you’ll start seeing more of that as time goes on.

‍

Karen Somal:

I hadn’t thought about that, but it’s something to bear in mind now. I hope not, because I don’t want my recruiters out of business.

‍

Francis Larson:

They’ll adapt. They’ll do different things.

We’re starting to see that a lot of recruitment agencies, and you probably see this too, have clients who want them to do more consultative work in addition to managing the candidate funnel. They want more collaboration and more industry expertise. They’re not accepting just being given three candidates. They want more of a consulting relationship.

What do you think is the hardest part right now with selling to recruitment agencies? What’s the biggest barrier to getting recruitment agencies to accept new technology?

‍

Karen Somal:

People reach out and want to see the product, and one of the biggest hurdles we come across is their contracts. They’re stuck in a two- or three-year contract. They’ve signed up with another supplier, but they still have 18 months left on their contract, and they don’t want to pay for two CRMs and ATSes at the same time. That’s one hurdle.

The second hurdle is that a lot of people are panicked about AI. I hear this a lot. They say, “We need to tidy up our database. We need to make sure it’s clean. We need to make sure all the field levels are there, the skill coding is done, and everything is ready.”

I say, “If you’re using true AI, that shouldn’t matter. As long as your telephone calls, the candidate’s resume, the emails, WhatsApps, and SMS messages are all synced, you don’t need to worry about manually skilling up a candidate’s profile.”

A lot of companies say, “I need to go back and make sure we’re AI-ready by tidying up our database.” With no disrespect, I don’t think they understand how AI works. That clever level of intelligence means you don’t need to do those manual things to get the correct output. People think, “It’s a really good system. I want to be onboarded. I need to go tidy up my data.” I’m saying, “No, you don’t need to do that.”

‍

Francis Larson:

Our company, Ascen, came out of Y Combinator a few years ago. It’s one of the top seed investors in the world, and lots of well-known companies come out of there. We’re really AI-pilled. We came out before AI took off, and now we’re transitioning the company to be AI-native, but that’s really hard to do and sometimes impossible to do.

One thing that’s interesting about the recruitment industry, and it sounds like you’re seeing this too, is that there are so many vendors offering some AI solutions. They were made two or three years ago when the models weren’t very good, and they don’t really work.

Recruitment agencies don’t seem to understand how powerful the models are this year and how good they are. They have this misconception, and a lot of it is vendor-led. The vendors are not good, so agencies have an overly negative view because of all the crappy vendors out there.

‍

Karen Somal:

I know exactly what you mean.

‍

Francis Larson: 

There are all these people peddling AI. I think you have to be careful if you’re a recruitment owner. Just because somebody says AI, or just because it’s a nice demo, you should actually learn specifically about it.

‍

Karen Somal: 

Being a technologist myself, I’m very specific about what tools we use. Our whole platform is built on AWS. I love AWS, and we wanted to remain in that AWS ecosystem. Our AI tools are AWS tools, like Bedrock, where we know their training models are up to date. It’s not like we’re having to maintain that either.

Every day, I wake up and think, “There’s something new.” The way technology is evolving now is incredible. A year ago, ChatGPT was amazing, and now everyone’s on the Claude bandwagon.

The vendors and technologies they’re using behind the scenes drive a lot of that. If you’re not adapting as a vendor to the latest technologies, you’re going to have massive issues and leave a bad taste in clients’ mouths, because they don’t see any benefit from what they’re getting. Then they think, “If I’m going to move, what am I moving for? It’s just going to be the same.” That’s something that needs to be addressed beforehand so people understand it.

‍

Francis Larson:

Last bit here. Roadmap-wise, what is the most exciting feature or product you think you’ll launch for your customers in the next year?

‍

Karen Somal:

It’s been pretty complex, but we have a lot of clients in the education and healthcare sectors, and licenses and compliance are a big pain.

I know most people listening are probably thinking, “Compliance is so boring. Why is she talking about that?” But being able to help businesses is important. I totally agree with what you said earlier about recruitment agencies having to be more consultative. Recruitment is not just recruitment anymore. You need to extend your services to your clients to show additional value.

There are so many recruitment agencies now. Before, you didn’t have as much choice. If I was looking for an IT developer, where could I go? If I Google that now, there would be 60 agencies within a 10-mile radius of me. So, it’s about what you offer as a service.

One of the things we want to do is more around compliance, including document checking, validation, checking expiries, and managing documentation. It has proved a bit more difficult than we thought, so it will probably take a little longer. That is probably a longer-term, more exciting feature we’ll have. The amount of admin it will take off agencies will allow their staff or recruiters to recruit better, because it removes that burden from them.

‍

Francis Larson: 

That goes in line with what you said about how agencies need to trust that you, as a vendor, are keeping up with these changes, and you clearly are. Karen, this was really great. Thank you for being on the show.

‍

Karen Somal:

Thank you for having me.

‍