OpenAI just rolled out advanced usage analytics and spend controls for ChatGPT Enterprise. For consulting firms, this isn’t another dashboard to ignore. It’s the first real visibility into which partners and senior staff are actually using AI tools and which ones are still burning 30 hours on every major proposal.

The analytics show seat-level activity, token consumption by team, and cost allocation across projects. More importantly, they expose the gap between the firm that says it’s “exploring AI” and the one that’s cutting proposal turnaround from three weeks to four days while competitors are still formatting decks by hand.

I’m Sam McKay. I founded Enterprise DNA and spent years helping firms turn messy data into operational advantage. The new ChatGPT Enterprise analytics matter because they make the cost of inaction visible. If your senior people aren’t using AI for research, synthesis, and proposal work, you’re paying full freight for tasks that competitors are automating. The analytics tell you exactly where that’s happening.

The Real Cost of Manual Work in Consulting

Most consulting firms know their win rate and average engagement size. Few track the internal cost-of-sale. A major proposal takes 20 to 40 hours of senior time. Research at the start of an engagement runs another 15 to 25 hours. Across a year, that’s $80K to $300K in partner and director time spent on work that could be templated, automated, or pulled from past projects.

The problem isn’t effort. It’s repetition. Every proposal starts from a blank slide deck. Every new client kicks off with the same secondary research. Every insight the firm produces lives in a folder that nobody searches. The firm pays for the same work twice, then pays again when a competitor delivers faster.

ChatGPT Enterprise analytics expose this waste at the individual level. You can see which teams are using AI to draft proposals, which partners are still doing it manually, and where token spend correlates with faster delivery. For firms that have rolled out enterprise seats, the data often shows a 70-30 split: a third of the team uses the tool daily, two-thirds touch it once a month or ignore it entirely.

That split is the gap your competitors are exploiting. The firms winning work right now aren’t smarter. They’re faster because they’ve automated the repetitive parts and freed senior people to focus on strategy and client relationships.

What the New Analytics Actually Show

OpenAI’s update gives enterprise admins three things they didn’t have before: seat-level usage tracking, cost allocation by team or project, and token consumption trends over time. You can see who’s using the tool, what they’re using it for, and how much it’s costing per engagement.

For consulting firms, the most useful view is usage by role. If your partners and directors aren’t logging regular activity, they’re still doing proposal work and research the old way. If your analysts are heavy users but senior staff aren’t, you’ve got a training problem or a culture problem.

The cost allocation feature lets you tag usage to specific clients or internal projects. That means you can track whether AI spend on a given engagement correlates with faster delivery or higher margins. Early data from firms using this feature shows that engagements with high AI usage typically close proposals 40% faster and report lower internal labor costs.

The token trend view is less immediately useful but matters for planning. If usage is flat month-over-month, your rollout stalled. If it’s spiking in one team and flat in others, you know where to focus training dollars and where to find early wins to showcase internally.

Where Consulting Firms Leak Time and Money

The analytics reveal three patterns that show up across most consulting firms. First, proposal generation is still manual. Senior people write decks from scratch, pull case studies by memory, and format slides one at a time. A $200-per-hour partner spends 25 hours on a proposal that recycles 60% of past content. That’s $5,000 in labor for work that an AI agent could draft in 90 minutes.

Second, research and synthesis happen in silos. Each engagement starts with the same secondary research: industry reports, competitor analysis, regulatory landscape. The firm has done this work before, but it’s buried in old decks and email threads. The new team starts over. That’s another 20 hours per engagement, repeated across every client in the same sector.

Third, knowledge management is a graveyard. Every project produces insights, frameworks, and proprietary analysis. Almost none of it is reusable because nobody can find it. The firm’s competitive advantage is locked in PDFs that don’t surface in search. When a partner needs a past case study or a pricing model, they either skip it or recreate it.

ChatGPT Enterprise analytics make these leaks visible. You can see which teams are drafting proposals with AI, which ones are using it for research, and which ones aren’t using it at all. The firms that act on this data are pulling ahead. The ones that ignore it are paying full cost for work that’s already been automated elsewhere.

What an AI Agent Does for Proposal Work

A Proposal Generation Agent doesn’t write proposals from scratch. It pulls past proposals, case studies, pricing models, and client context into a tailored first draft. The partner reviews, edits, and adds strategic nuance. The 25-hour process drops to four hours of senior time plus 90 minutes of agent work.

Here’s what that looks like in practice. The partner opens a new opportunity in the CRM. The agent reads the client brief, pulls relevant past proposals from the firm’s document library, identifies case studies that match the client’s industry and challenge, and generates a draft deck with an executive summary, approach, team bios, and pricing options.

The draft isn’t final. It’s 70% complete and needs strategic input. But the partner isn’t starting from zero. They’re editing, refining, and adding the insights that only a human can provide. The repetitive work is done. The high-value work gets full attention.

This is one of the named agents we build in Omni Ops. It connects to the firm’s existing document storage, CRM, and knowledge base. It doesn’t require a new system or a migration project. It works with what you already have and makes it searchable and reusable.

Firms that deploy this agent report proposal turnaround dropping from three weeks to one week. Win rates stay flat or improve slightly, but cost-of-sale drops by 50% or more. That’s the difference between paying $5,000 in partner time per proposal and paying $1,200.

Research Agents and the Cost of Starting Over

A Research Agent runs structured industry and company research at the start of every engagement. It pulls public filings, news, competitor analysis, and regulatory updates, then produces a one-page brief with sources and key takeaways. The analyst reviews it, adds proprietary context, and hands it to the partner. The 20-hour research phase drops to three hours of human time.

The agent doesn’t replace judgment. It replaces the repetitive work of finding, reading, and summarizing sources. The analyst still decides what matters and what doesn’t. But they’re not spending two days on Google and in PDF readers. They’re reviewing a structured brief and adding the insights that require domain expertise.

This is another Omni Ops agent. It connects to the firm’s research subscriptions, internal document library, and public data sources. It runs on a schedule or on-demand when a new engagement kicks off. The output is a structured brief that the team can edit, annotate, and share with the client.

Firms using this agent report research time dropping by 60% to 75%. More importantly, they report higher consistency across engagements. Every project starts with the same structured research process. Nothing gets missed because someone forgot to check a source or didn’t know the firm had already analyzed this competitor.

Knowledge Agents and the IP You’ve Already Paid For

A Knowledge Agent reads every deck, document, meeting transcript, and email the firm produces. It indexes the content, understands the context, and answers questions across the entire corpus. A partner preparing for a pitch can ask, “What pricing models have we used for financial services clients?” and get a list of past proposals with relevant excerpts.

This isn’t search. It’s synthesis. The agent doesn’t just find documents. It reads them, understands the relationships between them, and surfaces the specific insight you need. It turns the firm’s knowledge base from a graveyard into a competitive advantage.

This is the third Omni Ops agent we build most often for consulting firms. It connects to the firm’s document storage, CRM, and communication tools. It runs in the background, indexing new content as it’s created. The team interacts with it through a simple interface: ask a question, get an answer with sources.

Firms using this agent report that partners and directors spend 30% less time hunting for past work. More importantly, they report that junior staff can access senior expertise without interrupting senior people. The knowledge that used to live in one partner’s head is now available to the entire firm.

If you want a structured way to identify which of these agents would deliver the fastest ROI in your firm, we built a worksheet that walks through the decision framework. Download the Deploy Your First Business Agent guide and use it to map your highest-cost manual processes to the agent types that eliminate them.

What the Analytics Tell You About Training ROI

ChatGPT Enterprise analytics show which training investments are working and which ones are theater. If you ran a firm-wide AI workshop six months ago and usage is still concentrated in two teams, the training didn’t stick. If token consumption is spiking in one practice area and flat in others, you know where the early adopters are and where to focus follow-up.

The most useful metric is usage by role. If your partners and directors aren’t logging regular activity, they either don’t see the value or don’t know how to use the tool for their actual work. Generic training on “how to write a good prompt” doesn’t move the needle for senior people. They need to see how AI fits into their specific workflow: proposal generation, client research, or knowledge retrieval.

Firms that get this right run role-specific training. Partners learn how to use AI for proposal drafts and client research. Analysts learn how to use it for data synthesis and report generation. The training is two hours, not two days, and it focuses on one workflow at a time.

The analytics let you measure the impact. If usage spikes after training and stays elevated, the training worked. If it spikes and drops back to baseline, the workflow integration didn’t stick. You can iterate quickly and focus on the use cases that deliver measurable time savings.

Why Competitors Are Pulling Ahead Right Now

The firms winning work in 2026 aren’t better strategists. They’re faster operators. They deliver proposals in one week instead of three. They start engagements with research already done. They reuse past insights instead of recreating them. That speed compounds into a competitive advantage that’s hard to reverse.

ChatGPT Enterprise analytics make this gap visible. You can see which of your teams are operating at the new pace and which ones are still doing it the old way. You can see where competitors are likely investing and where you’re falling behind.

The cost of inaction isn’t hypothetical. It’s $80K to $300K per year in wasted senior time, plus the opportunity cost of slower delivery and lower win rates. The firms that act on this data in the next six months will pull ahead. The ones that wait will spend the next two years catching up.

What an Omni Audit Looks Like for Consulting Firms

The Omni Audit is 60 minutes. We review your current workflow for proposals, research, and knowledge management. We look at your ChatGPT Enterprise usage data if you have it, or we estimate the cost of manual work if you don’t. We identify the three processes where AI agents would deliver the fastest ROI.

You leave with three outputs. First, a one-page process map that shows where time is leaking and where an agent would intervene. Second, a cost-benefit model that quantifies the savings in partner time and faster delivery. Third, a 90-day deployment plan that prioritizes the highest-impact agent and outlines the integration work required.

This isn’t a sales call. It’s a working session. We don’t pitch a platform or a multi-year roadmap. We show you exactly what an agent would do for your specific workflow, how much it would cost to build, and what the payback period looks like. Most consulting firms see payback in 60 to 90 days for the first agent.

See the full Omni Audit for consulting firms and read what other partners and managing directors took away from the session. The audit is free. The only cost is 60 minutes of your time and the willingness to look honestly at where your firm is leaking money on repetitive work.

The Next 90 Days Matter More Than the Last Two Years

The consulting firms that deploy AI agents in the next 90 days will have a 12-month head start on competitors. They’ll deliver proposals faster, start engagements with better research, and reuse past insights instead of recreating them. That speed advantage compounds into higher win rates, better margins, and the ability to take on more work without hiring.

ChatGPT Enterprise analytics give you the data to make this decision with confidence. You can see where your team is already using AI, where they’re not, and where the biggest time savings are hiding. You can measure the impact of training, track ROI by engagement, and prove the business case for expanding usage.

The firms that act on this data will pull ahead. The ones that wait will spend the next two years catching up. The cost of inaction is $80K to $300K per year in wasted senior time, plus the opportunity cost of slower delivery and lost work to faster competitors.

If you’re building with Claude or Codex right now, grab the free Working With Claude field guide. Thirty-two pages on the full ecosystem, Claude Code in depth, and how to roll agents out properly. Get the free guide.

The window is open. The data is available. The tools are ready. The only question is whether you’ll use them before your competitors do.