AI will not save lazy SaaS. And I say that as a founder who has spent a decade building SaaS for key account management and eventually killed it to rebuild an AI-first key account management software.
The uncomfortable truth about SaaS and Key Account Management
When we started DemandFarm, our tools/products were like most enterprise SaaS tools of that era: sophisticated systems of record for Sales Leaders. Leaders loved them because they got governance, structure, and reports. Users hated them because they had no choice but to do manual data entry.
In key account management, that meant one thing: account managers spent more time feeding the system than getting help from it. The product behaved like a compliance layer, not a guide or a thinking partner. It was great for “What happened so far?” but terrible for “What should I do next in this account?”
You could say a good key account manager should anyway know their next best move. But that is true for the top 10-20%. I always believed in something bigger: Democratizing Account Management. Shrinking the gap between the best and the average AM. Spreading winning patterns across all key accounts so growth is not a happy accident in a few logos, but a repeatable outcome across the strategic accounts portfolio.
Then AI became real enough to matter, not just as a buzzword. And it forced a hard question in my mind: if we build on the same foundations and just add an AI wrapper, are we actually changing anything?
My answer: no.
Why “AI wrappers” for SaaS are a dead end
The easy path for any SaaS founder right now is obvious: keep the same product, sprinkle some AI:
-> Add a copilot bubble in the corner
-> Auto-generate a few summaries
-> Put “AI-powered” on the website
But if the core product is still a system of record, AI becomes cosmetics. The workflow is unchanged. The user is still typing, cleaning, tagging, and updating. AI just makes the screen look modern.
For key account management, that’s a waste.
Account managers are already juggling politics, relationships, renewals, expansions, product complexity, and internal stakeholders. They don’t need yet another place to document reality. They need a system that understands the account with them and nudges them toward better moves.
That’s where the idea crystallized for me:
-> SaaS gave sales leadership systems of record.
-> AI must give account managers/users systems of insight and action.
And if that is true, then many of our own products had to eventually die.
Deciding to “kill” our own product
Inside DemandFarm, this was not a theoretical discussion. We had a mature product line, paying customers, and years of UX patterns built for a pre‑AI world.
But every time I looked at how account managers actually worked, it bothered me:
-> They still ran to PowerPoint for QBRs.
-> They still kept “real” notes and strategies in their own docs.
-> They updated the product mainly before reviews, not as a natural part of their weekly rhythm.
If AI was going to be central, not decorative, we had two choices:
1. Retrofit AI into the old flows
2. Start again and design for an AI-first world
Retrofitting would make us feel productive but not change the nature of work. So we chose the painful route: rebuild DemandFarm Kampanion from scratch and be willing to kill patterns that had “worked” for years.
That decision changed everything.
How I redesigned Kampanion from the ground up
When we started redesigning Kampanion, I kept three constraints in my head:
1. Assume the account story already exists in scattered data and tools.
2. Assume the account manager’s time and attention are the scarcest resources.
3. Assume leadership wants the real truth, not staged theatre.
From there, the design question became very simple:
What if the account manager never typed a single thing into the system? What could AI figure out on its own?
That forced a very different product:
-> The starting point is data ingestion, not input forms for users.
-> The default mode is “suggest, build, and ask for correction,” not “ask for inputs.”
-> The primary artifacts are insights and actions, not fields and sections.
Examples of how this showed up in Kampanion’s design:
-> Instead of “Build an account plan,” Kampanion auto-assembles a living account narrative from CRM data, deals, products, emails, meetings, conversation intelligence, and external signals. The AM edits and adds nuance; they don’t start on a blank canvas.
-> Instead of “Draw an org chart and relationship map,” Kampanion proposes a relationship map based on roles, touchpoints, and interaction history, and suggests additional contacts to be continuously added to the canvas. The AM fixes it where the machine is wrong.
-> Instead of “Fill a white-space matrix,” Kampanion surfaces buying centers, where similar customers buy more, where adoption is shallow, where engagement is dropping, and turns those into “plays” the AM can pick up.
The mental model flipped from:
“System demands data, then gives reports”
to
“System mines data, then proposes moves.”
From system of record to system of insights and actions
The old world of KAM tools looked like this:
-> At the start of the year, everyone fills plans to satisfy the process.
-> Mid-year, reality diverges; plans stay static.
-> Before QBRs, teams scramble to reconcile reality with the tool and slides.
In that world, the “truth” of the account lived in decks, side conversations, and the AM’s head. The product was mostly an after-the-fact documentation tool.
With Kampanion rebuilt as AI-first, the loop I wanted was very different:
-> The system keeps learning about the account every week.
-> It flags risks and opportunities as they emerge, not at quarter-end.
-> It keeps a single, living, breathing narrative of the account that both AMs and leaders can trust.
You should be able to ask at any time:
-> “Where is this account likely to grow next?”
-> “Which relationships are strategic and at risk?”
-> “What are the goals and objectives that I should set for an account for the year?”
-> “Which commitments did we make last QBR and what changed since?”
And the system should answer from live data, not static plans in outdated slides.
That’s what I mean by a “system of insights and actions.” It’s not just dashboards with charts. It is a product that constantly:
-> Interprets what is happening in the account
-> Translates that into risk, opportunity, and priority
-> Suggests concrete actions or conversations to run next
The account manager’s job becomes: judge, refine, and execute. Not: collect, enter, and reconcile.
How Kampanion has evolved since the rebuild
The first AI-first Kampanion was, frankly, opinionated. It did a few things very deeply rather than trying to replicate every old screen.
Over time, three big evolutions have happened:
1. Deeper context, less asking
As we integrated more sources and refined the models, Kampanion needed fewer explicit inputs. It could infer more: buying centers, stakeholder influence, product penetration anomalies, early risk signals. The direction is clear: every release should reduce manual effort, not add “one more thing to fill.”
2. Closer to where work actually happens
Instead of being a parallel universe, Kampanion now lives closer to CRM and daily workflows. That was important to me. AI should not require a ritual: “Now I will go to the AI product.” It should be part of the natural motion of updating an opportunity, preparing for a meeting, or planning a quarter.
3. From single-account assistant to portfolio intelligence
Once we had an AI view at the account level, the natural next step was: how does this look across 20, 50, 100 key accounts? Leaders care about patterns:
-> Which accounts are “quietly at risk”?
-> Where are we under-invested despite strong potential?
-> Which AMs are consistently turning insights into action?
Kampanion had to serve that leadership lens without falling back into “governance-first” mode. The trick has been to keep the user we optimize for as the AM, while still giving leadership the visibility they need.
What this means for key account management
If you are leading KAM today, here is my blunt view:
-> If your “AI strategy” is just adding a copilot to a legacy system-of-record, you will get short-term excitement and long-term disappointment.
-> If your account managers still depend on slides and spreadsheets for the “real” story, your platform is not yet their system of insight.
-> If using the tool feels like a tax rather than an advantage, it is built for leaders, not for users.
An AI-native KAM product must:
-> Reduce data entry dramatically.
-> Continuously build, update, and refine the account plan on its own.
-> Turn complexity (orgs, politics, products, global scope) into a clear set of next best actions.
-> Make both AMs and sales leaders feel like they are looking at the same, live, real truth.
When I say “we have to kill our existing products,” I don’t mean being reckless. I mean being honest about what belongs to the pre‑AI era and having the courage to redesign from first principles, even if it hurts in the short term.
Kampanion is my attempt to do exactly that in key account management: move from policing to partnering, from records to insights, from reporting the past to shaping the next move.
