
AI and Sales Assessments: What You Need to Know

We’re hearing more questions lately about how artificial intelligence might affect sales candidate assessments and evaluations. Two of the most common are:
- Does Grow Sales and OMG use AI in its assessments?
- Can candidates use AI to cheat an assessment?
The short answers: No and Not easily.
Here’s why...
Why We Don’t Use AI in Our Assessments
AI models, especially generative ones, are often described as black boxes. You input a question, it produces an answer, but the reasoning behind that answer is hidden.
If you type “2 * 2” into Excel, you know the answer will always be “4” and you understand the logic. With AI, you receive a response without visibility into how it was generated or what influenced the outcome.
In hiring, that lack of transparency creates real problems:
- Hiring managers need clarity. They need to understand why a candidate is recommended or rejected so they can make informed decisions.
- Some jurisdictions require explainability. Employers may be legally obligated to explain to candidates why they were rejected if asked.
- AI can amplify bias. Because it is trained on imperfect real-world data, bias in that data can become embedded in the model’s recommendations. This can create significant risk around fairness, adverse impact, and compliance.
Our assessments are intentionally transparent and explainable. Every employer receives a clear breakdown of which criteria a candidate met or did not meet. If needed, hiring managers can point to the specific question and the candidate’s actual response.
We also conduct annual adverse impact audits to ensure our recommendations remain fair and consistent across demographic groups.
When it comes to hiring, novelty for novelty’s sake is not the goal. Accuracy, fairness, and explainability are what matter most, which is why AI is not part of our scoring or recommendation process.
Why AI Cheating Doesn’t Work with Our Assessments
Could a candidate attempt to use AI to answer our assessment questions? Technically yes, but the effort required is significant, and the results would likely be ineffective.
AI is too slow for the format. Our assessment contains approximately 250 multiple-choice questions, with an average completion time of 10 seconds per question. When we tested ChatGPT on a reworded question, it took 18 seconds to respond. Multiply that by 250 and you get a 75-minute test - an extreme outlier that our system flags automatically.
AI doesn’t know our scoring methodology. Our scoring process is highly specialised and not publicly available. It uses multiple scoring approaches, including true/false, conditional scoring, and partial credit. This means an AI model might choose an answer that seems “good” but does not achieve the highest possible score. Additionally, each client customises their assessments to reflect their own hiring criteria. A “good” answer for one company could be the wrong answer for another.
Our content changes regularly. We constantly monitor and update our questions to align with actual sales performance data. New questions are added, others are removed, and scoring is refined as needed. Even if an AI somehow learned our “right” answers, several would be outdated within weeks.
The Bigger Picture
Could a highly determined candidate still attempt to game the system with AI? Possibly. But the time, effort, and uncertainty involved make it a poor use of energy. In most cases, they would be far better off preparing for the role - researching the company, practicing their pitch, and showing up ready to sell.
AI can be an incredible tool for many business purposes, and OMG uses it internally for certain operational tasks such as drafting alternative versions of assessment questions or speeding up non-critical content work. However, when it comes to sales candidate assessments and evaluations, our priority is ensuring the results are accurate, fair, defensible, and predictive. That commitment is why AI does not, and will not, play a role in how we score or recommend candidates.



