The rise of deep research: How agencies are using AI for strategy, content and client pitches

As client demands grow more complex and timelines shrink, agencies are turning to new AI-powered research tools to streamline workflows and supercharge insight development.
In recent months, tech giants and startups alike have released new “deep research” tools that are already helping teams within holding companies and indie shops to move faster, think deeper and deliver better work. Agencies are integrating the capabilities across strategy, media, creative, and comms teams — with some seeing benefits in helping craft pitches, shape creative briefs, and even simulate audience personas.
Unlike basic search or dashboards, deep research tools from Perplexity and others use AI-enabled reasoning autonomously run multi-step workflows — scanning the open web, proprietary datasets, and niche databases — to generate rich, synthesized insights in minutes. As OpenAI put it, deep research equips ChatGPT to perform “at the level of a research analyst.”
The goal is to compress hours of manual research into automated briefings that can power everything from brand audits to creative platforms. This could have major implications for agencies as strategy teams rethink workflows, staffing, and even pricing models.
One adopter is Havas Group, which has integrated deep research into its proprietary data sources within its Converge platform. It’s “like a dream from 10 years ago sort of finally coming true,” said Dan Hagen, global chief data and technology officer at Havas, who recalled imagining a decade ago that he’d someday have an AI “planning buddy” to help him quickly learn new topics.
“We’re finding that the marketing and new business space is very much the sweet spot,” Hagen said. “You’ve got four weeks to turn something around and you want to be able to understand a brand, a category, maybe even some individual people… And the way that [these tools] are structured means it’s super accessible.”
One way Havas is experimenting with the tech is by turning deep research into interactive tools, such as custom GPTs that can simulate consumer behavior as “digital twins” of various audience segments or client types. These tools help by finding, synthesizing and uploading info from a range of external and internal sources such as internal data sets, cultural reports and client-specific data while also preserving data protection.
The custom audiences allow teams to explore different facets of consumer identity in a dynamic, conversational format with a wider and deeper data set: “Now you’ve got effectively a hardcore runner you can have a conversation with which has been really interesting to sort of explore some stuff,” Hagen said.
IPG-owned Golin is another agency using deep research to help with both client work and new business development. After initially being used for efficient background research, its subsidiary creative agency, The Brooklyn Brothers, has now integrated it into strategic development, creative brief writing, brand audits, and audience profiling.
“It has real utility,” said Paul Parton, Group Chief Strategy Officer of Golin and founding partner of The Brooklyn Brothers. “… It’s shifting a conversation from efficiency definitely to efficacy and it’s actively making work better.”
Golin is taking an iterative approach to improve the quality of insights with each round of strategy development — starting with broad prompts and refining them into detailed outputs such as company profiles, brand audits, and competitive analyses. Creative and strategy teams also toggle between AI tools to maximize results, often using Anthropic’s Claude generative AI platform for initial synthesis before turning to Waldo for deeper, sourced research that feeds back into the process with marketing-focused workflows like the “Four Cs.”
Austin-based agency Mighty & True is also integrating deep research into its proprietary tools like Flow, which help account managers and customers create briefs, said Mighty & True Founder and CEO Kevin Kerner. His agency also is developing agent-powered dashboards that analyze media campaign data from sources to compare performance against benchmarks, and surface strategic recommendations. It’s also using startups like Relay — an AI agent builder — to create smaller custom agents that help Mighty & True with its own marketing by automating content creation.
Kerne thinks B2B companies might also benefit from using deep research for more efficient, comprehensive, and potentially more unbiased information when conducting research and making purchasing decisions.
An overview of deep research
In the past few months, major AI players have introduced their own versions of deep research tools to provide individual users and enterprise customers with more detailed insights that would traditionally require significant time and effort. Following Google’s deep research debut in December, both Perplexity and OpenAI debuted new offerings in February with examples of how marketers might use the tools.
Deep research tools stand apart from standard generative AI search by operating at a radically larger scale — analyzing thousands to hundreds of thousands of documents, compared to the few hundred typically processed by general-purpose platforms like ChatGPT or Gemini. This volume-driven approach enables a deeper and more comprehensive understanding of complex topics that results in a shift from surface-level synthesis to true analytical depth.
Where Waldo found its niche
Of all the emerging deep research agents, Waldo is gaining traction within marketing agencies and consultancies, thanks to dozens of industry-specific workflows and what customers say are user-friendly features and helpful data citations.
Both Havas and Golin are already clients along with others like WPP, Accenture, BCG and smaller independent agencies that use Waldo to punch above their weight. Waldo also has a brand-based pricing model that mirrors how agencies actually operate and allows teams to collaborate across functions without worrying about seat limits.
Instead of simply scanning the open web, the startup integrates with social platforms, ad libraries, and other industry-specific databases. Its proprietary research engine, automation tools, and other features also help users extract trends, analyze audiences, and surface insights at scale.
According to Waldo Founder and CEO Justin Wohlstadter, the goal is to help tools no longer be siloed within specific teams. New features include Waldo’s “Brand Audit” and “Brief Agent” tools that further push the envelope, offering real-time brand analysis and dynamic, guided research experiences.
“Waldo has kind of become the start screen for strategy,” Wohlstadter said. “The one place where all these strategists can go and they can get the social, they can get the audience stuff, they can get whatever it is, all in one simple place.”
More in Media

Digiday+ Research’s 2025 ad snapshot: Are publishers becoming less dependent on ad revenue?
Publishers are working to diversify their revenue, and fewer say the vast majority of their revenue is coming from advertising this year.

Publishers watch warily as tariffs loom over ad budgets and print costs
Publishers are keeping a watchful eye on the outcome of the sustained macroeconomic shock that will arise thanks to Trump’s tariffs.

Media Briefing: The pros and cons of different AI revenue models for publishers
Here’s a list of some different AI revenue models that publishers are signing, and the pros and cons of each.