BLS Purchasing Projections: Limits for AI Staffing Decisions
“Keep the national projection beside the local workflow record. Name the decision owner, test what changed, and explain what the evidence cannot establish.”
What does BLS project for purchasing jobs in each industry?
BLS projects 34,400 more jobs for purchasing managers, buyers and purchasing agents in 2035 than in 2025, and for buyers and purchasing agents its sector projections run from −6.9% in government (excl. state and local education and hospitals) to +18.3% in utilities around the national +6.0%, which is why the check under the chart starts from your own industry’s line.
Zinit Research · BLS projections, 2025 to 2035
BLS projects jobs for buyers and purchasing agents to change +6.0% from 2025 to 2035, and from −6.9% to +18.3% across industry sectors
Projected change in employment by industry sector, from the BLS National Employment Matrix, for sectors with at least 1,000 jobs in the occupation in 2025. Self-employed workers are counted apart from the sectors, and BLS projects them to change −22.3% for buyers and purchasing agents and +10.0% for purchasing managers. Choose an occupation and your industry to redraw the chart and the check.
Set your staffing plan beside the BLS projection
- Your plan, a year
- −4.0%
- BLS projection, a year
- +0.6%
- Your team at the BLS pace, 2030
- 20.5
- Plan against that pace
- −3.5 people
Your plan takes a team of 20 to 17 by 2030. BLS projects buyers and purchasing agents across all industries to change +6.0% from 2025 to 2035, and at that pace a team of 20 would be 20.5 in 2030, so the plan sits 3.5 people below the national projection. That gap is a question for your own records, because the projection counts jobs across every U.S. employer under a full-employment assumption for 2035 and cannot isolate the effect of one technology; it holds nothing about your transaction volume, category mix, touch hours or exceptions.
How BLS reached +6.0%: all jobs in the economy change +3.5% (industry output and productivity), and the occupation’s share of them +2.4% (staffing patterns), the two channels through which the BLS FAQ says new technology enters the projections. Table 1.12 names the factor BLS applied to buyers and purchasing agents across all industries: “Demand change - share increases as supply chains become more complex and additional service contracts are needed”.
Formula: team at the BLS pace = team today × (1 + BLS ten-year change) ^ (years to the plan year ÷ 10), and your plan a year = (planned team ÷ team today) ^ (1 ÷ years) − 1. The pace applies BLS’s average for 2025 to 2035 from 2026; it is a national comparison line for the industry, never a staffing target.
Source: U.S. Bureau of Labor Statistics, Employment Projections, National Employment Matrix and occupation tables, 2025–35 projections released August 27, 2026, files as captured by the Internet Archive on September 26 and October 2, 2026. BLS suppresses matrix cells that are confidential, of poor quality or under 50 jobs, and the matrix has no real estate and rental and leasing row for buyers and purchasing agents. Chart: Zinit Research. Data and method
What National Purchasing Occupational Projections Actually Measure
The BLS outlook for purchasing managers, buyers and purchasing agents projects employment in the combined group to grow by 6 percent, from 604,400 jobs in 2025 to 638,800 in 2035: a net increase of 34,400 jobs. Purchasing managers account for 88,300 base-year jobs and 91,800 projected jobs, a rounded 4 percent increase of 3,500; buyers and purchasing agents account for 516,100 and 547,000, a rounded 6 percent increase of 30,800. The same outlook projects about 53,600 openings each year on average over the decade, many replacing people who change occupations or leave the labor force. These are separate measures: employment at the target date and opportunities arising during the intervening years.
The BLS methods documentation describes six linked stages, ending with occupational employment projected through industry staffing patterns. Its National Employment Matrix counts nonfarm wage and salary jobs alongside self-employed, agricultural and private-household workers. The distinction matters: the matrix is not uniformly a count of unduplicated people, and it is not a record of hours spent completing a purchasing workflow. Read the unit and employment class before comparing a national total with a local payroll or capacity measure.
Why Annual Openings Differ From Organizational Expansion
The BLS occupational separations overview distinguishes employment growth from opportunities created when people leave the labor force or transfer to another occupation. Its separations approach uses historical patterns and regression models; a change of employer within the same occupation is excluded. Neither an occupational transfer nor a labor-force exit is equivalent to a vacancy in a particular buyer team. Keep your own employer turnover, open requisitions and approved posts in a separate record, and do not read the national openings figure as an instruction to expand or shrink that team.
Growth is a small part of the projected openings
BLS expects about 53,600 openings a year for the two purchasing occupations from 2025 to 2035, and about 3,400 of them come from employment growth. Transfers to another occupation account for 30,100 and exits from the labor force, such as retirement, for 20,200, so growth supplies 6.4% of the openings here against 3.4% across all occupations. None of these counts a buyer who moves to another employer and stays a buyer, which BLS’s openings measure leaves out and a team’s own turnover includes, so a team’s vacancies belong in its own leaver records, split by reason.
Projected openings a year, 2025 to 2035 average, by source
Source: BLS Employment Projections, Table 1.10, occupational separations and openings, 2025–35. Growth is a tenth of the projected ten-year change; BLS rounds each figure, so the three parts can differ from its openings total by about 100. BLS puts the separations rate at 6.7% a year for purchasing managers and 8.3% for buyers and purchasing agents. Chart: Zinit Research. Data and method
Macroeconomic Technology Assumptions versus Direct Automation Claims
According to the BLS projections FAQ, technology enters the projections through industry productivity and occupational staffing patterns, alongside other economic influences. BLS combines quantitative analysis with qualitative review and cannot isolate one factor’s employment effect. The FAQ also discusses AI exposure categories informed by potential and observed large-language-model task performance. Those categories are not employment forecasts and do not establish that a particular organization has displaced labor. A task-performance observation, a national staffing projection and a measured change in a procurement team answer different questions; keep the evidence labels attached when comparing them.
How AI exposure and BLS’s staffing factors enter the purchasing projections
BLS places both purchasing occupations in its very high category of relative AI exposure, the highest of four, and projects them to grow 4.0% and 6.0%, faster than the +2.0% for the 206 very-high occupations together, 60 of which are projected to shrink. The note on BLS’s table says exposure “does not imply job loss, productivity gains, automation probability, or wage effects.”
Where BLS did change buyers’ staffing, Table 1.12 of the projections names the factor. In 12 public-sector industries it lowers buyers’ share of jobs, reasoning that “new technology allows workers to be more efficient” and that cooperative purchasing agreements spread, and in the 9 of them the matrix reports that share falls between 3.2% and 6.3% over the decade. In the federal government, the largest of them, buyer jobs go from 41,800 to 37,600 while all federal jobs outside the Postal Service change −3.9%. Across all industries, where the factor reads “share increases as supply chains become more complex and additional service contracts are needed”, buyers grow +6.0% against +3.5% for all jobs, and none of these figures measures what a tool did inside one team.
Projected employment change, 2025 to 2035, by BLS relative AI exposure category
Source: BLS, AI exposure categories and 2025–35 employment projections, and Table 1.12, factors affecting occupational utilization. Category changes are computed by Zinit Research from the summed employment of each category’s detailed occupations; the share change compares the occupation’s projected change with all jobs in the same industry. Chart: Zinit Research. Data and method
Diagnostic Framework: Evidence Register for Staffing and Automation Claims
| Example claim to test | Evidence to distinguish | Local record to request | Decision boundary |
|---|---|---|---|
| The national outlook tells us our team needs fewer buyers. | National occupational projection versus a local staffing proposal. | The local service scope, transaction mix, approved posts and workload baseline. | Reject the inference: the cited purchasing group is projected to grow nationally, and its outlook does not size this team. |
| Our vacancies prove that the sector is expanding. | Employer vacancies versus occupational growth and separations. | Vacancy reasons, internal transfers, employer exits and recruitment records. | Classify each local vacancy before drawing a conclusion; national openings do not explain individual vacancies. |
| After intake automation, our process became faster. | Observed elapsed time versus labor touch time and causal attribution. | Comparable completed transactions, timestamps, touch hours in every affected function, exceptions, rework and service quality. | A before-and-after record can document a change; examine alternative explanations before attributing it to the tool. |
| Agent deployment released capacity for other purchasing work. | Estimated saved time versus realized, reassigned capacity. | An explicit productive-hours convention, task-level effort records and evidence of the work actually reassigned. | Test whether effort moved elsewhere or quality changed; document the comparison design and its unresolved confounders. |
These are hypothetical claims to test, not reported findings. This authored register is a starting point for an evidence request, not a complete causal study, staffing formula or universal threshold.
Local Workflow Evidence Required for Procurement Restructuring
For a staffing proposal, request records that distinguish elapsed process time from the human effort needed to finish comparable work. Include requesters, procurement, accounts payable and other affected functions so that a proposed saving does not merely omit effort moved elsewhere. Use the guides to supplier risk assessment and supplier relationship management as planning references when assigning owners and reviewing supplier exceptions. They should help frame the local evidence request, rather than stand in for a measured automation result.
To verify whether an automated system alters staffing needs, procurement leaders should evaluate five operational questions across their workflow data:
- What changed in elapsed cycle time and human touch hours per comparable completed transaction, after accounting for volume and category mix?
- Does the effort record cover all affected functions, including requester corrections, accounts payable and system support?
- How did exceptions, rework, unresolved cases and service quality change over the same comparison period?
- What productive-hours convention is used to translate recorded effort into a capacity estimate, and is it consistent across the comparison?
- Which released hours were actually reassigned to useful work, who owns that work, and what alternative explanations remain for the observed change?
Who Remains Accountable When Automated Agents Enter the Procurement Workflow?
Analytical Scope and Statistical Program Limitations
Where does this data come from?
- What the numbers are
- BLS Employment Projections for 2025 to 2035, released August 27, 2026: employment in 2025 and projected for 2035 for purchasing managers (SOC 11-3061) and buyers and purchasing agents (13-1020, which BLS projects as one line covering wholesale and retail buyers, farm-product buyers and other purchasing agents), by industry from the National Employment Matrix; annual openings split into labor-force exits and occupational transfers (Table 1.10); the factors BLS lists for an occupation’s share of industry jobs (Table 1.12); and BLS’s relative AI exposure categories. The matrix counts wage and salary jobs, so a worker with two jobs counts twice, and self-employed workers as a separate class. Every figure is a national projection under a full-employment assumption for 2035; none describes one employer’s team.
- Source files
- Employment Projections occupation tables 1.1 to 1.12, 2025–35 (XLSX), U.S. Bureau of Labor Statistics. Retrieved October 22, 2026. Read from the Internet Archive capture of 2026-10-02 (web.archive.org/web/20261002185424/...), because bls.gov refuses automated downloads (HTTP 403). SHA-256 a7d060f0576f…
- 2025–35 National Employment Matrix (XLSX), U.S. Bureau of Labor Statistics. Retrieved October 22, 2026. Read from the Internet Archive capture of 2026-09-26 (web.archive.org/web/20260926215546/...). SHA-256 cf0a39bcefe3…
- AI exposure categories and 2025–35 employment projections (XLSX), U.S. Bureau of Labor Statistics. Retrieved October 22, 2026. Read from the Internet Archive capture of 2026-10-11 (web.archive.org/web/20261011085009/...). SHA-256 51f5779feea3…
- BLS Public Data API v1, Employment Projections series for the two occupations, U.S. Bureau of Labor Statistics. Retrieved October 22, 2026. Queried without a key on 2026-10-22; the API returns the 2025 base year only and refuses 2035 as an invalid year, so it confirms base-year employment and nothing else. SHA-256 35649c531e34…
- Occupational Outlook Handbook: Purchasing Managers, Buyers, and Purchasing Agents, U.S. Bureau of Labor Statistics. Retrieved October 22, 2026. Read from the Internet Archive capture of 2026-10-03. SHA-256 7bec2d7cba00…
- Employment Projections methods overview, U.S. Bureau of Labor Statistics. Retrieved October 22, 2026. Read from the Internet Archive capture of 2026-09-16. SHA-256 8672bb83834d…
- Occupational separations methodology, U.S. Bureau of Labor Statistics. Retrieved October 22, 2026. Read from the Internet Archive capture of 2026-09-18. SHA-256 4321f0bfbcf2…
- Employment Projections frequently asked questions, U.S. Bureau of Labor Statistics. Retrieved October 22, 2026. Read from the Internet Archive capture of 2026-10-05. SHA-256 5cb31df1885b…
- Artificial Intelligence (AI) exposure categories, U.S. Bureau of Labor Statistics. Retrieved October 22, 2026. Read from the Internet Archive capture of 2026-10-17. SHA-256 37ee175c4d11…
- What the BLS API carries
- The BLS public data API serves the matrix’s 2025 base-year employment for these series (for example EPU131020TE1000, 516.1 thousand) and refuses 2035 as an invalid year, so the projected figures come from the archived workbooks; the build checks the API’s 2025 values against them and checks the occupation totals against the Occupational Outlook Handbook page.
- How the page stays current
- BLS publishes a new projections vintage each year, in late summer, and revises the whole decade with it. Because bls.gov refuses automated downloads, this page reads Internet Archive captures of the official files and states their dates; a refresh needs a newer capture or a human download, and the page is rebuilt from whichever files are retained.
- Download the data
One ZIP holds the 3 CSV files below and a README that lists the sources and the citation.
- ai-exposure-and-projections-2025-35.csv
- purchasing-employment-by-industry-2025-35.csv
- purchasing-occupation-projections-2025-35.csv
Your data is on its way
The download should start on its own. If it doesn’t, use the link below.
Download the data (ZIP, 139 KB)- Related Zinit Research
- Cite this page
Zinit Research. “BLS Purchasing Projections: Limits for AI Staffing Decisions.” Data: U.S. Bureau of Labor Statistics, Employment Projections 2025–35. https://zinit.com/journal/research/purchasing-occupation-projection-limits, data retrieved October 22, 2026.
- How this page is produced
- Zinit Research downloads the source files listed above, keeps the originals, and computes every chart, table and figure in the data sections from them; an automated check rebuilds each number from the files before the page is published. The article text follows the Editorial Policy.
Sources
- Purchasing Managers, Buyers, and Purchasing Agents — U.S. Bureau of Labor Statistics, 2026. Current empirical evidence (official report): Separate occupational observation, projection, openings, technology assumptions and the buyer’s unresolved local workflow evidence.
- Employment Projections Methods Overview — U.S. Bureau of Labor Statistics, 2026. Foundational evidence (official report): Separate occupational observation, projection, openings, technology assumptions and the buyer’s unresolved local workflow evidence.
- Occupational Separations — U.S. Bureau of Labor Statistics, 2026. Foundational evidence (official report): Separate occupational observation, projection, openings, technology assumptions and the buyer’s unresolved local workflow evidence.
- Employment Projections Frequently Asked Questions — U.S. Bureau of Labor Statistics, 2026. Foundational evidence (official report): Separate occupational observation, projection, openings, technology assumptions and the buyer’s unresolved local workflow evidence.