One sprint or all four, the commitment frame.

The single-sprint retainer runs between seven and nine thousand a month per function, depending on the function and the scope. Sales sits at the higher end of that range because the touch volume and the CRM integration work absorb more operator time. Support sits at the lower end because the deflection workflow stabilizes faster against a defined KB. Content and Ops both land in the middle. Stack four of those separately and the run-rate lands around thirty-one and a half thousand a month, before any bundling discount.

The four-department bundle prices at twenty-five thousand a month, which is roughly a twenty percent discount against the separate retainers. The discount is not a marketing line item. It reflects real operating efficiency on our side. One operator can run two or three functions for a single client more efficiently than four operators running one function each across four clients, because the data access, the brand voice profile, the slack channel, and the weekly review cadence get shared across the functions. The savings get passed through.

The structural question is not whether the twenty percent discount is real. The discount is real. The question is whether your team is in a position to use four fractional departments effectively from day one, or whether spreading attention across four sprints during the first month produces worse outcomes per function than running one sprint at a time. We wrote about the underlying mechanics in What is a Fractional AI Department. This page is the commitment frame.

Run the math at the run-rate level. Four functions at the separate retainer floor of around thirty-one and a half thousand a month is three hundred and seventy-eight thousand a year. The bundle at twenty-five thousand a month is three hundred thousand a year. The annual delta is seventy-eight thousand dollars in retainer savings, before any annual prepay or twelve-month commit discount on top of that. For a funded team under fifty, seventy-eight thousand a year is a senior engineer salary, a full-time growth hire, or three quarters of runway on the next product bet. The savings are not nominal.

The savings are also not free. The bundle assumes you have all four functions running in production by month three at the latest, which means four sprints have kicked off, four operators are tuning four functions against your data, four senior team members are aligned on the function under their part of the org, and four sets of inputs are clean enough that the sprints can hit cadence on schedule. That is a real operating ask. Some teams have all four functions ready to run and the senior bandwidth to absorb four kickoffs in a single quarter. Most teams do not, and the bundle pricing only pays off when the operating capacity is real.

The right frame is to ask whether you would be willing to run four sprints in parallel in week one of the engagement. If yes, the bundle is the right starting commitment because the savings are real and the operating capacity is in place. If no, the phase-in approach lands the same bundle pricing over the course of ninety days without the operating strain of four parallel kickoffs. The destination is the same. The question is how fast you reach it.

For most funded teams under fifty, the single sprint is the right first commitment. The reason is structural. The fractional model is new to your team, your operating cadence, and your senior reviewers. Running one sprint produces a clean proof window inside the first month against one function, with one operator, against one set of inputs. The senior team gets to see the model run end to end on a single function before signing off on the spend for the other three. The math closes the conversation about whether the model fits your specific shape of company, instead of remaining theoretical across four parallel commitments.

The single sprint is also the lowest-risk commitment because the off-ramp is a thirty-day notice on a single retainer. If the function does not produce the expected output in week four or week six, the engagement is a six-figure rounding error against the full year-one plan. Compare that to the bundle commitment, where a misfit on the model produces a four-function unwind that is operationally expensive even when the contractual notice is the same thirty days. The reversibility math favors the single sprint for the first engagement.

The single sprint is the right starting commitment when any of the following is true. The team is new to fractional AI as a concept. The senior team is split on which function should be the priority. The board needs to see a proof artifact before signing off on a larger commitment. The cash flow model has constrained the next ninety days. The team has internal political friction on the spend decision. In each of these shapes the single sprint produces the proof and the political cover that the bundle commitment cannot produce in week one.

The bundle is the right starting commitment in a narrow but real set of cases. The team has already worked with a fractional layer before, or the founder has run multiple functions through agency or outsourced models and understands the operating shape. All four functions are bleeding badly enough that the sequencing question is not which function first but how fast can all four come online. The senior team has the bandwidth to absorb four kickoffs in a single quarter. The cash flow model can absorb the bundle retainer from month one, and the seventy-eight thousand a year in savings is meaningful to the year-one plan.

In these shapes the bundle is the right move. The twenty percent savings compound from day one. The four operators get to work in parallel, which means the AI Sales Department, the AI Content Department, the AI Ops Department, and the AI Support Department all hit full cadence by week four or week five instead of staggering across months. The compounding effects show up faster because the functions reinforce each other. Content traffic feeds Sales conversion. Ops dashboards give Sales the data to refine ICP. Support feedback feeds Content to address objections.

The honest filter is whether the founder has prior conviction on the model or is testing it for the first time. First-time conviction usually defaults to the single sprint regardless of how clear the four-function thesis seems on the spreadsheet. Prior conviction earned through previous fractional or outsourced engagements is the structural fit for the bundle commitment in week one. There is no shame in either path. The bundle is just the right move for fewer teams than the spreadsheet math suggests in isolation.

The phase-in path is the default recommendation for teams that are committed to running all four functions but want to earn the proof on each one before stacking the commitment. The sequence is straightforward. Sprint one in week one, scoped to the function that is bleeding worst. Sprint two in week eight when sprint one is in steady-state cadence and the senior team has bandwidth for a second kickoff. Sprint three in week twelve when sprint two is stabilizing. Sprint four in week sixteen when the third function is producing cleanly.

By month four all four functions are running, the bundle pricing kicks in retroactively against the next twelve months, and the team has earned the proof on each function instead of committing on the spreadsheet. The total annual cost on the phase-in path is slightly higher than the cold-bundle path because the first three months do not get the discount, but the difference is roughly fifteen to twenty thousand dollars across the year, against the cost of being wrong on the upfront commitment by tens of thousands. The phase-in path is the lower-variance choice for first engagements.

The phase-in path also has a natural sequencing logic that the bundle commitment does not preserve. Sprint one is the function that is bleeding worst right now. Sprint two is the function that pairs best with sprint one once it is producing. Sales pairs naturally with Content because the content surface feeds the outbound personalization. Content pairs naturally with Sales because the warm replies create demand on the conversion funnel. Ops pairs with whatever function is producing the most data that needs to roll into dashboards. Support pairs with whatever function is creating the most customer-facing surface area. The sequencing emerges from the proof of the prior sprint, not from the spreadsheet plan.

The trap teams fall into when they look at the bundle pricing is treating the seventy-eight thousand in annual savings as a pure win, without accounting for the operating cost of running four parallel kickoffs in the first month. The hidden cost of the bundle is senior team attention. Each function kickoff absorbs three to six hours a week of senior time across the first two weeks. Stack four kickoffs in the same window and the senior team is spending fifteen to twenty-five hours a week on engagement work during a quarter where they were planning to ship product, close a round, or run a campaign. That cost is not on the invoice. It shows up as the senior team running ten percent slower on everything else during the first quarter.

The other trap is the assumption that bundle pricing is the only way to land the twenty percent savings. The phase-in path lands the bundle pricing by month four as soon as the fourth function comes online, with the prior three months running at the single-sprint rate. The annual cost difference between the phase-in path and the cold-bundle path is small relative to the total spend, and the operating variance is much lower. For most teams that math is the right trade.

The teams who do choose the bundle from week one and run it well are the teams who have prior conviction, real operating bandwidth, and a senior layer that can absorb four conversations in parallel without losing focus on the underlying business. Those teams exist. They are not the majority. The honest default is the phase-in path, and the bundle commitment is the right choice for the smaller subset of teams where the prior conditions are clearly in place.