Agencies sell expertise but lose margin on admin. Account managers rebuild the same proposal from an old deck, analysts spend the first days of every month copying numbers from Google Ads, Meta and GA4 into slides, and project managers chase status across Asana, Slack and email. The tool bill grows each year, and nobody is quite sure which seats are still in use.
InTheory acts as your part-time technology lead. We would pick one recurring workflow that eats billable time, such as monthly client reporting or proposal assembly, measure it, and fix it with the stack you already pay for where possible. Your strategists still own the thinking, and nothing goes to a client without a person reviewing it.
Every retainer client expects a monthly report, and each one is built by hand: export from ad platforms, paste into a template, write commentary, fix formatting. Senior people review late at night. The work is rarely billed in full, it delays strategic conversations, and a single copy-paste error in front of a client costs more trust than the report earns.
Retainers are sold on hours or deliverables, but time tracking in Harvest or the project tool is patchy. Extra revisions and small requests pile up unrecorded. By the time an account is reviewed, it has been unprofitable for months. Owners know some clients lose money but cannot say which ones or by how much without a spreadsheet exercise.
Agencies accumulate software: two project tools, three reporting connectors, design apps, SEO platforms, AI writing tools on personal cards, and seats for people who left. Each was bought for a good reason, but together they cost more than anyone tracks and create duplicate work as data moves by hand from one to another.
Clients ask whether AI makes content cheaper, and some expect lower fees. Teams are already using ChatGPT, Claude and Jasper in different ways, with no shared standards on quality, disclosure or client confidentiality. Agencies need a clear position: where AI speeds up production, where human judgment is the product, and how to price the difference.
We fix one at a time, starting with the one that costs you most. Each page shows the steps, the tools, what stays human and what we would measure.
Turn discovery calls into reviewed proposals faster, with notes, scope modules and pricing assembled from approved building blocks and an account lead signing off on scope and fees.
See the workflow →Give new clients an organized first 30 days with intake forms, access requests for ad accounts and analytics, kickoff scheduling and a project template that starts automatically on signature.
See the workflow →Pull ad, analytics and CRM data into consistent monthly reports automatically, with drafted commentary that an account manager checks and edits before anything reaches the client inbox.
See the workflow →Plan, brief, draft, review and publish client content with a clear workflow, approved AI uses and a quality bar that keeps senior editors focused on judgment.
See the workflow →Send clients and internal teams regular status updates drawn from Asana, ClickUp or Monday.com, so account managers stop writing the same summary emails every Friday.
See the workflow →List every tool and seat the agency pays for, see what overlaps and who uses what, and decide what to keep, connect or cancel before the next renewal.
See the workflow →Spot at-risk retainers early with signals like falling engagement, missed approvals and over-servicing, and plan renewal conversations well before the contract end date or notice period.
See the workflow →See utilization, effective hourly rate, retainer profitability, pipeline and revenue concentration in one weekly view for partners and leads, instead of an end-of-quarter spreadsheet exercise that arrives too late to act on.
See the workflow →How we deliver this: AI automation for small business or fractional CTO leadership, scoped in writing and run over 90 days. See pricing or the sample scope.
We start with what you already pay for. New tools have to earn their place.
Agencies hold client ad accounts, analytics, customer lists and confidential plans, usually under NDAs and contracts with confidentiality and IP clauses. Before using AI tools on client material, check those contracts and the vendor's data terms, including retention and model training. Customer data from clients may fall under privacy laws such as the Colorado Privacy Act. Disclosure of AI-assisted or sponsored content can raise FTC questions. Counsel should advise on specifics.
General information, not legal advice.
Defaults are illustrative for marketing agencies (new retainer clients per month). Change them to your numbers.
a month in staff time, before any lost sales
Illustration only, not a forecast or guarantee. Recovered staff time becomes a cash saving only if spending actually declines or the time goes into paid work. Tool costs are an assumption; we replace every number here with yours before recommending a scope.
Some production tasks get faster, and clients will notice. The answer is usually to price on outcomes and expertise rather than hours, and to be clear about where AI speeds up drafting and where strategy and judgment remain the product. We would help you decide which workflows to speed up first and what that means for margins and pricing.
It depends on the tool and your contracts. Business plans for tools like ChatGPT, Claude or Microsoft Copilot typically offer stronger data terms than free versions, but you should check retention and training settings. Client NDAs may also restrict use. We would build an approved tool list and simple rules so staff know what is allowed.
Often nothing new. Many agencies have reporting tools set up for a few clients and manual decks for the rest, or dashboards nobody reads. We would look at how reporting actually happens, fix templates and connectors, and add AI-drafted commentary only where account managers find it saves real time.
The workflow that costs the most unbilled senior time and can be measured. For many agencies that is monthly reporting or proposals; for others it is scope tracking or the tool bill. In the first 30 days we would map it with your team, baseline the hours and agree one pilot with clear acceptance criteria.
Plans are $2,500 per month for advisory support, $5,000 per month for Lead plus delivery covering one scoped workstream, and $7,000 per month for embedded support with team sessions. Engagements start with 90 days and a written scope. Tools and any large custom build are budgeted separately, and a software audit sometimes covers part of the fee in cancelled subscriptions.
In 30 minutes we look at one recurring workflow, what it costs you today and whether there is a credible case to improve it. If there is not, we will say so.
30 minutes with Phil or Martin. No technical brief, no pitch deck. Phil or Martin replies personally, usually within one business day.