Copyright Forbes
Dennis Consorte founded Consorte Marketing, a digital marketing and publicity company that helps startups and small public companies. I believe the agency of the future won’t be defined by who has the best creative director, but who has the smartest partnership between humans and machines. For many agencies, AI means faster copywriting or automated reports. But the next wave combines intelligent data analysis with agentic capabilities to create a self-guided research assistant that works hand in hand with human intuition. Best of all, you can do this with a basic ChatGPT Pro or Claude account without API access. Empower your strategists by turning generative AI into a junior team member that can explore, summarize and synthesize data from multiple sources with their guidance. The MedTech Project That Changed How I Use AI A medtech client once asked me to build a “data story” around a new product. They handed us spreadsheets, survey results and PDFs containing anonymized data. Although the client provided valuable insight during our initial interview, there was no clear understanding of what the numbers actually meant to patients, doctors and executives. Instead of assigning multiple analysts to crunch data, we used generative AI. Here's a step-by-step look at what we did: Step 1: Upload and understand. We uploaded the client’s anonymized files and asked the AI to summarize the data and identify stakeholder sentiment. I had it compare executive beliefs with user feedback from patients and clinicians. We refined our prompts and results until we had a data table that tracked trust, effectiveness, ease of use and emotional response. Step 2: Expand context using agentic browsing. Next, we gave it this prompt: “Search the web for how doctors, patients and medtech investors currently talk about [product] adoption barriers. Summarize live discussions and trending language.” We had the AI collect insights from forums, medical publications and other online resources. It discovered new patterns, including the fear of new tech replacing clinicians. That was the emotional gap. The numbers showed friction, and the web search revealed why. We then had the AI scan peer-reviewed articles for supplemental data and enhanced credibility. Step 3: Add a human touch. We curated results based on project requirements and emotional impact. We examined the product through the lens of each stakeholder and used empathy to translate data into a human story. With all of this information in mind, our team built a communications plan to position the product as empowering clinicians to provide patients with a better experience. By combining machine efficiency with human understanding, we improved adoption rates. How Your Agency Can Replicate This Workflow Replicating this workflow doesn't require a data science team or API access—just a pro license to your favorite generative AI with agentic capabilities and a clear process. Let's break it down: • Define the brief in plain language. I'd recommend beginning your project with the following prompt to train your AI to act like a partner rather than a text generator: “You are a strategy assistant. Your goal is to turn unstructured data into insights about human motivations. I’ll provide files and feedback as we go.” • Upload and extract. Upload files for raw client data, such as spreadsheets, PDFs and transcripts. Ask for a “stakeholder empathy map” or “theme clustering.” I recommend reviewing your AI platform's data retention policies and any enhanced security options available. Then, remove any personally identifiable information and make sure all client data is anonymized before uploading. • Activate browsing for live context. Give the AI agent autonomy to explore: “Find the latest articles, reviews or discussions that reflect how [target audience] feels about [product or trend]. Summarize emotional themes, not just stats.” This transforms AI into a live cultural scanner that you might currently do manually or with expensive analysis software. • Validate with human judgment. Don’t delegate empathy. Review findings through your human lens. Which insights feel real, which feel shallow and which deserve deeper exploration? Remember, generative AI can hallucinate or provide bad data. So, cross-check high-stakes statistics or quotes against the original source documents. • Build a repeatable framework. Save the best prompts as templates. Some of the ones we've saved and repeat include a client discovery prompt for initial audits, voice of the customer for feedback synthesis and market pulse for competitor and sentiment scanning. Why This Matters For Agency Leaders The old agency model rewarded creative genius. The new one rewards systems that scale with technology and human oversight. AI can handle most of the discovery, but results improve when humans inject emotional intelligence, ethical reasoning and creative translation. Leaders should now ask: How do we make human feedback part of every AI cycle? How do we use browsing agents to stay culturally current? How do we turn those insights into actionable campaigns that clients find valuable? Next Steps If you run an agency, start small. Give your generative AI one project this week. Ask it to analyze a dataset, audit content or review the market. Let it browse, summarize and visualize. Then, step in. Question it, and humanize the results. Follow this plan to build your human-in-the-loop agency today.
