Forrester Predictions 2026 — How AI Agents Reshape Enterprise Software

1. Three Numbers Boards Must Remember — From Forrester Predictions 2026

Forrester's three core predictions all carry direct implications for IT/HR budgeting and vendor selection at Thai organizations:

Prediction	What It Means	What the Board Should Watch
1. Top 5 HCM platforms will offer Digital Employee Management	HR tech becomes the "sophisticated employee system of record" tracking and optimizing a hybrid workforce — humans + AI agents in one place	Ask the CHRO: do we have a policy for managing AI agents like employees? (job description, access rights, KPIs, performance reviews)
2. 30% of enterprise app vendors will launch MCP servers	MCP = an open standard letting external AI agents collaborate across vendor platforms — Forrester calls it "an open ecosystem where businesses aren't locked into a single AI provider"	Ask the CIO: which of our vendors has an MCP server on the roadmap? — none = future "data island" that AI can't reach
3. 50% of ERP vendors will release autonomous governance modules	Combining explainable AI + automated audit trails + real-time compliance — half of ERP vendors will have it, half won't	Ask the CIO + DPO: can our ERP justify every AI decision? + is the audit log tamper-proof?

Forrester's quiet implication — "half of ERP vendors will have a governance module" means the other half won't. In 2026, the laggards lose ground on regulatory compliance — particularly the Thai PDPA Crackdown 2025-2026 that's already imposing real fines.

2. Service-as-Software — The Business Model That's Shifting

Traditionally, enterprise software is sold by per-seat license — pay per human user. Once AI agents start doing the work, that model penalizes the vendor (fewer headcount = less revenue). Constellation Research observes that vendors are now experimenting with "Agentic Enterprise License Agreements (AELAs)" — Salesforce, for instance, calls its approach "all you can eat with Agentforce" — a flat fee with unlimited usage, giving enterprises budget predictability.

Model	How It's Priced	Implication for the Board / CFO
Per-seat license (legacy)	User count × annual fee	Predictable budgeting — but headcount cuts mean fewer seats sold → vendor's interest no longer aligns with the customer
Per-action / Consumption	Pay per task an AI agent performs	Overage risk — hard to forecast, especially when agents misfire and retry
AELA (flat-fee, "all you can eat")	Fixed annual platform fee, unlimited usage	Higher list price but predictable — Salesforce's Agentforce pioneered this
Service-as-Software (outcome)	Pay per outcome (e.g., invoices closed, tickets resolved)	Best alignment — but requires the vendor to take outcome risk; still in early prototype today

Practical implication for Thai boards — vendor proposals in 2026-2027 will not be straightforward "users × annual fee". You'll need to compare more complex pricing tiers and define your ROI threshold before signing, not after — see 90% of AI investments fail to break even for the full argument.

3. Governance Becomes a Board-Level Issue — Not an IT Task

Forrester is explicit — "tech leaders should evaluate vendors' governance roadmaps immediately" — at higher priority than functional capability. The reason — when an AI agent makes autonomous decisions, who carries the legal liability? Under Thai law, the executive and the organization are liable, not the vendor.

This aligns with the Stanford AI Index 2026 finding — 89% of AI agent implementations don't reach production. Not because AI isn't capable, but because ERP governance is the bottleneck. Read AI Adoption Gap 2026 for the full breakdown.

Governance Issue	Board Decision Required	Further Reading
Explainable AI	Every agent decision must be explainable — who validates this?	AI Governance
Tamper-proof Audit Trail	Log every agent action — immutable + retained 1+ year	Agentic AI
Cyber Risk from AI Agents	Agents with access to financial/HR data are new attack surfaces	Cyber Top Business Risk
Vendor Lock-in vs Open Standards	Pick MCP-supporting vendors — to swap AI providers without migration	Claude Managed Agents
Compliance + PDPA	Agents handling personal data = data processors under PDPA	PDPA Crackdown 2026

4. Workforce in 2026 — HR Must Rethink

Forrester frames this shift as — "from enabling employees with digital tools to accommodating a digital workforce of AI agents." HR must start treating AI agents as "employees" who need:

• A job title + scope — what the agent can and cannot do
• Access rights — same RBAC discipline as human employees
• KPIs + performance reviews — what % of tasks does this agent close?
• An offboarding process — retire the agent + revoke access immediately
• A human supervisor — Forrester is clear: "full autonomous business units without human involvement remain years away"

Importantly — Forrester emphasizes "role-based" agents working across systems become the new standard, replacing the old "task-based" agents stuck on a single workflow. So HR + IT must plan jointly, not in silos.

5. The Remaining Bottleneck — Process Standardization + Data Fragmentation

For years, everyone assumed AI's bottleneck was "compute" and "storage". Forrester is explicit — those two have been solved. What's left is:

Real 2026 Bottleneck	Source	What ERP Must Solve
Business Process Standardization	Each department runs its own workflow → AI doesn't know whose process to automate	An ERP that enforces standardized workflow + uniform approvals across the organization
Data Fragmentation	Data scattered across systems + Excel — agents must reconcile = slow + error-prone	An ERP storing HR + Finance + Inventory + Sales in a single database
Connection / Integration Cost	Constellation Research flags vendors charging "data access tolls" for external agents (Salesforce, SAP have active litigation)	Pick an ERP with free APIs + open MCP server support

The implication — before investing in AI agents, fix data + process inside your ERP first. See Data Fragmentation and AI in ERP 2026 for what an "AI-ready" ERP must look like.

6. Questions the Board Must Ask CIO + CFO + CHRO

Forrester's recommendation — boards must move from "listening to reports" to "asking the right questions." This is the checklist for the first board meeting of 2026:

Audience	Question	"Pass" Criteria
CIO	Do our core vendors (ERP, HR, CRM) have an MCP server roadmap within 12 months?	Documented roadmap + timeline + supplied to the board
CIO	Does our ERP have an autonomous governance module today?	Audit log + explainable AI + real-time compliance — all present
CFO	If the vendor moves to AELA pricing, what is next year's cost?	Active vendor dialogue + worst/best case scenarios prepared
CFO	Do we capitalize or expense AI agent costs? + is there an ROI threshold per project?	Documented policy + ROI threshold signed by CEO before each project
CHRO	Do we have a policy for managing AI agents like employees? (title, access, KPIs)	Agent job descriptions + RBAC policy + 6-monthly reviews
CHRO	Do we have an upskill/reskill timeline for human employees?	Training budget + attendance KPIs + selected courses

7. The Thai Lens — Why Forrester's Predictions Hit Home

Forrester wrote for global enterprises — but three points hit Thailand directly:

• 1. PDPA enforcement is accelerating — Forrester's autonomous governance prediction maps almost 1:1 onto the PDPC's 8-fine enforcement wave in April-August 2025. Thai organizations face a tighter timeline than the global average.
• 2. Cyber risk is the top business risk in Thailand — Allianz Risk Barometer 2026 ranks cyber #1 for Thailand. AI agents without governance become a new entry point for attackers (see Cyber Top Business Risk Thailand 2026).
• 3. Foreign vendor pricing ≠ what Thai customers can absorb — flat-fee structures like Salesforce Agentforce often start higher than mid-sized Thai companies' budgets — making Thai-built, AI-ready ERPs a more reasonable choice.

Additional context — see Anthropic Enterprise AI Firm $1.5B as a real-world case study of the "service-as-software" model Forrester predicts. Anthropic isn't selling licenses — they're selling outcomes from AI agents working inside enterprises.

Summary

Area	What Boards Must Do in 2026
3 Forrester numbers	Top 5 HCM = Digital Employee Mgmt, 30% vendors = MCP, 50% ERP vendors = governance module
Pricing	Per-seat → AELA / outcome-based — CFO must engage vendors before any 2026 signing
Governance	Now a board-level issue — every vendor must show an explainable-AI roadmap
HR	Manage AI agents like employees — job description, RBAC, KPIs, supervisor
Data Fragmentation	Fix ERP + processes before investing in AI agents — otherwise 89% never reach production
Thai lens	PDPA + cyber risk compress the timeline — AI-ready ERP shifts from "nice-to-have" to "must-have"

"Forrester's 2026 predictions aren't really about 'AI is coming' — they're about 'the business model behind enterprise software is changing — and boards must prepare.' The Top 5 HCM platforms will offer Digital Employee Management, 30% of vendors will open MCP servers, 50% of ERP vendors will release governance modules — which means the other 50% won't. Executives who don't pick vendors with governance + open standards risk becoming the organization that AI agents can't reach and PDPA enforcement catches off guard. Start with the six questions in your first 2026 board meeting."

References

• Forrester — Predictions 2026: AI Agents Changing Business Models and Workplace Culture Impact Enterprise Software
• Constellation Research — Enterprise Technology 2026: 15 AI, SaaS, Data Business Trends to Watch
• Stanford HAI — The 2026 AI Index Report
• ERP Today — AI Is Expanding, But ERP Systems May Not Be Ready (Stanford AI Index 2026)